Article(id=1322962977804136796, tenantId=1146029695717560320, journalId=1249024232475115590, issueId=1322962941087195876, articleNumber=null, orderNo=null, doi=10.11834/jig.250521, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1761580800000, receivedDateStr=2025-10-28, revisedDate=1772380800000, revisedDateStr=2026-03-02, acceptedDate=null, acceptedDateStr=null, onlineDate=1793348553371, onlineDateStr=2026-10-30, pubDate=1792080000000, pubDateStr=2026-10-16, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1793348553371, onlineIssueDateStr=2026-10-30, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1793348553371, creator=13701087609, updateTime=1793348553371, updator=13701087609, issue=Issue{id=1322962941087195876, tenantId=1146029695717560320, journalId=1249024232475115590, year='2026', volume='31', issue='10', pageStart='3441', pageEnd='3818', issueExtLink='null', onlineDate='null', pubDate='1792080000000', pubDateStr='2026-10-16', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=1, specialIssue=null, createTime=1793348544618, creator='13701087609', updateTime=1793348641113, updator='13701087609', preIssue=null, nextIssue=null, articleTotal=null, ext={EN=IssueExt(id=1322963345913021357, tenantId=1146029695717560320, journalId=1249024232475115590, issueId=1322962941087195876, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1322963345913021358, tenantId=1146029695717560320, journalId=1249024232475115590, issueId=1322962941087195876, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null, downloadFileDto=null}, startPage=3590, endPage=3605, ext={EN=ArticleExt(id=1322962978051600733, articleId=1322962977804136796, tenantId=1146029695717560320, journalId=1249024232475115590, language=EN, title=Methods and a benchmark dataset for mountain road extraction from high-resolution remote-sensing imagery: AsiaMountain-Road, columnId=1322961445507440644, journalTitle=Journal of Image and Graphics, columnName=Dataset, runingTitle=null, highlight=null, articleAbstract=
Objective

Mountain roads are vital infrastructure for transportation,logistics,and socioeconomic development in remote,topographically challenging regions. They provide essential connectivity for education,healthcare,resource transport,and disaster relief,yet their construction and maintenance are extremely difficult because of steep slopes,unstable geology,and frequent natural hazards. Despite their importance,automatically extracting mountain roads from high-resolution remote-sensing imagery remains a longstanding challenge. Complex terrain introduces severe occlusion and illumination variation,while the spectral appearance of roads is easily confused with bare rock,dry riverbeds,and agricultural paths. In addition,mountain roads are typically narrow,winding,and discontinuous,which differs substantially from the wide,regular patterns found in urban or rural road networks. Existing public road-extraction datasets,such as SpaceNet and DeepGlobe,were developed primarily for flat or urbanized environments. As a result,they fail to capture the spectral,geometric,and contextual complexity of mountain roads,leading to poor generalization when deep-learning models are transferred to high-relief regions. To bridge this gap,we construct AsiaMountain-Road,the first high-resolution,multi-country benchmark dataset dedicated to mountain-road extraction,and propose a novel multi-resolution adaptive feature dual aggregation network (MR-AFDANet) to evaluate and enhance model adaptability in diverse mountainous environments.

Method

The AsiaMountain-Road dataset was built from 29 high-resolution satellite-image areas selected from representative mountainous regions in 17 Asian countries,covering a total of 1 836 km² and providing detailed surface information for fine-scale mapping. All imagery underwent radiometric calibration,orthorectification,and mosaicking,followed by manual annotation by trained experts. Each region was labeled with fine-grained binary road masks to ensure pixel-level accuracy. To improve usability,auxiliary metadata,including elevation maps and land-cover references,were also provided. For algorithmic evaluation,we benchmark several mainstream deep-learning models—D-LinkNet,OareNet,C2Net,AFDANet,RCFSNet,and RS-Mamba—to assess their adaptability to mountainous terrain. Building on these findings,we propose MR-AFDANet,a model specifically designed for complex terrain mapping. MR-AFDANet uses a ResNet34 encoder as its backbone,consisting of four downsampling convolutional blocks. Each block is followed by a SEAF (spatial-enhanced adaptive feature fusion) module,which adaptively fuses intra-block and residual features while strengthening spatial information through multi-scale convolutions and attention mechanisms. In the decoder,an strip refinement attention decoder (SRADecoder) progressively fuses multi-resolution features to refine road boundaries,preserve continuity in narrow and winding segments,and suppress background noise caused by shadows and vegetation. This design enables the network to preserve thin road topology even under severe spectral interference.

Result

Experimental evaluation on the AsiaMountain-Road benchmark demonstrates the substantial difficulty of road extraction in mountainous regions. Compared with urban or plain datasets,mountain imagery contains more complex topography,stronger illumination variation,and more severe occlusion from vegetation and terrain shadows,all of which significantly degrade segmentation accuracy. Across six state-of-the-art baselines,all models show varying degrees of road omission and fragmentation,particularly on narrow,winding,and shaded mountain roads. The results further show that models trained only on source-domain data struggle to generalize to the mountainous target domain because of large distributional discrepancies. Quantitative comparisons indicate that MR-AFDANet achieves the best overall performance,obtaining the highest intersection over union(IoU),F1-score,and recall among all evaluated methods. These improvements confirm that combining multi-scale feature fusion with the SRADecoder enhances both global consistency and fine-grained boundary precision. Visual analysis further verifies this advantage: MR-AFDANet produces more continuous and complete road extractions under severe occlusion,maintains connectivity along steep and curved routes,and reduces false detections on spectrally similar surfaces.

Conclusion

We present AsiaMountain-Road,the first high-resolution,multi-country dataset dedicated to road extraction in mountainous regions. Covering 29 representative areas across 17 Asian countries,it provides detailed road-mask annotations together with auxiliary contextual and topographic data,filling an important gap in current remote-sensing resources. Baseline experiments show that existing state-of-the-art methods struggle with narrow,winding,and occluded roads in complex terrain. To address these challenges,MR-AFDANet integrates SEAF modules,a dual-resolution feature-fusion mechanism,and the SRADecoder to recover thin road structures more precisely and improve robustness across highly variable mountain landscapes. Overall,AsiaMountain-Road narrows the gap between real-world geospatial challenges and current computer-vision capabilities by offering both a demanding benchmark for algorithmic innovation and a valuable resource for applied geospatial analysis. The dataset and code associated with this paper are available at https://cstr.cn/31253.11.sciencedb.j00240.00109 and https://github.com/wyl-ucas/AsianMountain-DataSet.

, authors=Yilin Wang1, 2, Yu Meng1, 2, *, Zhihao Xi1, Yupeng Deng1, Diyou Liu1, Jingbo Chen1, 2, authorsList=Yilin Wang, Yu Meng, Zhihao Xi, Yupeng Deng, Diyou Liu, Jingbo Chen, authorCompany=null, correspAuthors=Yu Meng, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, fund=null), CN=ArticleExt(id=1322962984611492204, articleId=1322962977804136796, tenantId=1146029695717560320, journalId=1249024232475115590, language=CN, title=高分辨率遥感影像山区道路提取方法与基准数据集:AsiaMountain-Road, columnId=1322961445683601414, journalTitle=中国图象图形学报, columnName=数据集, runingTitle=null, highlight=null, articleAbstract=
目的

针对高分辨率遥感影像中山区道路因地形复杂、形态多样及背景干扰严重而难以准确提取的问题,以及现有道路提取数据集对山区道路适应性不足的现状,构建了首个覆盖亚洲多国典型山区的高分辨率遥感影像山区道路数据集(AsiaMountain-Road),并提出多分辨率自适应双聚合网络(multi-resolution adaptive feature dual aggregation network,MR-AFDANet)以支持山区道路自动提取算法的研究与应用。

方法

数据集从亚洲多国典型山区选取了29个高分辨率遥感影像区域,总覆盖面积达1 836 km²。结合道路稀疏、弯曲以及遮挡等复杂场景特征进行采样与标注,最终形成涵盖多种地貌和道路类型的山区道路影像与精细标注。基于该数据集,对多种主流遥感道路提取算法进行了系统基准评测,并验证了所提MR-AFDANet模型在复杂山区环境下的分割性能。

结果

实验表明,现有主流道路提取方法在AsiaMountain-Road数据集上表现出显著差异,其性能受到山区道路狭窄、弯曲、遮挡及多尺度特征的影响较大。该数据集有效揭示了现有算法在复杂山区环境中的局限性,并为算法鲁棒性提升与结构优化提供了基准支持。

结论

AsiaMountain-Road数据集填补了高分辨率遥感影像山区道路提取领域的基准数据空白,MR-AFDANet模型有效解决了复杂地形下的道路断裂与漏检问题。两者共同为山区遥感影像的智能化解译提供了高质量的数据基础与方法参考。论文相关数据集与代码下载地址:https://cstr.cn/31253.11.sciencedb.j00240.00109和https://github.com/wyl-ucas/AsianMountain-DataSet。

, authors=王怡琳1, 2, 孟瑜1, 2, *, 席智浩1, 邓毓弸1, 刘帝佑1, 陈静波1, 2, authorsList=王怡琳, 孟瑜, 席智浩, 邓毓弸, 刘帝佑, 陈静波, authorCompany=null, correspAuthors=孟瑜, authorNote=

王怡琳,女,硕士研究生,主要研究方向为遥感图像语义分割。E-mail:

孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:

席智浩,男,助理研究员,主要研究方向为计算机视觉、领域自适应和遥感图像解译。E-mail:

邓毓弸,男,特别研究助理,主要研究方向为遥感与变化检测。E-mail:

刘帝佑,男,特别研究助理,主要研究方向为高分辨率遥感智能解译与矢量制图。E-mail:

陈静波,男,研究员,主要研究方向为遥感图像智能处理和遥感大数据分析。E-mail:

, correspAuthorsNote=
孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=JGT1HF9Qubb/5SYKQCW78Q==, magXml=b9EumHv3qHr7nU7bwVf4cw==, pdfUrl=null, pdf=WsfvMlQh2tMiKddaeU12Ow==, pdfFileSize=13185275, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=oMcE+c7DldycO97WGlDtGg==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=tLEoYsoGTTQfKwpqjh+i6g==, mapNumber=null, fund=null)}, authors=[Author(id=1322962985056088436, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=wangyilin23@mails.ucas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962985135780215, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985056088436, language=EN, stringName=Yilin Wang, firstName=Yilin, middleName=null, lastName=Wang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985253220728, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985056088436, language=CN, stringName=王怡琳, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 中国科学院空天信息创新研究院,北京 100094
2 中国科学院大学电子电气与通信工程学院,北京 100049, bio={"content":"

王怡琳,女,硕士研究生,主要研究方向为遥感图像语义分割。E-mail:

"}, bioImg=null, bioContent=

王怡琳,女,硕士研究生,主要研究方向为遥感图像语义分割。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)]), AuthorCompany(id=1322962984963813744, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=2, ext=[AuthorCompanyExt(id=1322962984972202353, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China), AuthorCompanyExt(id=1322962984980590962, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国科学院大学电子电气与通信工程学院,北京 100049)])]), Author(id=1322962985328718202, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=mengyu@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1322962985404215677, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985328718202, language=EN, stringName=Yu Meng, firstName=Yu, middleName=null, lastName=Meng, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, *, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985475518846, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985328718202, language=CN, stringName=孟瑜, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, *, address=1 中国科学院空天信息创新研究院,北京 100094
2 中国科学院大学电子电气与通信工程学院,北京 100049, bio={"content":"

孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:

"}, bioImg=null, bioContent=

孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)]), AuthorCompany(id=1322962984963813744, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=2, ext=[AuthorCompanyExt(id=1322962984972202353, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China), AuthorCompanyExt(id=1322962984980590962, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国科学院大学电子电气与通信工程学院,北京 100049)])]), Author(id=1322962985538433408, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=xizh@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962985605542274, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985538433408, language=EN, stringName=Zhihao Xi, firstName=Zhihao, middleName=null, lastName=Xi, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985681039747, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985538433408, language=CN, stringName=席智浩, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国科学院空天信息创新研究院,北京 100094, bio={"content":"

席智浩,男,助理研究员,主要研究方向为计算机视觉、领域自适应和遥感图像解译。E-mail:

"}, bioImg=null, bioContent=

席智浩,男,助理研究员,主要研究方向为计算机视觉、领域自适应和遥感图像解译。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)])]), Author(id=1322962985773314437, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=dengyp@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962985853006215, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985773314437, language=EN, stringName=Yupeng Deng, firstName=Yupeng, middleName=null, lastName=Deng, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985920115080, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985773314437, language=CN, stringName=邓毓弸, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国科学院空天信息创新研究院,北京 100094, bio={"content":"

邓毓弸,男,特别研究助理,主要研究方向为遥感与变化检测。E-mail:

"}, bioImg=null, bioContent=

邓毓弸,男,特别研究助理,主要研究方向为遥感与变化检测。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)])]), Author(id=1322962986004001162, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=liudiyou@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962986079498636, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986004001162, language=EN, stringName=Diyou Liu, firstName=Diyou, middleName=null, lastName=Liu, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962986150801805, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986004001162, language=CN, stringName=刘帝佑, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国科学院空天信息创新研究院,北京 100094, bio={"content":"

刘帝佑,男,特别研究助理,主要研究方向为高分辨率遥感智能解译与矢量制图。E-mail:

"}, bioImg=null, bioContent=

刘帝佑,男,特别研究助理,主要研究方向为高分辨率遥感智能解译与矢量制图。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)])]), Author(id=1322962986226299279, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=5, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=chenjb@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962986314379666, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986226299279, language=EN, stringName=Jingbo Chen, firstName=Jingbo, middleName=null, lastName=Chen, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962986377294227, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986226299279, language=CN, stringName=陈静波, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 中国科学院空天信息创新研究院,北京 100094
2 中国科学院大学电子电气与通信工程学院,北京 100049, bio={"content":"

陈静波,男,研究员,主要研究方向为遥感图像智能处理和遥感大数据分析。E-mail:

"}, bioImg=null, bioContent=

陈静波,男,研究员,主要研究方向为遥感图像智能处理和遥感大数据分析。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)]), AuthorCompany(id=1322962984963813744, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=2, ext=[AuthorCompanyExt(id=1322962984972202353, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China), AuthorCompanyExt(id=1322962984980590962, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国科学院大学电子电气与通信工程学院,北京 100049)])])], keywords=[Keyword(id=1322962986482151828, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, orderNo=1, keyword=mountainous road extraction), Keyword(id=1322962986545066389, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, orderNo=2, keyword=high-resolution remote sensing), Keyword(id=1322962986612175254, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, orderNo=3, keyword=remote sensing benchmark dataset), Keyword(id=1322962986687672727, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, orderNo=4, keyword=complex terrain road detection), Keyword(id=1322962986767364504, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, orderNo=5, keyword=deep learning semantic segmentation), Keyword(id=1322962986838667673, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, orderNo=1, keyword=山区道路提取), Keyword(id=1322962986888999322, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, orderNo=2, keyword=高分辨率遥感), Keyword(id=1322962986951913883, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, orderNo=3, keyword=遥感基准数据集), Keyword(id=1322962987023217052, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, orderNo=4, keyword=复杂地形道路检测), Keyword(id=1322962987086131613, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, orderNo=5, keyword=深度学习语义分割)], refs=[Reference(id=1322962991305601466, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2020, volume=12, issue=9, pageStart=null, pageEnd=1444, url=null, language=null, rfNumber=null, rfOrder=0, authorNames=Abdollahi A, Pradhan B, Shukla N, Chakraborty S, Alamri A, journalName=Remote Sensing, refType=null, unstructuredReference=Abdollahi A,Pradhan B,Shukla N,Chakraborty S and Alamri A. 2020. Deep learning approaches applied to remote sensing datasets for road extraction: a state-of-the-art review. Remote Sensing,12(9): #1444 [DOI: 10.3390/rs12091444], articleTitle=Deep learning approaches applied to remote sensing datasets for road extraction: a state-of-the-art review, refAbstract=null), Reference(id=1322962991368516027, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=4720, pageEnd=4728, url=null, language=null, rfNumber=null, rfOrder=1, authorNames=Bastani F, He S T, Abbar S, Alizadeh M, Balakrishnan H, Chawla S, journalName=null, refType=null, unstructuredReference=Bastani F,He S T,Abbar S,Alizadeh M,Balakrishnan H,Chawla S,et al. 2018. RoadTracer: automatic extraction of road networks from aerial images//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City,USA: IEEE: 4720-4728 [DOI: 10.1109/CVPR.2018.00496], articleTitle=RoadTracer: automatic extraction of road networks from aerial images, refAbstract=null), Reference(id=1322962991444013500, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=39, issue=5, pageStart=658, pageEnd=667, url=null, language=null, rfNumber=null, rfOrder=2, authorNames=Chen R N, Peng L, Liu Y F, Wei Z C, Lyu B R, Chen D Y, journalName=Journal of University of Chinese Academy of Sciences, refType=null, unstructuredReference=Chen R N,Peng L,Liu Y F,Wei Z C,Lyu B R and Chen D Y. 2022. Road information extraction and application in the suburban mountainous area based on remote sensing images. Journal of University of Chinese Academy of Sciences,39(5): 658-667, articleTitle=Road information extraction and application in the suburban mountainous area based on remote sensing images, refAbstract=null), Reference(id=1322962991548871101, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=39, issue=5, pageStart=658, pageEnd=667, url=null, language=null, rfNumber=null, rfOrder=3, authorNames=陈若男, 彭玲, 刘玉菲, 卫志超, 吕蓓茹, 陈德跃, journalName=中国科学院大学学报, refType=null, unstructuredReference=陈若男,彭玲,刘玉菲,卫志超,吕蓓茹,陈德跃. 2022. 引入空间距离信息的城郊山区道路提取与应用. 中国科学院大学学报,39(5): 658-667 [DOI: 10.7523/j.ucas.2021.0004], articleTitle=引入空间距离信息的城郊山区道路提取与应用, refAbstract=null), Reference(id=1322962993222398398, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2016, volume=13, issue=4, pageStart=545, pageEnd=549, url=null, language=null, rfNumber=null, rfOrder=4, authorNames=Cheng G L, Zhu F Y, Xiang S M, Pan C H, journalName=IEEE Geoscience and Remote Sensing Letters, refType=null, unstructuredReference=Cheng G L,Zhu F Y,Xiang S M and Pan C H. 2016. Road centerline extraction via semisupervised segmentation and multidirection nonmaximum suppression. IEEE Geoscience and Remote Sensing Letters,13(4): 545-549 [DOI: 10.1109/2016.2524618], articleTitle=Road centerline extraction via semisupervised segmentation and multidirection nonmaximum suppression, refAbstract=null), Reference(id=1322962993297895871, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=172, pageEnd=183, url=null, language=null, rfNumber=null, rfOrder=5, authorNames=Demir I, Koperski K, Lindenbaum D, Pang G, Huang J, Basu S, journalName=null, refType=null, unstructuredReference=Demir I,Koperski K,Lindenbaum D,Pang G,Huang J,Basu S,et al. 2018. DeepGlobe 2018: a challenge to parse the earth through satellite images//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Salt Lake City,USA: IEEE: 172-183 [DOI: 10.1109/CVPRW.2018.00031], articleTitle=DeepGlobe 2018: a challenge to parse the earth through satellite images, refAbstract=null), Reference(id=1322962993385976256, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2025, volume=22, issue=null, pageStart=null, pageEnd=6010505, url=null, language=null, rfNumber=null, rfOrder=6, authorNames=Deng X L, Li M Z, Chen Z H, journalName=IEEE Geoscience and Remote Sensing Letters, refType=null, unstructuredReference=Deng X L,Li M Z and Chen Z H. 2025. AFDANet: an adaptive full-stage feature fusion and directional-aware network for road extraction in remote sensing images. IEEE Geoscience and Remote Sensing Letters,22: #6010505 [DOI: 10.1109/LGRS.2025.3583214], articleTitle=AFDANet: an adaptive full-stage feature fusion and directional-aware network for road extraction in remote sensing images, refAbstract=null), Reference(id=1322962993453085121, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2020, volume=58, issue=9, pageStart=6494, pageEnd=6508, url=null, language=null, rfNumber=null, rfOrder=7, authorNames=Ding L, Bruzzone L, journalName=IEEE Transactions on Geoscience and Remote Sensing, refType=null, unstructuredReference=Ding L and Bruzzone L. 2020. DiResNet: direction-aware residual network for road extraction in VHR remote sensing images. IEEE Transactions on Geoscience and Remote Sensing,58(9): 6494–6508 [DOI: 10.1109/TGRS.2020.3034011], articleTitle=DiResNet: direction-aware residual network for road extraction in VHR remote sensing images, refAbstract=null), Reference(id=1322962993528582594, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2020, volume=12, issue=18, pageStart=null, pageEnd=2985, url=null, language=null, rfNumber=null, rfOrder=8, authorNames=Lin Y N, Xu D Y, Wang N, Shi Z, Chen Q X, journalName=Remote Sensing, refType=null, unstructuredReference=Lin Y N,Xu D Y,Wang N,Shi Z and Chen Q X. 2020. Road extraction from very-high-resolution remote sensing images via a nested SE-Deeplab model. Remote Sensing,12(18): #2985 [DOI: 10.3390/rs12182985], articleTitle=Road extraction from very-high-resolution remote sensing images via a nested SE-Deeplab model, refAbstract=null), Reference(id=1322962993604080067, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2024, volume=16, issue=12, pageStart=null, pageEnd=2056, url=null, language=null, rfNumber=null, rfOrder=9, authorNames=Liu R Y, Wu J H, Lu W Y, Miao Q G, Zhang H, Liu X Z, journalName=Remote Sensing, refType=null, unstructuredReference=Liu R Y,Wu J H,Lu W Y,Miao Q G,Zhang H,Liu X Z,et al. 2024. A review of deep learning-based methods for road extraction from high-resolution remote sensing images. Remote Sensing,16(12): #2056 [DOI: 10.3390/rs16122056], articleTitle=A review of deep learning-based methods for road extraction from high-resolution remote sensing images, refAbstract=null), Reference(id=1322962993683771844, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2024, volume=62, issue=null, pageStart=null, pageEnd=4503415, url=null, language=null, rfNumber=null, rfOrder=10, authorNames=Luo H, Wang Z J, Du B, Dong Y N, journalName=IEEE Transactions on Geoscience and Remote Sensing, refType=null, unstructuredReference=Luo H,Wang Z J,Du B and Dong Y N. 2024. A deep cross-modal fusion network for road extraction with high-resolution imagery and LiDAR data. IEEE Transactions on Geoscience and Remote Sensing,62: #4503415 [DOI: 10.1109/TGRS.2024.3360963], articleTitle=A deep cross-modal fusion network for road extraction with high-resolution imagery and LiDAR data, refAbstract=null), Reference(id=1322962993759269317, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2013, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=null, rfOrder=11, authorNames=Mnih V, journalName=Machine Learning for Aerial Image Labeling, refType=null, unstructuredReference=Mnih V. 2013. Machine Learning for Aerial Image Labeling. Toronto:University of Toronto:[s.n.], articleTitle=null, refAbstract=null), Reference(id=1322962993838961094, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2010, volume=null, issue=null, pageStart=210, pageEnd=223, url=null, language=null, rfNumber=null, rfOrder=12, authorNames=Mnih V, Hinton G E, journalName=null, refType=null, unstructuredReference=Mnih V and Hinton G E. 2010. Learning to detect roads in high-resolution aerial images//Proceedings of the 11th European Conference on Computer Vision. Heraklion,Greece: Springer: 210-223 [DOI: 10.1007/978-3-642-15561-1_16], articleTitle=Learning to detect roads in high-resolution aerial images, refAbstract=null), Reference(id=1322962993901875655, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=null, pageEnd=null, url=https://arxiv.org/pdf/1807.01232.pdf, language=null, rfNumber=null, rfOrder=13, authorNames=Van Etten A, Lindenbaum D, Bacastow T M, journalName=null, refType=null, unstructuredReference=Van Etten A,Lindenbaum D and Bacastow T M. 2018. SpaceNet: a remote sensing dataset and challenge series [EB/OL]. [2025-10-10]., articleTitle=SpaceNet: a remote sensing dataset and challenge series, refAbstract=null), Reference(id=1322962993973178824, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=27, issue=10, pageStart=3102, pageEnd=3115, url=null, language=null, rfNumber=null, rfOrder=14, authorNames=Wang Y, Zeng X Q, journalName=Journal of Image and Graphics, refType=null, unstructuredReference=Wang Y and Zeng X Q. 2022. Road extraction model derived from integrated attention mechanism and dilated convolution. Journal of Image and Graphics,27(10): 3102-3115, articleTitle=Road extraction model derived from integrated attention mechanism and dilated convolution, refAbstract=null), Reference(id=1322962994048676297, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=27, issue=10, pageStart=3102, pageEnd=3115, url=null, language=null, rfNumber=null, rfOrder=15, authorNames=王勇, 曾祥强, journalName=中国图象图形学报, refType=null, unstructuredReference=王勇,曾祥强. 2022. 集成注意力机制和扩张卷积的道路提取模型. 中国图象图形学报,27(10): 3102-3115 [DOI: 10.11834/jig.210226], articleTitle=集成注意力机制和扩张卷积的道路提取模型, refAbstract=null), Reference(id=1322962994136756682, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=26, issue=9, pageStart=1872, pageEnd=1885, url=null, language=null, rfNumber=null, rfOrder=16, authorNames=Wu Q Q, Wang S, Wang B, Wu Y L, journalName=National Remote Sensing Bulletin, refType=null, unstructuredReference=Wu Q Q,Wang S,Wang B and Wu Y L. 2022. Road extraction method of high-resolution remote sensing image on the basis of the spatial information perception semantic segmentation model. National Remote Sensing Bulletin,26(9): 1872-1885, articleTitle=Road extraction method of high-resolution remote sensing image on the basis of the spatial information perception semantic segmentation model, refAbstract=null), Reference(id=1322962994203865547, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=26, issue=9, pageStart=1872, pageEnd=1885, url=null, language=null, rfNumber=null, rfOrder=17, authorNames=吴强强, 王帅, 王彪, 吴艳兰, journalName=遥感学报, refType=null, unstructuredReference=吴强强,王帅,王彪,吴艳兰. 2022. 空间信息感知语义分割模型的高分辨率遥感影像道路提取. 遥感学报,26(9): 1872-1885 [DOI: 10.11834/jrs.20210021], articleTitle=空间信息感知语义分割模型的高分辨率遥感影像道路提取, refAbstract=null), Reference(id=1322962994283557324, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2023, volume=43, issue=5, pageStart=273, pageEnd=276, url=null, language=null, rfNumber=null, rfOrder=18, authorNames=Yang C F, Yu S Y, journalName=Journal of China and Foreign Highway, refType=null, unstructuredReference=Yang C F and Yu S Y. 2023. Research on route scheme of highway in difficult and dangerous mountainous area based on construction and rescue factors. Journal of China and Foreign Highway,43(5): 273-276, articleTitle=Research on route scheme of highway in difficult and dangerous mountainous area based on construction and rescue factors, refAbstract=null), Reference(id=1322962994354860493, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2023, volume=43, issue=5, pageStart=273, pageEnd=276, url=null, language=null, rfNumber=null, rfOrder=19, authorNames=杨昌凤, 余蜀予, journalName=中外公路, refType=null, unstructuredReference=杨昌凤,余蜀予. 2023. 基于施工及救援因素的艰险山区公路路线方案研究. 中外公路,43(5): 273-276 [DOI: 10.14048/j.issn.1671-2579.2023.05.046], articleTitle=基于施工及救援因素的艰险山区公路路线方案研究, refAbstract=null), Reference(id=1322962994417775054, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2021, volume=26, issue=12, pageStart=3005, pageEnd=3014, url=null, language=null, rfNumber=null, rfOrder=20, authorNames=Yang J L, Guo X J, Chen Z H, journalName=Journal of Image and Graphics, refType=null, unstructuredReference=Yang J L,Guo X J and Chen Z H. 2021. Road extraction method from remote sensing images based on improved U-Net network. Journal of Image and Graphics,26(12): 3005-3014, articleTitle=Road extraction method from remote sensing images based on improved U-Net network, refAbstract=null), Reference(id=1322962994489078223, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2021, volume=26, issue=12, pageStart=3005, pageEnd=3014, url=null, language=null, rfNumber=null, rfOrder=21, authorNames=杨佳林, 郭学俊, 陈泽华, journalName=中国图象图形学报, refType=null, unstructuredReference=杨佳林,郭学俊,陈泽华. 2021. 改进U-Net型网络的遥感图像道路提取. 中国图象图形学报,26(12): 3005-3014 [DOI: 10.11834/jig.202863], articleTitle=改进U-Net型网络的遥感图像道路提取, refAbstract=null), Reference(id=1322962994564575696, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2024a, volume=62, issue=null, pageStart=null, pageEnd=5619316, url=null, language=null, rfNumber=null, rfOrder=22, authorNames=Yang R Y, Zhong Y F, Liu Y H, Lu X Y, Zhang L P, journalName=IEEE Transactions on Geoscience and Remote Sensing, refType=null, unstructuredReference=Yang R Y,Zhong Y F,Liu Y H,Lu X Y and Zhang L P. 2024a. Occlusion-aware road extraction network for high-resolution remote sensing imagery. IEEE Transactions on Geoscience and Remote Sensing,62: #5619316 [DOI: 10.1109/TGRS.2024.3387945], articleTitle=Occlusion-aware road extraction network for high-resolution remote sensing imagery, refAbstract=null), Reference(id=1322962994631684561, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2024b, volume=62, issue=null, pageStart=null, pageEnd=5647011, url=null, language=null, rfNumber=null, rfOrder=23, authorNames=Yang Z G, Zhang W, Li Q, Ni W P, Wu J Z, Wang Q, journalName=IEEE Transactions on Geoscience and Remote Sensing, refType=null, unstructuredReference=Yang Z G,Zhang W,Li Q,Ni W P,Wu J Z and Wang Q. 2024b. C2Net: road extraction via context perception and cross spatial-scale feature interaction. IEEE Transactions on Geoscience and Remote Sensing,62: #5647011 [DOI: 10.1109/TGRS.2024.3491755], articleTitle=C2Net: road extraction via context perception and cross spatial-scale feature interaction, refAbstract=null), Reference(id=1322962994698793426, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=20, issue=null, pageStart=null, pageEnd=8000405, url=null, language=null, rfNumber=null, rfOrder=24, authorNames=Yang Z G, Zhou D X, Yang Y, Zhang J P, Chen Z H, journalName=IEEE Geoscience and Remote Sensing Letters, refType=null, unstructuredReference=Yang Z G,Zhou D X,Yang Y,Zhang J P and Chen Z H. 2022. Road extraction from satellite imagery by road context and full-stage feature. IEEE Geoscience and Remote Sensing Letters,20: #8000405 [DOI: 10.1109/LGRS.2022.3228967], articleTitle=Road extraction from satellite imagery by road context and full-stage feature, refAbstract=null), Reference(id=1322962994782679507, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2024, volume=62, issue=null, pageStart=null, pageEnd=5633314, url=null, language=null, rfNumber=null, rfOrder=25, authorNames=Zhao S J, Chen H, Zhang X L, Xiao P F, Bai L, Ouyang W L, journalName=IEEE Transactions on Geoscience and Remote Sensing, refType=null, unstructuredReference=Zhao S J,Chen H,Zhang X L,Xiao P F,Bai L and Ouyang W L. 2024. RS-Mamba for large remote sensing image dense prediction. IEEE Transactions on Geoscience and Remote Sensing,62: #5633314 [DOI: 10.1109/TGRS.2024.3425540], articleTitle=RS-Mamba for large remote sensing image dense prediction, refAbstract=null), Reference(id=1322962994870759892, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=60, issue=null, pageStart=null, pageEnd=5614115, url=null, language=null, rfNumber=null, rfOrder=26, authorNames=Zhou G D, Chen W T, Gui Q S, Li X J, Wang L Z, journalName=IEEE Transactions on Geoscience and Remote Sensing, refType=null, unstructuredReference=Zhou G D,Chen W T,Gui Q S,Li X J and Wang L Z. 2022. Split depth-wise separable graph-convolution network for road extraction in complex environments from high-resolution remote-sensing images. IEEE Transactions on Geoscience and Remote Sensing,60: #5614115 [DOI: 10.1109/TGRS.2021.3128033], articleTitle=Split depth-wise separable graph-convolution network for road extraction in complex environments from high-resolution remote-sensing images, refAbstract=null), Reference(id=1322962994963034581, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=192, pageEnd=1924, url=null, language=null, rfNumber=null, rfOrder=27, authorNames=Zhou L C, Zhang C, Wu M, journalName=null, refType=null, unstructuredReference=Zhou L C,Zhang C and Wu M. 2018. D-LinkNet: LinkNet with pretrained encoder and dilated convolution for high resolution satellite imagery road extraction//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Salt Lake City,USA: IEEE: 192-1924 [DOI: 10.1109/CVPRW.2018.00034], articleTitle=D-LinkNet: LinkNet with pretrained encoder and dilated convolution for high resolution satellite imagery road extraction, refAbstract=null), Reference(id=1322962995055309270, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2020, volume=168, issue=null, pageStart=288, pageEnd=306, url=null, language=null, rfNumber=null, rfOrder=28, authorNames=Zhou M T, Sui H G, Chen S X, Wang J D, Chen X, journalName=ISPRS Journal of Photogrammetry and Remote Sensing, refType=null, unstructuredReference=Zhou M T,Sui H G,Chen S X,Wang J D and Chen X. 2020. BT-RoadNet: a boundary and topologically-aware neural network for road extraction from high-resolution remote sensing imagery. ISPRS Journal of Photogrammetry and Remote Sensing,168: 288-306 [DOI: 10.1016/j.isprsjprs.2020.08.019], articleTitle=BT-RoadNet: a boundary and topologically-aware neural network for road extraction from high-resolution remote sensing imagery, refAbstract=null), Reference(id=1322962995130806743, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2022, volume=184, issue=null, pageStart=63, pageEnd=78, url=null, language=null, rfNumber=null, rfOrder=29, authorNames=Zhu Q Q, Guo X, Deng W H, Shi S N, Guan Q F, Zhong Y F, journalName=ISPRS Journal of Photogrammetry and Remote Sensing, refType=null, unstructuredReference=Zhu Q Q,Guo X,Deng W H,Shi S N,Guan Q F,Zhong Y F,et al. 2022. Land-use/land-cover change detection based on a Siamese global learning framework for high spatial resolution remote sensing imagery. ISPRS Journal of Photogrammetry and Remote Sensing,184: 63-78 [DOI: 10.1016/j.isprsjprs.2021.12.005], articleTitle=Land-use/land-cover change detection based on a Siamese global learning framework for high spatial resolution remote sensing imagery, refAbstract=null), Reference(id=1322962995202109912, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, doi=null, pmid=null, pmcid=null, year=2021, volume=175, issue=null, pageStart=353, pageEnd=365, url=null, language=null, rfNumber=null, rfOrder=30, authorNames=Zhu Q Q, Zhang Y N, Wang L Z, Zhong Y F, Guan Q F, Lu X Y, journalName=ISPRS Journal of Photogrammetry and Remote Sensing, refType=null, unstructuredReference=Zhu Q Q,Zhang Y N,Wang L Z,Zhong Y F,Guan Q F,Lu X Y,et al. 2021. A global context-aware and batch-independent network for road extraction from VHR satellite imagery. ISPRS Journal of Photogrammetry and Remote Sensing,175: 353-365 [DOI: 10.1016/j.isprsjprs.2021.03.016], articleTitle=A global context-aware and batch-independent network for road extraction from VHR satellite imagery, refAbstract=null)], funds=[Fund(id=1322962991142023608, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, awardId=null, language=EN, fundingSource=Pre-Research Project of Civil Aerospace Technology under the 14th Five-Year Plan (D040404), fundOrder=null, country=null), Fund(id=1322962991221715385, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, awardId=null, language=CN, fundingSource=“十四五”民用航天技术预先研究项目(D040404), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)]), AuthorCompany(id=1322962984963813744, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=2, ext=[AuthorCompanyExt(id=1322962984972202353, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China), AuthorCompanyExt(id=1322962984980590962, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国科学院大学电子电气与通信工程学院,北京 100049)])], figs=[ArticleFig(id=1322962988814184862, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Fig.1, caption=Data annotation flowchart, figureFileSmall=d+lFqTUfWzbFSxIW692TwA==, figureFileBig=oMcE+c7DldycO97WGlDtGg==, tableContent=null), ArticleFig(id=1322962988885488031, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=图1, caption=数据标注流程, figureFileSmall=d+lFqTUfWzbFSxIW692TwA==, figureFileBig=oMcE+c7DldycO97WGlDtGg==, tableContent=null), ArticleFig(id=1322962989074231712, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Fig.2, caption=Examples of images and their annotations in the AsiaMountain-Road dataset, figureFileSmall=BPQIsaxrHF4Ddb+qTNeunA==, figureFileBig=fTAeZMUerE5Fc7+rv7DzIw==, tableContent=null), ArticleFig(id=1322962989166506401, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=图2, caption=AsiaMountain-Road数据集部分影像及其对应的标注示例, figureFileSmall=BPQIsaxrHF4Ddb+qTNeunA==, figureFileBig=fTAeZMUerE5Fc7+rv7DzIw==, tableContent=null), ArticleFig(id=1322962989330084258, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Fig.3, caption=Examples of challenges faced by the AsiaMountain-Road dataset ((a) image inconsistency caused by pronounced terrain relief; (b) roads exhibit complex morphology and poor structural connectivity; (c) the extreme sparsity of road samples leads to severe class imbalance; (d) roads exhibit high spectral similarity with natural backgrounds), figureFileSmall=Ba/NLhQXx42u0xfhiW0G/A==, figureFileBig=Fk2nuJ0hMyAni5IWDwMjOw==, tableContent=null), ArticleFig(id=1322962989409776035, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=图3, caption=AsiaMountain-Road数据集所面临的挑战示例, figureFileSmall=Ba/NLhQXx42u0xfhiW0G/A==, figureFileBig=Fk2nuJ0hMyAni5IWDwMjOw==, tableContent=null), ArticleFig(id=1322962989481079204, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Fig.4, caption=The network architecture of MR-AFDANet, figureFileSmall=VsKQg/Y0XCRDalcEQVIGLQ==, figureFileBig=PpDJm1d3cjkCTssbO1Aijg==, tableContent=null), ArticleFig(id=1322962989569159589, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=图4, caption=MR-AFDANet网络架构, figureFileSmall=VsKQg/Y0XCRDalcEQVIGLQ==, figureFileBig=PpDJm1d3cjkCTssbO1Aijg==, tableContent=null), ArticleFig(id=1322962989640462758, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Fig.5, caption=Inference results obtained on the AsiaMountain-Road test set for models trained on the DeepGlobe dataset ((a) true images;(b) ground truths(GT);(c) D-LinkNet;(d) OareNet;(e) RCFSNet;(f) C2Net;(g) AFDANet;(h) RS-Mamba;(i) ours), figureFileSmall=4T8kktlzH8GjsEMGOMVxxA==, figureFileBig=tgwKbLE/d3zoMCWgOroU0Q==, tableContent=null), ArticleFig(id=1322962989732737447, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=图5, caption=在DeepGlobe数据集上训练的模型在AsiaMountain-Road测试集上获得的推理结果, figureFileSmall=4T8kktlzH8GjsEMGOMVxxA==, figureFileBig=tgwKbLE/d3zoMCWgOroU0Q==, tableContent=null), ArticleFig(id=1322962989795652008, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Fig.6, caption=Qualitative results of different road extraction methods on the AsiaMountain-Road dataset, figureFileSmall=TulcbgrhC//fv+9b5Q97uQ==, figureFileBig=7iJoiS+wz/gsiP8koE4B2g==, tableContent=null), ArticleFig(id=1322962989871149481, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=图6, caption=不同道路提取方法在AsiaMountain-Road数据集的定性结果

((a) true images;(b) GT;(c) D-LinkNet;(d) OareNet;(e) RCFSNet;(f) C2Net;(g) AFDANet;(h) RS-Mamba;(i) ours)

, figureFileSmall=TulcbgrhC//fv+9b5Q97uQ==, figureFileBig=7iJoiS+wz/gsiP8koE4B2g==, tableContent=null), ArticleFig(id=1322962989946646954, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Fig.7, caption=Generalization results of different road extraction methods on the Massachusetts Roads dataset, figureFileSmall=XnpC72P+QCOrKqwVsJR9ag==, figureFileBig=an+vHIhHGGYPZbo+LCUKVA==, tableContent=null), ArticleFig(id=1322962990034727339, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=图7, caption=不同道路提取方法在Massachusetts Roads数据集上的泛化性结果

((a) true images;(b) GT;(c) D-LinkNet;(d) OareNet;(e) RCFSNet;(f) C2Net;(g) AFDANet;(h) RS-Mamba;(i) ours)

, figureFileSmall=XnpC72P+QCOrKqwVsJR9ag==, figureFileBig=an+vHIhHGGYPZbo+LCUKVA==, tableContent=null), ArticleFig(id=1322962990093447596, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Table 1, caption=The existing high-resolution remote sensing public road extraction datasets, figureFileSmall=null, figureFileBig=null, tableContent=
数据集影像尺寸/像素空间分辨率/m训练/验证影像测试影像覆盖面积/km2
DeepGlobe(Demir等,2018)1 024 × 1 0240.54 6961 5301 632.11
SpaceNet(Van Etten等,2018)1 300 × 1 3000.32 2135 670422.84
Massachusetts Roads(Mnih,2013)1 500 × 1 5001.01 108/14492 600
CHN6-CUG(Zhu等,2022)512 × 5120.53 608903295.63
GanSu-Road(Zhou等,2022)256 × 2562.0204/11402 500
WHU(Zhou等,2020)512 × 5120.8 × 2.05 1211 7071 145
AsiaMountain-Road(本文)

1 024 × 1 024

2 048 × 2 048

1.0

7 183/1 100

1 213/160

1 936

333

1 836
), ArticleFig(id=1322962990173139373, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=表1, caption=

现有的高分辨率遥感公共道路提取数据集

, figureFileSmall=null, figureFileBig=null, tableContent=
数据集影像尺寸/像素空间分辨率/m训练/验证影像测试影像覆盖面积/km2
DeepGlobe(Demir等,2018)1 024 × 1 0240.54 6961 5301 632.11
SpaceNet(Van Etten等,2018)1 300 × 1 3000.32 2135 670422.84
Massachusetts Roads(Mnih,2013)1 500 × 1 5001.01 108/14492 600
CHN6-CUG(Zhu等,2022)512 × 5120.53 608903295.63
GanSu-Road(Zhou等,2022)256 × 2562.0204/11402 500
WHU(Zhou等,2020)512 × 5120.8 × 2.05 1211 7071 145
AsiaMountain-Road(本文)

1 024 × 1 024

2 048 × 2 048

1.0

7 183/1 100

1 213/160

1 936

333

1 836
), ArticleFig(id=1322962990257025454, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Table 2, caption=Distribution of mountain types in the AsiaMountain-Road dataset, figureFileSmall=null, figureFileBig=null, tableContent=
山区类型数量占比/%主要分布区域
高海拔雪山—冰川型413.79中国青藏高原及周边高山区(喜马拉雅、喀喇昆仑、帕米尔、昆仑)
峡谷—河谷型517.24中国(横断山区、川陕鄂渝山地)、阿富汗和巴基斯坦(兴都库什)、俄罗斯(西高加索)
丘陵—山地农业型620.69中国东南丘陵、中国台湾、中国山西、菲律宾、哈萨克丘陵
草原—山前盆地型310.35中国西北及中亚山前地带(天山、阿尔泰、祁连)
火山—岩层台地型26.90日本列岛
热带—高原聚落型413.79东南亚热带高原及岛屿、阿拉伯半岛南部
山地城镇型517.24伊朗高原、印度半岛东西两侧山地
), ArticleFig(id=1322962990345105839, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=表2, caption=

AsiaMountain-Road山区类型分布

, figureFileSmall=null, figureFileBig=null, tableContent=
山区类型数量占比/%主要分布区域
高海拔雪山—冰川型413.79中国青藏高原及周边高山区(喜马拉雅、喀喇昆仑、帕米尔、昆仑)
峡谷—河谷型517.24中国(横断山区、川陕鄂渝山地)、阿富汗和巴基斯坦(兴都库什)、俄罗斯(西高加索)
丘陵—山地农业型620.69中国东南丘陵、中国台湾、中国山西、菲律宾、哈萨克丘陵
草原—山前盆地型310.35中国西北及中亚山前地带(天山、阿尔泰、祁连)
火山—岩层台地型26.90日本列岛
热带—高原聚落型413.79东南亚热带高原及岛屿、阿拉伯半岛南部
山地城镇型517.24伊朗高原、印度半岛东西两侧山地
), ArticleFig(id=1322962990424797616, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Table 3, caption=Typical mountain types and core features of the AsiaMountain-Road dataset, figureFileSmall=null, figureFileBig=null, tableContent=
山区类别核心地形特征道路分布与形态特征主导道路类型典型代表山脉
高海拔雪山—冰川型海拔极高(多 > 4 000 m),冰川/荒漠广布,植被覆盖率 < 5%道路密度极低(每100 km² < 1 km),仅沿山口、冰川边缘或简易盘山路径延伸,无分支盘山公路、乡村山路喜马拉雅山、喀喇昆仑山、帕米尔高原、昆仑山
峡谷—河谷型峡谷深切(相对高差 > 1 000 m),聚落沿河谷狭长分布1)河谷主线型:道路沿河谷线性展开,方向受河道限制,顺流向延伸,主干道为主;2)盘山道路型:绕山盘旋、弯曲层叠,多回头弯,衔接河谷与山顶区域河谷主线道路、水坝施工道、盘山公路、乡村山路横断山脉、巫山、兴都库什、西高加索山脉、大巴山
丘陵—山地农业型低山丘陵(相对高差200~800 m),茶园/梯田连片,耕地穿插1)河谷依托型:以河谷主干道为核心,向两侧梯田区延伸支线(如武夷山);2)盘山细密型:乡村道路密集,多窄幅小路,随梯田等高线弯曲,分支发达(如吕宋山地)乡村山路、林道、村际联络线武夷山、吕宋山地、天目山、恒山、哈萨克丘陵
草原—山前盆地型山前平缓盆地,草原/荒漠草原过渡带,聚落分散主干道通达性强(连接盆地核心聚落),次级道路稀疏,呈放射状或“十字形”交汇,无密集分支乡村山路、盘山公路、矿区道路天山、阿尔泰山、祁连山
火山—岩层台地型火山地貌 + 岩层台地,针阔混交林覆盖,地形起伏和缓林道与盘山路结合,主干道绕台地边缘,支线深入林区,整体呈“网格状”分支(覆盖林区与台地)林道、盘山公路、乡村山路日本阿尔卑斯山、日高山脉
热带—高原聚落型热带高原(海拔1 000~2 000 m),高温多雨,村落密集(每10 km² >5个)道路网络化程度高,主干道连接村落集群,支线随聚落分布延伸,多与耕地、林地交错乡村山路、村际联络线、林道巴新高地、阿拉伯山地、若开山、克罗克山
山地城镇型山地腹地或山谷间小型城镇聚集,地形以低山(相对高差500~1 000 m)为主,周边有少量耕地/林地城镇内部道路呈网格状,外部道路以连接周边村落或主干道为主;部分路段沿山脚或缓坡修建,弯曲度适中,无复杂盘山盘山公路、乡村山路、矿区道路阿尔波兹山、伊朗中部山地、札格罗斯山、西高止山脉、东高止山脉
), ArticleFig(id=1322962990500295089, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=表3, caption=

AsiaMountain-Road数据集典型山区类型及核心特征

, figureFileSmall=null, figureFileBig=null, tableContent=
山区类别核心地形特征道路分布与形态特征主导道路类型典型代表山脉
高海拔雪山—冰川型海拔极高(多 > 4 000 m),冰川/荒漠广布,植被覆盖率 < 5%道路密度极低(每100 km² < 1 km),仅沿山口、冰川边缘或简易盘山路径延伸,无分支盘山公路、乡村山路喜马拉雅山、喀喇昆仑山、帕米尔高原、昆仑山
峡谷—河谷型峡谷深切(相对高差 > 1 000 m),聚落沿河谷狭长分布1)河谷主线型:道路沿河谷线性展开,方向受河道限制,顺流向延伸,主干道为主;2)盘山道路型:绕山盘旋、弯曲层叠,多回头弯,衔接河谷与山顶区域河谷主线道路、水坝施工道、盘山公路、乡村山路横断山脉、巫山、兴都库什、西高加索山脉、大巴山
丘陵—山地农业型低山丘陵(相对高差200~800 m),茶园/梯田连片,耕地穿插1)河谷依托型:以河谷主干道为核心,向两侧梯田区延伸支线(如武夷山);2)盘山细密型:乡村道路密集,多窄幅小路,随梯田等高线弯曲,分支发达(如吕宋山地)乡村山路、林道、村际联络线武夷山、吕宋山地、天目山、恒山、哈萨克丘陵
草原—山前盆地型山前平缓盆地,草原/荒漠草原过渡带,聚落分散主干道通达性强(连接盆地核心聚落),次级道路稀疏,呈放射状或“十字形”交汇,无密集分支乡村山路、盘山公路、矿区道路天山、阿尔泰山、祁连山
火山—岩层台地型火山地貌 + 岩层台地,针阔混交林覆盖,地形起伏和缓林道与盘山路结合,主干道绕台地边缘,支线深入林区,整体呈“网格状”分支(覆盖林区与台地)林道、盘山公路、乡村山路日本阿尔卑斯山、日高山脉
热带—高原聚落型热带高原(海拔1 000~2 000 m),高温多雨,村落密集(每10 km² >5个)道路网络化程度高,主干道连接村落集群,支线随聚落分布延伸,多与耕地、林地交错乡村山路、村际联络线、林道巴新高地、阿拉伯山地、若开山、克罗克山
山地城镇型山地腹地或山谷间小型城镇聚集,地形以低山(相对高差500~1 000 m)为主,周边有少量耕地/林地城镇内部道路呈网格状,外部道路以连接周边村落或主干道为主;部分路段沿山脚或缓坡修建,弯曲度适中,无复杂盘山盘山公路、乡村山路、矿区道路阿尔波兹山、伊朗中部山地、札格罗斯山、西高止山脉、东高止山脉
), ArticleFig(id=1322962990575792562, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Table 4, caption=Quantitative results of different road extraction methods on the AsiaMountain-Road dataset, figureFileSmall=null, figureFileBig=null, tableContent=
方法精确率召回率F1分数交并比
D-LinkNet78.2369.2773.4858.34
OareNet78.6367.8672.8557.30
RCFSNet77.2260.0667.5751.02
C2Net78.6868.9073.4758.33
AFDANet83.0261.8870.9054.93
RS-Mamba80.2867.1873.1557.67
本文80.1071.4574.5359.40
), ArticleFig(id=1322962990659678643, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=表4, caption=

不同道路提取方法在AsiaMountain-Road数据集的定量结果

, figureFileSmall=null, figureFileBig=null, tableContent=
方法精确率召回率F1分数交并比
D-LinkNet78.2369.2773.4858.34
OareNet78.6367.8672.8557.30
RCFSNet77.2260.0667.5751.02
C2Net78.6868.9073.4758.33
AFDANet83.0261.8870.9054.93
RS-Mamba80.2867.1873.1557.67
本文80.1071.4574.5359.40
), ArticleFig(id=1322962990760341940, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Table 5, caption=Ablation study of MR-AFDANet on the AsiaMountain-Road dataset, figureFileSmall=null, figureFileBig=null, tableContent=
实验编号网络结构SEAF模块高低分辨率融合模块SRA解码器精确率召回率F1分数交并比
1Baseline---73.9758.8365.5448.74
2+ SEAF模块√--77.8866.3571.7053.05
3+ 高低分辨率融合模块-√-76.4270.2373.1955.41
4+ SRA解码器--√77.6368.8773.0255.18
5+ SEAF模块 + 高低分辨率融合模块 + SRA解码器√√√80.1071.4574.5359.40
), ArticleFig(id=1322962990835839413, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=表5, caption=

MR-AFDANet模块消融实验结果

, figureFileSmall=null, figureFileBig=null, tableContent=
实验编号网络结构SEAF模块高低分辨率融合模块SRA解码器精确率召回率F1分数交并比
1Baseline---73.9758.8365.5448.74
2+ SEAF模块√--77.8866.3571.7053.05
3+ 高低分辨率融合模块-√-76.4270.2373.1955.41
4+ SRA解码器--√77.6368.8773.0255.18
5+ SEAF模块 + 高低分辨率融合模块 + SRA解码器√√√80.1071.4574.5359.40
), ArticleFig(id=1322962990898753974, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=EN, label=Table 6, caption=Quantitative results of different road extraction methods on the Massachusetts-Roads dataset, figureFileSmall=null, figureFileBig=null, tableContent=
方法精确率召回率F1分数交并比
D-LinkNet63.8780.9271.3955.51
OareNet64.5855.1859.5142.36
RCFSNet60.2653.1256.4639.34
C2Net68.0848.8456.8839.74
AFDANet63.9662.1263.0346.01
RS-Mamba69.7065.9367.7651.24
本文76.4075.0575.7260.93
), ArticleFig(id=1322962990957474231, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, language=CN, label=表6, caption=

不同道路提取方法在Massachusetts-Roads数据集上的泛化性定量结果

, figureFileSmall=null, figureFileBig=null, tableContent=
方法精确率召回率F1分数交并比
D-LinkNet63.8780.9271.3955.51
OareNet64.5855.1859.5142.36
RCFSNet60.2653.1256.4639.34
C2Net68.0848.8456.8839.74
AFDANet63.9662.1263.0346.01
RS-Mamba69.7065.9367.7651.24
本文76.4075.0575.7260.93
)], attaches=null, journal=Journal(id=1249023527618129992, delFlag=0, nameCn=中国图象图形学报, nameEn=Journal of Image and Graphics, nameHistory1=null, nameHistory2=null, issn=1006-8961, eissn=null, cn=11-3758, coden=CODEN ZTTXFZ, periodic=0, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=uirXtX858YS3zEpFXZttJA==, journalPrice=null, startedYear=null, abbrevIsoEn=Journal of Image and Graphics, journalRemark=null, publicationField=null, createdTime=1775720014721, updatedTime=1788951488513, createdBy=18614031015, updatedBy=13041195026, firstLetterCn=Z, firstLetterEn=Z, subjectCode=Electronic Information, subjectName=null, subjectCodeEn=Electronic Information, subjectNameEn=null, picCn=uirXtX858YS3zEpFXZttJA==, picEn=bud7qaxfvWHeFsbyBTAiKQ==, jcr=null, cjcr=null, exts=[JournalExt(id=1304520351199814396, language=CN, name=中国图象图形学报, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1788951488542, updatedTime=1788951488542, createdBy=13041195026, updatedBy=13041195026, submissionGuidelinesUrl=, submissionAuthorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-cjig-author&redirect_uri=https%3A%2F%2Fcjig.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=e6369def-2842-41d8, submissionEditorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-cjig-editor&redirect_uri=https%3A%2F%2Fcjigeditor.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=9ccec05b-6bd, submissionReviewUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-cjig-author&redirect_uri=https%3A%2F%2Fcjig.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=1e8a31c8-5434-4f78, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1304520351250146045, language=EN, name=Journal of Image and Graphics, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1788951488554, updatedTime=1788951488554, createdBy=13041195026, updatedBy=13041195026, submissionGuidelinesUrl=, submissionAuthorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-cjig-author&redirect_uri=https%3A%2F%2Fcjig.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=e6369def-2842-41d8, submissionEditorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-cjig-editor&redirect_uri=https%3A%2F%2Fcjigeditor.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=9ccec05b-6bd, submissionReviewUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-cjig-author&redirect_uri=https%3A%2F%2Fcjig.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=1e8a31c8-5434-4f78, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1249024232475115590, websiteList=[Website(id=1249025782459334881, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1249024232475115590, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/zgtxtxxb/CN, language=CN, createTime=1775720552315, createBy=18614031015, updateTime=1775720586268, updateBy=18614031015, name=中国图象图形学报-中文, tplId=1146099689490845704, title=中国图象图形学报, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1249026166254928133, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=articleTextType, value=kx, createTime=1775720643819, updateTime=1775720643819, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166221373698, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=banner, value=null, createTime=1775720643811, updateTime=1775720643811, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166271705352, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=grayFlag, value=0, createTime=1775720643823, updateTime=1775720643823, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166212985089, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=logo, value=https://castjournals.cast.org.cn/joweb/zgtxtxxb/CN/file/pic?fileId=TDRjKTHfgAnvFKZaDA70wA==, createTime=1775720643809, updateTime=1775720643809, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166288482570, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=minRunFlag, value=0, createTime=1775720643827, updateTime=1775720643827, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166246539524, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/zgtxtxxb/CN/file/pic, createTime=1775720643817, updateTime=1775720643817, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166280093961, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=silenceFlag, value=0, createTime=1775720643825, updateTime=1775720643825, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166233956611, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1775720643814, updateTime=1775720643814, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166259122438, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=themeColor, value=null, createTime=1775720643820, updateTime=1775720643820, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026166267511047, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782459334881, code=themeStyle, value=null, createTime=1775720643822, updateTime=1775720643822, creator=18614031015, updator=18614031015)]), Website(id=1249025782681633001, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1249024232475115590, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/zgtxtxxb/EN, language=EN, createTime=1775720552368, createBy=18614031015, updateTime=1775720607118, updateBy=18614031015, name=中国图象图形学报-英文, tplId=1146101810881728533, title=Journal of Image and Graphics, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1249026195371786511, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=articleTextType, value=kx, createTime=1775720650761, updateTime=1775720650761, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195355009292, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=banner, value=null, createTime=1775720650757, updateTime=1775720650757, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195392758034, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=grayFlag, value=0, createTime=1775720650766, updateTime=1775720650766, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195342426379, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=logo, value=https://castjournals.cast.org.cn/joweb/zgtxtxxb/EN/file/pic?fileId=TDRjKTHfgAnvFKZaDA70wA==, createTime=1775720650754, updateTime=1775720650754, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195409535252, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=minRunFlag, value=0, createTime=1775720650770, updateTime=1775720650770, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195367592206, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/zgtxtxxb/EN/file/pic, createTime=1775720650760, updateTime=1775720650760, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195401146643, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=silenceFlag, value=0, createTime=1775720650768, updateTime=1775720650768, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195359203597, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1775720650758, updateTime=1775720650758, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195380175120, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=themeColor, value=null, createTime=1775720650763, updateTime=1775720650763, creator=18614031015, updator=18614031015), WebsiteProps(id=1249026195388563729, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1249025782681633001, code=themeStyle, value=null, createTime=1775720650765, updateTime=1775720650765, creator=18614031015, updator=18614031015)])], journalTitle=中国图象图形学报, weixinUrl=null, journalUrl=https://www.cjig.cn/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Journal of Image and Graphics, journalPhotoCn=uirXtX858YS3zEpFXZttJA==, journalPhotoEn=bud7qaxfvWHeFsbyBTAiKQ==, journalFirstLetter=Z, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false, interPubPlatform=, interPubPlatformUrl=null), detailUrlCn=https://castjournals.cast.org.cn/joweb/zgtxtxxb/CN/10.11834/jig.250521, detailUrlEn=https://castjournals.cast.org.cn/joweb/zgtxtxxb/EN/10.11834/jig.250521, pdfUrlCn=https://castjournals.cast.org.cn/joweb/zgtxtxxb/CN/PDF/10.11834/jig.250521, pdfUrlEn=https://castjournals.cast.org.cn/joweb/zgtxtxxb/EN/PDF/10.11834/jig.250521, aliStartDate=0, aliEndDate=0, collectionFlag=false, citedCount=null, citedUrl=null, previewStatus=0, delFlag=0, hasFullText=1, orderTime=1792080000000, fullTextJson=null, articleText=null, reference=null)
收藏切换
高分辨率遥感影像山区道路提取方法与基准数据集:AsiaMountain-Road
收藏切换
PDF下载
王怡琳 1, 2 , 孟瑜 1, 2, * , 席智浩 1 , 邓毓弸 1 , 刘帝佑 1 , 陈静波 1, 2
中国图象图形学报 | 数据集 2026,31(10): 3590-3605
收起
收藏切换
中国图象图形学报 |数据集 2026 , 31 (10) : 3590 -3605
高分辨率遥感影像山区道路提取方法与基准数据集:AsiaMountain-Road
全屏
[Author(id=1322962985056088436, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=wangyilin23@mails.ucas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962985135780215, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985056088436, language=EN, stringName=Yilin Wang, firstName=Yilin, middleName=null, lastName=Wang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985253220728, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985056088436, language=CN, stringName=王怡琳, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 中国科学院空天信息创新研究院,北京 100094
2 中国科学院大学电子电气与通信工程学院,北京 100049, bio={"content":"

王怡琳,女,硕士研究生,主要研究方向为遥感图像语义分割。E-mail:

"}, bioImg=null, bioContent=

王怡琳,女,硕士研究生,主要研究方向为遥感图像语义分割。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)]), AuthorCompany(id=1322962984963813744, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=2, ext=[AuthorCompanyExt(id=1322962984972202353, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China), AuthorCompanyExt(id=1322962984980590962, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国科学院大学电子电气与通信工程学院,北京 100049)])]), Author(id=1322962985328718202, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=mengyu@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1322962985404215677, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985328718202, language=EN, stringName=Yu Meng, firstName=Yu, middleName=null, lastName=Meng, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, *, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985475518846, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985328718202, language=CN, stringName=孟瑜, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, *, address=1 中国科学院空天信息创新研究院,北京 100094
2 中国科学院大学电子电气与通信工程学院,北京 100049, bio={"content":"

孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:

"}, bioImg=null, bioContent=

孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)]), AuthorCompany(id=1322962984963813744, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=2, ext=[AuthorCompanyExt(id=1322962984972202353, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China), AuthorCompanyExt(id=1322962984980590962, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国科学院大学电子电气与通信工程学院,北京 100049)])]), Author(id=1322962985538433408, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=xizh@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962985605542274, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985538433408, language=EN, stringName=Zhihao Xi, firstName=Zhihao, middleName=null, lastName=Xi, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985681039747, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985538433408, language=CN, stringName=席智浩, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国科学院空天信息创新研究院,北京 100094, bio={"content":"

席智浩,男,助理研究员,主要研究方向为计算机视觉、领域自适应和遥感图像解译。E-mail:

"}, bioImg=null, bioContent=

席智浩,男,助理研究员,主要研究方向为计算机视觉、领域自适应和遥感图像解译。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)])]), Author(id=1322962985773314437, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=dengyp@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962985853006215, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985773314437, language=EN, stringName=Yupeng Deng, firstName=Yupeng, middleName=null, lastName=Deng, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962985920115080, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962985773314437, language=CN, stringName=邓毓弸, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国科学院空天信息创新研究院,北京 100094, bio={"content":"

邓毓弸,男,特别研究助理,主要研究方向为遥感与变化检测。E-mail:

"}, bioImg=null, bioContent=

邓毓弸,男,特别研究助理,主要研究方向为遥感与变化检测。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)])]), Author(id=1322962986004001162, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=liudiyou@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962986079498636, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986004001162, language=EN, stringName=Diyou Liu, firstName=Diyou, middleName=null, lastName=Liu, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962986150801805, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986004001162, language=CN, stringName=刘帝佑, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国科学院空天信息创新研究院,北京 100094, bio={"content":"

刘帝佑,男,特别研究助理,主要研究方向为高分辨率遥感智能解译与矢量制图。E-mail:

"}, bioImg=null, bioContent=

刘帝佑,男,特别研究助理,主要研究方向为高分辨率遥感智能解译与矢量制图。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)])]), Author(id=1322962986226299279, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, orderNo=5, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=chenjb@aircas.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1322962986314379666, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986226299279, language=EN, stringName=Jingbo Chen, firstName=Jingbo, middleName=null, lastName=Chen, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1322962986377294227, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, authorId=1322962986226299279, language=CN, stringName=陈静波, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1 中国科学院空天信息创新研究院,北京 100094
2 中国科学院大学电子电气与通信工程学院,北京 100049, bio={"content":"

陈静波,男,研究员,主要研究方向为遥感图像智能处理和遥感大数据分析。E-mail:

"}, bioImg=null, bioContent=

陈静波,男,研究员,主要研究方向为遥感图像智能处理和遥感大数据分析。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1322962984871539053, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=1, ext=[AuthorCompanyExt(id=1322962984879927662, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China), AuthorCompanyExt(id=1322962984888316271, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984871539053, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国科学院空天信息创新研究院,北京 100094)]), AuthorCompany(id=1322962984963813744, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, xref=2, ext=[AuthorCompanyExt(id=1322962984972202353, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China), AuthorCompanyExt(id=1322962984980590962, tenantId=1146029695717560320, journalId=1249024232475115590, articleId=1322962977804136796, companyId=1322962984963813744, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国科学院大学电子电气与通信工程学院,北京 100049)])])]
王怡琳1, 2 , 孟瑜1, 2, * , 席智浩1 , 邓毓弸1 , 刘帝佑1 , 陈静波1, 2
作者信息
  • 1 中国科学院空天信息创新研究院,北京 100094
  • 2 中国科学院大学电子电气与通信工程学院,北京 100049
通讯作者:
孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:
作者简介:

王怡琳,女,硕士研究生,主要研究方向为遥感图像语义分割。E-mail:

孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:

席智浩,男,助理研究员,主要研究方向为计算机视觉、领域自适应和遥感图像解译。E-mail:

邓毓弸,男,特别研究助理,主要研究方向为遥感与变化检测。E-mail:

刘帝佑,男,特别研究助理,主要研究方向为高分辨率遥感智能解译与矢量制图。E-mail:

陈静波,男,研究员,主要研究方向为遥感图像智能处理和遥感大数据分析。E-mail:

Methods and a benchmark dataset for mountain road extraction from high-resolution remote-sensing imagery: AsiaMountain-Road
Yilin Wang1, 2 , Yu Meng1, 2, * , Zhihao Xi1 , Yupeng Deng1 , Diyou Liu1 , Jingbo Chen1, 2
Affiliations
  • 1 Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
  • 2 School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China
出版时间: 2026-10-16 doi: 10.11834/jig.250521
文章导航
收藏切换
目的

针对高分辨率遥感影像中山区道路因地形复杂、形态多样及背景干扰严重而难以准确提取的问题,以及现有道路提取数据集对山区道路适应性不足的现状,构建了首个覆盖亚洲多国典型山区的高分辨率遥感影像山区道路数据集(AsiaMountain-Road),并提出多分辨率自适应双聚合网络(multi-resolution adaptive feature dual aggregation network,MR-AFDANet)以支持山区道路自动提取算法的研究与应用。

方法

数据集从亚洲多国典型山区选取了29个高分辨率遥感影像区域,总覆盖面积达1 836 km²。结合道路稀疏、弯曲以及遮挡等复杂场景特征进行采样与标注,最终形成涵盖多种地貌和道路类型的山区道路影像与精细标注。基于该数据集,对多种主流遥感道路提取算法进行了系统基准评测,并验证了所提MR-AFDANet模型在复杂山区环境下的分割性能。

结果

实验表明,现有主流道路提取方法在AsiaMountain-Road数据集上表现出显著差异,其性能受到山区道路狭窄、弯曲、遮挡及多尺度特征的影响较大。该数据集有效揭示了现有算法在复杂山区环境中的局限性,并为算法鲁棒性提升与结构优化提供了基准支持。

结论

AsiaMountain-Road数据集填补了高分辨率遥感影像山区道路提取领域的基准数据空白,MR-AFDANet模型有效解决了复杂地形下的道路断裂与漏检问题。两者共同为山区遥感影像的智能化解译提供了高质量的数据基础与方法参考。论文相关数据集与代码下载地址:https://cstr.cn/31253.11.sciencedb.j00240.00109和https://github.com/wyl-ucas/AsianMountain-DataSet。

山区道路提取  /  高分辨率遥感  /  遥感基准数据集  /  复杂地形道路检测  /  深度学习语义分割
Objective

Mountain roads are vital infrastructure for transportation,logistics,and socioeconomic development in remote,topographically challenging regions. They provide essential connectivity for education,healthcare,resource transport,and disaster relief,yet their construction and maintenance are extremely difficult because of steep slopes,unstable geology,and frequent natural hazards. Despite their importance,automatically extracting mountain roads from high-resolution remote-sensing imagery remains a longstanding challenge. Complex terrain introduces severe occlusion and illumination variation,while the spectral appearance of roads is easily confused with bare rock,dry riverbeds,and agricultural paths. In addition,mountain roads are typically narrow,winding,and discontinuous,which differs substantially from the wide,regular patterns found in urban or rural road networks. Existing public road-extraction datasets,such as SpaceNet and DeepGlobe,were developed primarily for flat or urbanized environments. As a result,they fail to capture the spectral,geometric,and contextual complexity of mountain roads,leading to poor generalization when deep-learning models are transferred to high-relief regions. To bridge this gap,we construct AsiaMountain-Road,the first high-resolution,multi-country benchmark dataset dedicated to mountain-road extraction,and propose a novel multi-resolution adaptive feature dual aggregation network (MR-AFDANet) to evaluate and enhance model adaptability in diverse mountainous environments.

Method

The AsiaMountain-Road dataset was built from 29 high-resolution satellite-image areas selected from representative mountainous regions in 17 Asian countries,covering a total of 1 836 km² and providing detailed surface information for fine-scale mapping. All imagery underwent radiometric calibration,orthorectification,and mosaicking,followed by manual annotation by trained experts. Each region was labeled with fine-grained binary road masks to ensure pixel-level accuracy. To improve usability,auxiliary metadata,including elevation maps and land-cover references,were also provided. For algorithmic evaluation,we benchmark several mainstream deep-learning models—D-LinkNet,OareNet,C2Net,AFDANet,RCFSNet,and RS-Mamba—to assess their adaptability to mountainous terrain. Building on these findings,we propose MR-AFDANet,a model specifically designed for complex terrain mapping. MR-AFDANet uses a ResNet34 encoder as its backbone,consisting of four downsampling convolutional blocks. Each block is followed by a SEAF (spatial-enhanced adaptive feature fusion) module,which adaptively fuses intra-block and residual features while strengthening spatial information through multi-scale convolutions and attention mechanisms. In the decoder,an strip refinement attention decoder (SRADecoder) progressively fuses multi-resolution features to refine road boundaries,preserve continuity in narrow and winding segments,and suppress background noise caused by shadows and vegetation. This design enables the network to preserve thin road topology even under severe spectral interference.

Result

Experimental evaluation on the AsiaMountain-Road benchmark demonstrates the substantial difficulty of road extraction in mountainous regions. Compared with urban or plain datasets,mountain imagery contains more complex topography,stronger illumination variation,and more severe occlusion from vegetation and terrain shadows,all of which significantly degrade segmentation accuracy. Across six state-of-the-art baselines,all models show varying degrees of road omission and fragmentation,particularly on narrow,winding,and shaded mountain roads. The results further show that models trained only on source-domain data struggle to generalize to the mountainous target domain because of large distributional discrepancies. Quantitative comparisons indicate that MR-AFDANet achieves the best overall performance,obtaining the highest intersection over union(IoU),F1-score,and recall among all evaluated methods. These improvements confirm that combining multi-scale feature fusion with the SRADecoder enhances both global consistency and fine-grained boundary precision. Visual analysis further verifies this advantage: MR-AFDANet produces more continuous and complete road extractions under severe occlusion,maintains connectivity along steep and curved routes,and reduces false detections on spectrally similar surfaces.

Conclusion

We present AsiaMountain-Road,the first high-resolution,multi-country dataset dedicated to road extraction in mountainous regions. Covering 29 representative areas across 17 Asian countries,it provides detailed road-mask annotations together with auxiliary contextual and topographic data,filling an important gap in current remote-sensing resources. Baseline experiments show that existing state-of-the-art methods struggle with narrow,winding,and occluded roads in complex terrain. To address these challenges,MR-AFDANet integrates SEAF modules,a dual-resolution feature-fusion mechanism,and the SRADecoder to recover thin road structures more precisely and improve robustness across highly variable mountain landscapes. Overall,AsiaMountain-Road narrows the gap between real-world geospatial challenges and current computer-vision capabilities by offering both a demanding benchmark for algorithmic innovation and a valuable resource for applied geospatial analysis. The dataset and code associated with this paper are available at https://cstr.cn/31253.11.sciencedb.j00240.00109 and https://github.com/wyl-ucas/AsianMountain-DataSet.

mountainous road extraction  /  high-resolution remote sensing  /  remote sensing benchmark dataset  /  complex terrain road detection  /  deep learning semantic segmentation
王怡琳, 孟瑜, 席智浩, 邓毓弸, 刘帝佑, 陈静波. 高分辨率遥感影像山区道路提取方法与基准数据集:AsiaMountain-Road. 中国图象图形学报, 2026 , 31 (10) : 3590 -3605 . DOI: 10.11834/jig.250521
Yilin Wang, Yu Meng, Zhihao Xi, Yupeng Deng, Diyou Liu, Jingbo Chen. Methods and a benchmark dataset for mountain road extraction from high-resolution remote-sensing imagery: AsiaMountain-Road[J]. Journal of Image and Graphics, 2026 , 31 (10) : 3590 -3605 . DOI: 10.11834/jig.250521
山区道路是连接偏远村落、促进区域经济发展和保障国防安全的重要基础设施,在资源开发、灾害救援及区域互联互通中发挥着关键作用(陈若男 等,2022;杨昌凤和余蜀予,2023)。山区道路不仅是山区与外界联系的唯一交通通道,也是实现区域协调发展、乡村振兴和应急救援保障的关键支撑。科学识别与高效管理山区道路,对于提升山区交通韧性和灾害防御能力具有重大科学与应用价值。高分辨率遥感对地观测技术凭借其大范围覆盖、高精度和周期性优势,能够提供详细的山区道路空间信息,不仅为地图更新、交通规划和灾害应急响应等应用提供重要支撑,同时也是解决山区道路提取难题的有效手段(Bastani等,2018;Cheng等,2016)。然而,山区道路如盘山公路、林道和乡村小道,因分布分散、宽度狭窄,且易受植被、阴影、积雪及风沙等因素遮挡,使其在高分辨率遥感影像中的识别与提取面临较大挑战(Abdollahi等,2020)。
因此,亟需构建覆盖典型复杂山区地貌、具有高分辨率和多样道路类型的大规模基准数据集,并探索适应山区场景的新型提取方法,以提升道路自动化提取的精度与可靠性。
近年来,基于深度学习的遥感影像道路提取研究取得了显著进展,主要集中在多尺度特征增强、注意力机制融合及拓扑结构优化等方向。
在特征表征与网络结构改良方面,研究者致力于解决高分辨率影像中的背景干扰与尺度差异问题。Zhou等人(2018)与Ding和Bruzzone(2020)分别从多尺度特征融合与道路方向感知的角度出发,增强了对道路纹理的判别能力。与此同时,杨佳林等人(2021)针对传统U-Net的局限性改进编码器结构,有效提升了对细微道路特征的捕获能力;王勇和曾祥强(2022)则提出集成注意力机制与扩张卷积的提取模型,进一步解决了复杂背景下的特征混淆问题。此外,Lin等人(2020)通过引入空间注意力模块,显著抑制了非道路区域的噪声响应。
在复杂场景适应与拓扑保持方面,针对遮挡、阴影及细小道路的断裂问题,多种针对性策略相继提出。吴强强等人(2022)利用加权边缘损失强化了对细长结构的回归;Yang等人(2024a)和Luo等人(2024)分别通过遮挡感知机制与级联修复网络改善了道路的连通性。近期,图卷积网络(Liu等,2024)及Mamba架构(Zhao等,2024)的引入,进一步增强了模型对长距离依赖关系和微弱信号的建模能力。
然而,这些方法大多基于城市或平原道路数据训练,其训练样本在道路宽度、连续性以及背景复杂度等方面与山区场景存在显著差异,因而在迁移至山区道路时常出现泛化能力不足、断裂严重以及误检、漏检等问题。
如表1所示,近年来相继涌现了多个具有不同空间分辨率和覆盖范围的道路提取基准数据集,包括DeepGlobe(Demir等,2018)、SpaceNet(Van Etten等,2018)、Massachusetts Roads(Mnih和Hinton,2010)、GanSu-Road(Zhou等,2022)、CHN6-CUG(China 6 cities-China University of Geosciences road dataset)(Zhu等,2022)、WHU(Wuhan University road dataset)(Zhou等,2020)等。这些数据集为城市道路自动化提取提供了大量基础数据支持,也极大地促进了基于深度学习的道路提取方法的发展。
然而,这些主流方法和数据集大多针对结构规则、边界清晰的城市道路场景,缺乏对山区道路的适配能力。山区道路具有显著的形态多样性(如宽度变化大、狭窄弯曲、结构断续)以及环境复杂性(地形起伏、植被遮挡、阴影干扰),且标注困难,导致现有方法在实际应用中常出现漏检和断裂提取问题。尽管如Massachusetts Roads等数据集中包含少量山区样本,但其覆盖面与环境复杂性仍不足以支撑针对高海拔、多地貌山区的建模研究(Zhu等,2021)。本文将现有道路数据集应用于山区道路提取的局限性总结为3点:
1)地理区域覆盖不均衡。现有数据集大多选自城市或平缓丘陵地区,其影像特征、地形地貌与高海拔、大起伏的复杂山区存在显著差异。专门覆盖高海拔雪山—冰川、深切峡谷、丘陵农业区、草原盆地、火山台地、热带高原及山地城镇等典型复杂山区地貌的样本严重不足,导致基于此类数据训练的模型难以适应真实的山区环境。
2)道路类型定义单一。现有数据集的标注规范通常针对宽阔、连续的城市道路网络设计,未能充分考虑山区道路的多样性和特殊性。对于宽度多变、线形曲折、铺装材料多样(如沥青、砂石、土路)甚至结构断续的山区道路(如盘山公路、林道、乡村小道等),缺乏精确的定义和标注,导致模型无法有效学习其独特特征。
3)高质量山区标注数据规模不足。前述的地貌覆盖偏差与标注规范问题,最终导致了可用于训练的高质量山区道路数据在规模上极度匮乏。深度学习模型需要海量数据学习复杂特征,而现有数据集中真正适用于山区的有效样本数量有限,难以支撑模型充分学习如何应对植被遮挡、地形起伏以及阴影干扰等挑战,成为发展高性能山区道路提取模型的主要瓶颈。
因此,为弥补现有高分辨率道路提取数据集在山区场景下的不足,本文构建了首个面向亚洲典型山区的大规模道路提取基准数据集——AsiaMountain-Road。该数据集采集了129景1 m分辨率的大幅面原始影像,经过统一坐标转换与质量清洗后,利用固定滑窗策略裁切生成,最终数据集包含1 706对2 048 × 2 048像素和10 219对1 024 × 1 024像素的精细标注样本。在此基础上,针对AsiaMountain-Road数据集的特点及其所面临的地形起伏与植被遮挡等挑战,本文提出了一种多分辨率自适应特征双聚合网络(multi-resolution adaptive feature dual aggregation network,MR-AFDANet)。该方法设计了空间增强自适应特征融合(spatial-enhanced adaptive feature fusion,SEAF)模块与多尺度条带卷积机制,通过高效整合全局上下文与局部细节,实现了对细窄、曲折山路的逐级精细提取。实验验证表明,该方法在AsiaMountain-Road数据集上取得了优异的提取效果,可为跨境交通规划、灾害预警及国防监测等应用提供关键技术支撑。
AsiaMountain-Road数据集覆盖了亚洲17个国家的29座典型山脉(包括喜马拉雅、横断及喀喇昆仑山脉等)。选区囊括了从高海拔冰雪、深切河谷到热带雨林的多种地貌(如表2所示),从源头上确保了场景的多样性。为确保模型在真实场景中的泛化能力,本数据集完整保留了山区道路面临的典型干扰特征,包括:地形起伏引发的投影畸变、植被冠层导致的道路断续、山体阴影与强光反射带来的光照剧变,以及碎石坡与冲沟等易混淆的伪纹理背景。在此复杂的地理环境下,数据集涵盖了乡村山路、盘山公路、林道及水坝施工道等多种道路类型(详见表3),这些道路在材质、曲率及视域可见性上呈现显著差异,为算法的鲁棒性验证提供了高难度基准。
为确保标签的准确性与一致性,本研究构建了标准化的数据标注体系。如图1所示,该流程主要包含数据输入、分类处理及融合输出3个核心阶段。首先,在数据输入与预筛选阶段,对原始影像进行严格质检,剔除云量覆盖超过15%或成像质量低下的数据,并将筛选的影像经几何校正与坐标统一化处理后,作为标注基础。随后进入差异化分类标注阶段,针对山区道路显著的形态差异采取混合标注策略:对于边缘清晰、路面较宽的主干道,采取“中心线矢量化 + 缓冲区生成”的方式制作道路面状图斑(依据实测宽度3~6 m);对于受植被遮挡严重、边界模糊或路面狭窄的乡村土路与林道,则直接进行像素级面状精细描绘,并引入OpenStreetMap矢量数据辅助推断遮挡区域的潜在连通性。最后,在融合输出与标准化阶段,对完成标注的所有矢量图层进行空间融合与多轮质量校验,重点修复断裂、伪影及拓扑错误,并将校验无误的矢量真值在统一坐标系下转换为1 m分辨率栅格标签,依据固定步长裁切为标准样本块(2 048 × 2 048像素与1 024 × 1 024像素),最终生成高精度对齐的影像—标签对。
AsiaMountain-Road数据集共包含129景1 m空间分辨率、空间尺寸6 144 × 4 352像素的高分辨率遥感影像,经过影像切片与清洗处理,最终获取1 706对2 048 × 2 048像素大小的影像切片与其对应的道路标注以及10 219对1 024 × 1 024像素大小的影像切片与其对应的道路标注。
为了评估深度学习模型的泛化性能,本研究将两种尺度的样本均按照7∶2∶1的比例随机划分为训练集、测试集和验证集。其中,2 048 × 2 048像素尺度数据集包含1 213幅训练样本、333幅测试样本和160幅验证样本;1 024 × 1 024像素尺度数据集包含7 183幅训练样本、1 936幅测试样本和1 100幅验证样本。数据集对应的山区类型分布如表2所示,部分标注示例如图2所示,AsiaMountain-Road数据集典型山区类型及核心特征如表3所示。
与现有道路数据集相比,本文提出的AsiaMountain-Road数据集具有以下4个显著特点:
1)更高的空间分辨率与更广泛的地理覆盖。数据集采用1 m分辨率影像,有效保留了细窄山路的空间形态。研究区涵盖亚洲29个典型山脉(跨越中国、尼泊尔、伊朗等17国)。
2)更强的地形异质性与道路类型多样性。为全面反映山区复杂场景,数据集涵盖了6类形态各异的典型道路:占比最高的乡村山路(38.7%)表现为窄而弯曲的土质线状结构,极易与裸土或耕地纹理混淆;盘山公路(22.1%)呈现为宽度稳定的深灰色连续铺装带,线性显著但常受山体阴影干扰;林道(16.6%)提取难度最高,表现为林冠间断续的细窄缝隙;此外,还包含受建筑纹理干扰的村际联络线(8.1%),以及受同质背景与施工纹理影响的水坝施工道(7.5%)和矿区道路(7.0%)。这种从“窄弯土质”到“深灰铺装”、从“林间缝隙”到“粗糙条带”的多样化特征,为验证模型在强干扰背景下的鲁棒性提供了高价值样本。
3)更系统的复杂场景覆盖与精准标注。数据集系统纳入了森林树冠遮蔽、山体阴影、积雪云雾等多样化干扰场景,以及排水沟、地块边界等易混淆负样本。同时,采用精细化像素级面状标注策略,经多轮校验,平均边界误差仅为0.205像素,确保了标签与实际道路几何形态的高度匹配。
4)显著的“细长稀疏”结构特征与背景高异质性。统计表明,数据集中的道路像素仅占全部像素的0.56%,背景像素占99.44%,道路总长度约3 433.50 km,呈现出显著的“细长且稀疏”特征。道路平均断裂比例为0.126%,说明整体标注连续性良好。从背景像素的内部构成来看,植被、裸土、岩石、建筑物、积雪和云影分别占45.07%、32.58%、20.83%、0.75%、0.73%和0.04%,合计为100%。其中,植被与裸土是最主要的背景类型,其较高占比反映了亚洲典型山区林地覆盖、坡耕台地及碎石裸露地物广泛分布的特点,也构成了具有高度异质性的干扰环境。
本文提出的亚洲山区道路数据集面临的挑战包括以下4个方面:
1)地形起伏剧烈导致影像变化显著。亚洲山区地势复杂、高差起伏大,导致遥感影像中存在大量阴影、遮挡和强烈的地形畸变。如图3(a)所示,同一类型道路在不同坡度和照射角度下呈现出显著的光谱和纹理变化。这种变化加剧了模型对道路类别的判别难度,因此,如何增强模型对地形变化鲁棒性的表达能力,是该数据集的重要挑战之一。
2)道路形态复杂且结构连通性弱。山区道路通常具有窄、弯曲、断裂和分叉多的特点,尤其在林区、陡坡或建筑阴影遮挡区域,道路的可见性极低。如图3(b)所示,道路可能被山体、植被、建筑或云影部分遮挡,造成明显的结构中断。这要求模型具备较强的形状先验和长距离依赖建模能力,以恢复道路全局结构与连通性,是该数据集的又一关键挑战。
3)道路样本极度稀疏导致类别不平衡严重。山区居民分布分散,道路网络密度远低于城市或平原地区。根据统计,本数据集中道路像素占比通常低于5%,甚至低于1%。如图3(c)所示,大面积背景(如山体、林地)覆盖图像,仅有细线状的道路穿插其间。前景—背景比例极度失衡导致模型易产生过拟合或前景忽略问题,因此,如何提升模型对极小目标道路的敏感性与鲁棒性,是该数据集的重要研究问题。
4)道路与自然背景的光谱相似性高。山区道路常为泥土路、砂石路等未经硬化的简易道路,其光谱特征与周围土地(如裸地、梯田、枯草地)高度相似。如图3(d)所示,这种相似性极易导致模型发生误检(非道路被识别为道路)或漏检(道路被误判为背景)。尤其是在植被覆盖率高或阴影区域,道路边缘模糊不清。因此,如何在光谱与纹理区分度低的背景下准确提取道路边界和主干结构,也是本数据集的一大挑战。
为了提升高分辨率山区道路的语义分割精度,本文提出了多分辨率自适应双聚合网络(MR-AFDANet),结合ResNet34编码器的多尺度特征提取能力与SEAF模块的自适应融合能力,并通过全局上下文增强和高低分辨率融合逐步恢复细节特征,实现对细长道路结构的高精度分割。网络整体架构包括编码器、高低分辨率融合模块和SRADecoder(strip refinement attention decoder),数据流从编码器输出开始,依次经过基于低分辨率分支的全局上下文增强与特征融合,最终生成高分辨率分割结果,如图4所示。
编码器采用ResNet34作为主干,由4层下采样卷积块组成,每层下采样后特征图尺寸依次为H/4 × W/4、H/8 × W/8、H/16 × W/16、H/32 × W/32,通道数分别为64、128、256、512。每个卷积块后接改进的空间增强自适应特征融合(SEAF)模块,用于自适应融合编码器内部和主干残差特征,同时结合多尺度卷积和注意力机制增强空间信息表达。该设计能够有效捕捉高分辨率道路的局部细节,同时保留全局结构信息。需指出,本研究虽然沿用SEAF术语(Deng等,2025),但在结构上进行了针对性改良。本文通过引入深度可分离卷积重构了模块,以弥补原始设计在山区细小道路提取上的不足,从而实现了更优的特征融合与抗噪性能。
为了兼顾局部细节与全局上下文信息,网络设计了高低分辨率双流分支。低分辨率分支以特定倍率下采样的影像为输入,使用轻量ResNet18提取全局上下文特征,通过1 × 1卷积投影到与高分辨率特征相同的通道维度。随后,将低分辨率特征上采样后与高分辨率特征进行拼接,并通过融合卷积进一步整合信息。该模块实现了局部细节与全局结构的有效结合,使网络在处理细长、曲折道路时能够保持连续性与完整性。
解码器部分采用条带细化注意力解码器(strip refinement attention decoder,SRADecoder)。该模块首先利用多尺度条带卷积分别在水平与垂直方向上提取细长道路的结构化响应,减少常规卷积在捕获细线状目标时出现的模糊与断裂。随后,引入卷积块注意力模块(convolutional block attention module,CBAM)对通道与空间维度进行联合注意力增强,使网络能够自适应突出道路区域并抑制复杂背景干扰。
在经过条带卷积与注意力增强后,特征通过逐级上采样(转置卷积)恢复空间分辨率,并与编码器的多尺度特征进行融合,以保证结构连续性与边界精细化。最终输出经Sigmoid激活映射为像素级道路概率图。该解码器在高分辨率场景中能够有效保持道路的拓扑连通性,并增强对细长道路边界与弱纹理区域的识别能力。
在推理阶段,针对大尺寸影像,模型采用基于重叠裁切的滑动窗口预测策略。通过设定固定步长与窗口重叠率,对整幅影像进行逐块预测,并在重叠区域采用加权平均融合,以消除切割边界处的伪影与断裂现象。得益于网络内部的高低分辨率融合机制,模型能够利用低分辨率分支的全局上下文信息约束局部预测,从而确保了大幅面影像分割结果的全局一致性。该机制使得模型能够在GPU显存受限的条件下,高效处理2 048 × 2 048像素乃至更大尺寸的输入,兼顾了计算效率与拓扑连通性。
评价指标包括:精确率(precision,P)、召回率(recall,R)、F1分数(F1-score,F1)和交并比(intersection over union,IoU)。精确率为预测到属于道路像素的结果中正确预测结果所占的比例。召回率是检测到的正确道路像素与所有真实道路像素的比值。F1分数和交并比是2个综合性的评价指标,F1分数是精确率和召回率的调和平均值,交并比是预测值和标签的重合比率。
本文选取PyTorch为深度学习框架,服务器显卡采用NVIDIA 3090 GPU(24 GB),并选取6种主流的遥感道路提取方法作为对比方法,分别为D-LinkNet(LinkNet with pretrained encoder and dilated convolution)(Zhou等,2018)、OareNet(Yang等,2024a)、C2Net(context perception and cross spatial-scale feature interaction network)(Yang等,2024b)、RCFSNet(road context and full-stage features from satellite imagery network)(Yang等,2022)、AFDANet(adaptive full-stage feature fusion and directional-aware network)(Deng等,2025)和RS-Mamba(Zhao等,2024)。在模型训练阶段,所有训练样本采用水平翻转、垂直翻转和角度旋转等方式进行数据增强。所有模型均使用Adam优化器训练100轮,批次大小为2,初始学习率为10-4,在第{60,80}轮时将学习率衰减为原来的1/5。损失函数为二类交叉熵损失与Dice损失的加权和。在推理阶段,不采用数据增强,最终输出的概率图以0.5为阈值转为二值图。
在山区道路提取任务中,深度学习模型同样面临显著挑战。与平原或城市区域相比,山区影像往往存在更为复杂的地形起伏和影像干扰因素。本文在AsiaMountain-Road测试集上对在DeepGlobe上训练的模型进行评估,结果见图5。
可以观察到,在山区场景中,模型存在一定程度的道路漏检和断裂问题,尤其是在阴影覆盖、林地遮挡以及狭窄弯曲道路区域,识别精度明显下降。造成这一现象的根本原因在于山区场景与常见的平原或城市数据存在明显的分布差异。具体而言,山区影像受地形起伏、植被覆盖和影像阴影等因素影响,道路纹理特征更复杂,光谱表现更不稳定。此外,山区道路普遍较窄,空间分布弯曲且断续,常常被植被或建筑物遮挡,从而导致模型在预测时难以保持道路的连续性。
这一实验结果充分说明,仅依赖现有公开数据集训练的模型无法满足山区道路提取需求,也凸显了构建专门针对山区道路的高分辨率数据集(AsiaMountain-Road)的必要性。在跨域场景下,结合目标域的山区样本进行训练或采用跨域适配方法,能显著提升模型在复杂山区环境下的道路提取性能。
在此基础上,为了全面评估不同深度学习方法在山区道路提取任务中的性能,本文在AsiaMountain-Road数据集上对以上具有代表性的基于深度学习的道路提取方法进行了基准测试,结果如表4所示。其中,本文方法的总体表现最好,在交并比、F1分数和召回率等指标上均取得了最优结果,交并比达到59.40%,F1分数达到74.53%。这表明多尺度注意力、高低分辨率特征融合以及条带卷积结构化细线提取策略在复杂山区场景中具有显著优势。结合图6可视化结果,可以更细致地阐明各类道路(宽主干道、窄乡村路、遮挡道路、光谱相似背景道路)在不同算法下的提取表现,进一步揭示不同方法在不同道路类型上的适应性差异。
首先,在宽阔且结构清晰的主干道路上,各模型整体表现均较为稳定,基本能够保持道路的完整边界。然而,D-LinkNet、C2Net和RCFSNet在局部弯折处仍存在轻微边界抖动,而AFDANet与RS-Mamba的提取结果较为平滑。MR-AFDANet在该类道路上的表现最为稳健,能够同时保持边界的规整性与形态一致性。
其次,在狭窄、弯曲的乡村道路上,模型性能差异显著。D-LinkNet、OareNet与C2Net在窄路段普遍出现断裂、漏检,难以维持拓扑连续性。RS-Mamba在部分区域能够保持一定连通性,但仍有局部缺失。相比之下,MR-AFDANet依托SEAF特征增强与多尺度细节融合,能够更完整地恢复细道路的连续走向,明显降低断裂率。
对于被阴影、林地或地形遮挡的道路,多数卷积主导的模型受到局部纹理干扰,提取结果常出现道路中断。AFDANet和RS-Mamba在全局建模方面有所提升,但重度遮挡下仍存在缺口。MR-AFDANet借助低分辨率全局语义与高分辨率细节的互补性,能够更好地推理出被部分遮挡的道路结构,实现较为完整的恢复。
在光谱特征与道路相似的裸岩、碎石坡或山体背景区域,传统模型易产生误检。“假正类”现象在C2Net、OareNet中尤为明显。RS-Mamba和AFDANet在一定程度上降低了误检,但仍会受到局部纹理的干扰。MR-AFDANet通过引入更强的上下文约束,显著减少了此类高背景干扰场景下的误检概率。
综合来看,MR-AFDANet在不同类型的道路场景中均展现出更优的鲁棒性与泛化能力,特别是在狭窄道路连续性保持、遮挡场景恢复以及高相似背景下减少误检方面具有明显优势。
为验证本文提出的MR-AFDANet中各个关键模块的有效性,在AsiaMountain-Road数据集上进行了系统的消融实验。实验以基础的编码器—解码器结构作为基线模型,逐步引入SEAF模块、高低分辨率融合模块和SRADecoder模块等核心组件,定量结果如表5所示。
实验结果显示,Baseline模型(实验1)的精确率为73.97%,召回率为58.83%,F1分数为65.54%,IoU为48.74%。该模型在各项指标上均表现较低,特别是召回率与IoU明显不足,表明基础编码器—解码器结构难以充分捕捉山区道路的复杂空间特征与上下文关系,尤其在弱纹理、多遮挡的山区场景下漏检严重。
当引入SEAF模块后(实验2),模型精确率显著提升至77.88%(↑3.91%),召回率提升至66.35%(↑7.52%),F1分数提升至71.70%,IoU提升至53.05%。这表明SEAF模块通过自适应特征融合机制,增强了模型对道路关键特征的聚焦能力,有效抑制了背景干扰,从而提高了像素级分类的准确性。
当单独引入高低分辨率融合模块后(实验3),模型召回率大幅提升至70.23%(↑11.40%),是所有单模块中召回率提升最显著的组件。同时,F1分数提升至73.19%,IoU提升至55.41%。这说明多尺度特征融合机制能够有效整合不同层次的特征信息,特别是利用低分辨率分支的全局上下文信息弥补了高分辨率细节的缺失,增强模型对窄小道路、模糊边界及多尺度道路片段的检测能力,显著减少了漏检现象。
单独加入SRA解码器(实验4)后,模型在精确率(77.63%)、召回率(68.87%)和IoU(55.18%)上均取得均衡提升,F1分数达到73.02%。这表明SRA解码器通过增强空间上下文建模与特征重建能力,有助于缓解因阴影、植被覆盖造成的道路断裂与边缘模糊问题,提升分割结果的连续性。
最终,完整模型(实验5)整合了SEAF模块、高低分辨率融合模块与SRA解码器,在各项指标上均取得最优表现:精确率80.10%、召回率71.45%、F1分数74.53%、IoU59.40%。与Baseline相比,完整模型在F1分数和IoU上分别提升了8.99%和10.66%,且精确率与召回率得到更为均衡的改善。这充分验证了各模块在提升山区道路提取性能方面的有效性与互补性:SEAF模块增强特征判别力,高低分辨率融合模块扩展感知尺度,SRA解码器优化空间重建,三者协同工作,共同解决了山区道路提取中的多尺度感知、特征干扰与连续性保持等关键问题。
为验证所提出模型在不同地貌与影像分布下的跨域泛化能力,本文在AsiaMountain-Road数据集上完成模型训练后,直接在Massachusetts-Roads道路数据集上进行测试评估。该实验设置体现了从复杂山区地形到规则城市道路场景的显著分布迁移差异,用以检验模型在极端域间变化下的鲁棒性。
Massachusetts-Roads数据集主要包含美国波士顿地区的城市影像,道路形态规整、宽度均匀。与之相比,AsiaMountain-Road数据集影像来源于亚洲典型山区地区,道路狭窄、形态弯曲、受地形与阴影干扰显著,两者在影像风格、纹理分布及语义上下文方面存在较大域间差异。
表6展示了各典型方法在跨域测试中的性能对比结果。由表可见,所有模型在迁移到城市数据集后性能均出现不同程度下降。值得注意的是,D-LinkNet虽然召回率较高,但精确率偏低,表明其容易将非道路背景误检为道路。相比之下,本文提出的MR-AFDANet取得了最佳结果,交并比达60.93%、F1分数为75.72%,分别较D-LinkNet提升5.42%和4.33%。这表明MR-AFDANet具备较强的跨域泛化能力,其多尺度上下文融合与SEAF自适应融合模块有效增强了模型对不同地貌场景的适应性与稳健性。
图7展示了不同方法在Massachusetts-Roads测试集上的可视化结果。可以观察到,现有对比方法在跨域道路场景中容易出现道路边缘模糊与结构断裂问题,而MR-AFDANet能够较好保持道路结构连续性和边界清晰度,显示出其良好的跨域适应能力。
本文针对亚洲复杂山区道路智能提取的现实需求,构建了首个跨17国的大规模山区道路基准数据集AsiaMountain-Road,并系统评估了6种主流深度学习模型在该场景下的适应性。在此基础上,针对山区道路在高差、阴影与遮挡条件下易出现漏检与断裂的问题,进一步提出了增强山区适应性的道路提取模型MR-AFDANet,以提升在复杂地貌中的道路检测稳定性与连续性。研究主要结论如下:
1)数据集贡献与特性。AsiaMountain-Road在覆盖范围、标注质量与任务难度方面均具有显著优势。(1)覆盖范围广泛:数据集涵盖高山、高原、峡谷、山谷城市等7类典型山区道路场景,具有较高的区域多样性与地貌代表性。(2)标注精度高:基于1 m分辨率遥感影像构建,并采用统一规范进行像素级道路标注,经多轮质量校验确保几何精度和拓扑一致性,为山区道路研究提供高可信度基准。(3)任务难度突出:针对山区特有的光照变化、植被遮挡和复杂地形影响,数据集涵盖了山体阴影、林下遮挡及碎石坡等高难度场景。这种对“局部边界模糊”与“弱可见性”特征的真实保留,相比现有数据集显著增强了挑战性,可全面检验模型在极端环境下的鲁棒性与泛化能力。
2)模型适应性分析与本文方法的改进。实验结果表明,现有道路提取模型在山区场景中仍存在以下方面局限。(1)对高差导致的道路形态畸变适应不足;(2)易受山体阴影影响,在遮挡区域产生明显漏检;(3)在狭窄盘山道路上难以保持连续、完整的拓扑结构。为解决上述问题,本文提出了MR-AFDANet,通过多尺度注意力机制、高低分辨率特征融合以及条带卷积结构化细线提取方式,增强了模型在复杂地形、植被遮挡与阴影干扰环境下的特征表达能力。实验结果表明,MR-AFDANet在多项评估指标上均显著优于现有对比方法,验证了模型在复杂山区高分辨率遥感场景中的适应能力与稳健性。
3)潜在应用价值与应用场景。得益于其高分辨率、广覆盖与高挑战性,AsiaMountain-Road数据集在国家战略与实用场景中具有重要应用潜力,包括: “一带一路”沿线山区交通基础设施监测与规划;偏远山区道路普查、抢险救援路线识别与应急调度;开源地图(如OSM(open street map))道路要素缺失区域的补全与更新;面向山区的智能交通、遥感制图自动化与地理要素矢量化研究。
4)未来展望。后续研究可围绕以下方向展开:(1)提升模型在多海拔梯度下的特征泛化能力;(2)加强对山体阴影、滑坡等动态环境的稳健识别;(3)探索小样本与迁移学习方法,缓解山区标注样本不足的问题;(4)融合多时相遥感与激光雷达等多源数据,构建新一代山区智能交通监测体系。
综上,AsiaMountain-Road数据集不仅揭示了山区道路智能提取的核心挑战,也为后续算法改进和应用推广奠定了重要基础。
支撑本研究的科学数据已在中国科学院科学数据银行ScienceDB以及GitHub平台公开发布,访问地址分别为https://cstr.cn/31253.11.sciencedb.j00240.00109和https://github.com/wyl-ucas/AsianMountain-DataSet。
  • “十四五”民用航天技术预先研究项目(D040404)
参考文献 引证文献
排序方式:
Abdollahi A,Pradhan B,Shukla N,Chakraborty S and Alamri A. 2020. Deep learning approaches applied to remote sensing datasets for road extraction: a state-of-the-art review. Remote Sensing,12(9): #1444 [DOI: 10.3390/rs12091444]
Bastani F,He S T,Abbar S,Alizadeh M,Balakrishnan H,Chawla S,et al. 2018. RoadTracer: automatic extraction of road networks from aerial images//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City,USA: IEEE: 4720-4728 [DOI: 10.1109/CVPR.2018.00496]
Chen R N,Peng L,Liu Y F,Wei Z C,Lyu B R and Chen D Y. 2022. Road information extraction and application in the suburban mountainous area based on remote sensing images. Journal of University of Chinese Academy of Sciences,39(5): 658-667
陈若男,彭玲,刘玉菲,卫志超,吕蓓茹,陈德跃. 2022. 引入空间距离信息的城郊山区道路提取与应用. 中国科学院大学学报,39(5): 658-667 [DOI: 10.7523/j.ucas.2021.0004]
Cheng G L,Zhu F Y,Xiang S M and Pan C H. 2016. Road centerline extraction via semisupervised segmentation and multidirection nonmaximum suppression. IEEE Geoscience and Remote Sensing Letters,13(4): 545-549 [DOI: 10.1109/2016.2524618]
Demir I,Koperski K,Lindenbaum D,Pang G,Huang J,Basu S,et al. 2018. DeepGlobe 2018: a challenge to parse the earth through satellite images//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Salt Lake City,USA: IEEE: 172-183 [DOI: 10.1109/CVPRW.2018.00031]
Deng X L,Li M Z and Chen Z H. 2025. AFDANet: an adaptive full-stage feature fusion and directional-aware network for road extraction in remote sensing images. IEEE Geoscience and Remote Sensing Letters,22: #6010505 [DOI: 10.1109/LGRS.2025.3583214]
Ding L and Bruzzone L. 2020. DiResNet: direction-aware residual network for road extraction in VHR remote sensing images. IEEE Transactions on Geoscience and Remote Sensing,58(9): 6494–6508 [DOI: 10.1109/TGRS.2020.3034011]
Lin Y N,Xu D Y,Wang N,Shi Z and Chen Q X. 2020. Road extraction from very-high-resolution remote sensing images via a nested SE-Deeplab model. Remote Sensing,12(18): #2985 [DOI: 10.3390/rs12182985]
Liu R Y,Wu J H,Lu W Y,Miao Q G,Zhang H,Liu X Z,et al. 2024. A review of deep learning-based methods for road extraction from high-resolution remote sensing images. Remote Sensing,16(12): #2056 [DOI: 10.3390/rs16122056]
Luo H,Wang Z J,Du B and Dong Y N. 2024. A deep cross-modal fusion network for road extraction with high-resolution imagery and LiDAR data. IEEE Transactions on Geoscience and Remote Sensing,62: #4503415 [DOI: 10.1109/TGRS.2024.3360963]
Mnih V. 2013. Machine Learning for Aerial Image Labeling. Toronto:University of Toronto:[s.n.]
Mnih V and Hinton G E. 2010. Learning to detect roads in high-resolution aerial images//Proceedings of the 11th European Conference on Computer Vision. Heraklion,Greece: Springer: 210-223 [DOI: 10.1007/978-3-642-15561-1_16]
Van Etten A,Lindenbaum D and Bacastow T M. 2018. SpaceNet: a remote sensing dataset and challenge series [EB/OL]. [2025-10-10]. https://arxiv.org/pdf/1807.01232.pdf
Wang Y and Zeng X Q. 2022. Road extraction model derived from integrated attention mechanism and dilated convolution. Journal of Image and Graphics,27(10): 3102-3115
王勇,曾祥强. 2022. 集成注意力机制和扩张卷积的道路提取模型. 中国图象图形学报,27(10): 3102-3115 [DOI: 10.11834/jig.210226]
Wu Q Q,Wang S,Wang B and Wu Y L. 2022. Road extraction method of high-resolution remote sensing image on the basis of the spatial information perception semantic segmentation model. National Remote Sensing Bulletin,26(9): 1872-1885
吴强强,王帅,王彪,吴艳兰. 2022. 空间信息感知语义分割模型的高分辨率遥感影像道路提取. 遥感学报,26(9): 1872-1885 [DOI: 10.11834/jrs.20210021]
Yang C F and Yu S Y. 2023. Research on route scheme of highway in difficult and dangerous mountainous area based on construction and rescue factors. Journal of China and Foreign Highway,43(5): 273-276
杨昌凤,余蜀予. 2023. 基于施工及救援因素的艰险山区公路路线方案研究. 中外公路,43(5): 273-276 [DOI: 10.14048/j.issn.1671-2579.2023.05.046]
Yang J L,Guo X J and Chen Z H. 2021. Road extraction method from remote sensing images based on improved U-Net network. Journal of Image and Graphics,26(12): 3005-3014
杨佳林,郭学俊,陈泽华. 2021. 改进U-Net型网络的遥感图像道路提取. 中国图象图形学报,26(12): 3005-3014 [DOI: 10.11834/jig.202863]
Yang R Y,Zhong Y F,Liu Y H,Lu X Y and Zhang L P. 2024a. Occlusion-aware road extraction network for high-resolution remote sensing imagery. IEEE Transactions on Geoscience and Remote Sensing,62: #5619316 [DOI: 10.1109/TGRS.2024.3387945]
Yang Z G,Zhang W,Li Q,Ni W P,Wu J Z and Wang Q. 2024b. C2Net: road extraction via context perception and cross spatial-scale feature interaction. IEEE Transactions on Geoscience and Remote Sensing,62: #5647011 [DOI: 10.1109/TGRS.2024.3491755]
Yang Z G,Zhou D X,Yang Y,Zhang J P and Chen Z H. 2022. Road extraction from satellite imagery by road context and full-stage feature. IEEE Geoscience and Remote Sensing Letters,20: #8000405 [DOI: 10.1109/LGRS.2022.3228967]
Zhao S J,Chen H,Zhang X L,Xiao P F,Bai L and Ouyang W L. 2024. RS-Mamba for large remote sensing image dense prediction. IEEE Transactions on Geoscience and Remote Sensing,62: #5633314 [DOI: 10.1109/TGRS.2024.3425540]
Zhou G D,Chen W T,Gui Q S,Li X J and Wang L Z. 2022. Split depth-wise separable graph-convolution network for road extraction in complex environments from high-resolution remote-sensing images. IEEE Transactions on Geoscience and Remote Sensing,60: #5614115 [DOI: 10.1109/TGRS.2021.3128033]
Zhou L C,Zhang C and Wu M. 2018. D-LinkNet: LinkNet with pretrained encoder and dilated convolution for high resolution satellite imagery road extraction//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Salt Lake City,USA: IEEE: 192-1924 [DOI: 10.1109/CVPRW.2018.00034]
Zhou M T,Sui H G,Chen S X,Wang J D and Chen X. 2020. BT-RoadNet: a boundary and topologically-aware neural network for road extraction from high-resolution remote sensing imagery. ISPRS Journal of Photogrammetry and Remote Sensing,168: 288-306 [DOI: 10.1016/j.isprsjprs.2020.08.019]
Zhu Q Q,Guo X,Deng W H,Shi S N,Guan Q F,Zhong Y F,et al. 2022. Land-use/land-cover change detection based on a Siamese global learning framework for high spatial resolution remote sensing imagery. ISPRS Journal of Photogrammetry and Remote Sensing,184: 63-78 [DOI: 10.1016/j.isprsjprs.2021.12.005]
Zhu Q Q,Zhang Y N,Wang L Z,Zhong Y F,Guan Q F,Lu X Y,et al. 2021. A global context-aware and batch-independent network for road extraction from VHR satellite imagery. ISPRS Journal of Photogrammetry and Remote Sensing,175: 353-365 [DOI: 10.1016/j.isprsjprs.2021.03.016]
2026年第31卷第10期
PDF下载
30
13
引用本文
BibTeX
文章信息
doi: 10.11834/jig.250521
  • 接收时间:2025-10-28
  • 首发时间:2026-10-30
  • 出版时间:2026-10-16
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2025-10-28
  • 修回日期:2026-03-02
基金
Pre-Research Project of Civil Aerospace Technology under the 14th Five-Year Plan (D040404)
“十四五”民用航天技术预先研究项目(D040404)
作者信息
    1 中国科学院空天信息创新研究院,北京 100094
    2 中国科学院大学电子电气与通信工程学院,北京 100049

通讯作者:

孟瑜,通信作者,女,研究员,主要研究方向为跨域数据智能处理。E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/zgtxtxxb/CN/10.11834/jig.250521
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数
科
Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)
属
Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏