智能化农业装备学报（中英文）

基于GOP-YOLO11n光伏清洗无人机的轻量化多尺度表面污渍检测方法

预览PDF | 下载PDF

赵立军 ¹^,² , 胡繁森 ¹ , 代凡程 ¹ , 李强 ¹^,² , 李建 ³ , 黎斌 ¹^,²^,^*^,^*

智能化农业装备学报（中英文） | 2026,7(2): 66-76

收起

智能化农业装备学报（中英文） | 2026 , 7 (2) : 66 -76

基于GOP-YOLO11n光伏清洗无人机的轻量化多尺度表面污渍检测方法

全屏

赵立军¹^,², 胡繁森¹, 代凡程¹, 李强¹^,², 李建³, 黎斌¹^,²^,^*^,^*

作者信息

^1.重庆文理学院，重庆市，402160

^2.丘陵山地智能农机装备重庆市高校工程研究中心，重庆市，402160

^3.重疆无人机科技（重庆）有限公司，重庆市，402160

通讯作者:

^* 黎斌，男，1991年生，重庆綦江人，博士，副教授；研究方向为山地智能农业装备及机器人。E-mail：cncqlibin@163.com

作者简介:

赵立军，男，1980年生，黑龙江克山人，博士，教授；研究方向为山地智能农业装备及机器人。E-mail：20190005@cqwu.edu.cn

GOP-YOLO11n-Based lightweight multi-scale surface soiling detection method for photovoltaic cleaning UAVs

Lijun Zhao¹^,², Fansen Hu¹, Fancheng Dai¹, Qiang Li¹^,², Jian Li³, Bin Li¹^,²

Affiliations

^1.Chongqing University of Arts and Sciences, Chongqing 402160, China

^2.Chongqing Municipal Engineering Research Center for Intelligent Agricultural Machinery and Equipment in Hilly and Mountainous Areas, Chongqing 402160, China

^3.Chongjiang UAV Technology (Chongqing) Co., Ltd., Chongqing 402160, China

Corresponding Author:

^* Li Bin, E-mail: cncqlibin@163.com

Zhao Lijun, E-mail: 20190005@cqwu.edu.cn

出版时间: 2026-05-15 doi: 10.12398/j.issn.2096-7217.2026.02.006

文章导航

摘要

收起

针对光伏电站中污渍积累导致发电效率降低，以及清洗无人机受限于边缘算力和复杂背景干扰难以精准定位的问题，提出一种基于Jetson Orin Nano边缘计算平台的清洗无人机解决方案。在YOLO11n模型框架下，提出一种基于GOP-YOLO11n的轻量化光伏板污渍检测算法。引入GroupDC结构，并对骨干网络中的C3k2模块进行轻量化重构，在降低模型计算开销的同时增强复杂背景下小目标特征的表达能力。在特征融合网络中加入OptimizedDC多尺度特征提取模块以提升对多尺度目标（微小鸟粪、长条水渍）的感知能力。检测头通过上采样和特征聚合块UP-Block，强化多像素图形聚合能力。试验采用无人机采集的光伏板可见光图像构建数据集，共获得4 190张有效图像，并按照8∶1∶1划分训练集、验证集和测试集，输入图像统一调整为960像素×960像素。在相同训练与测试条件下，GOP-YOLO11n的准确率、召回率、mAP@50和mAP@50：0.95分别达到91.1%、85.8%、92.5%和76.8%，较YOLO11n基线模型分别提高2.9、3.4、1.8和4.4个百分点。模型参数量由2.582 M降至2.068 M，浮点运算量为6.7×10⁹。进一步在Jetson Orin Nano Super 8G平台上进行TensorRT FP16部署测试，模型推理速度达到114.7帧/s，单帧延时为8.7 ms，显存占用为1 685 MB。结果表明，该方法在提升复杂背景下多尺度污渍检测精度的同时，兼顾模型轻量化与边缘端实时推理需求，可为光伏清洗无人机的污渍识别、区域化清洗决策和智能运维提供视觉感知支撑。

关键词

光伏板污渍检测 / GOP-YOLO11n / 轻量化 / 多尺度融合 / 边缘部署

Abstract

收起

To address the degradation of power generation efficiency caused by soiling accumulation in photovoltaic (PV) power plants, as well as the localization challenges faced by cleaning unmanned aerial vehicles (UAVs) due to limitations in edge computing power and complex background interference, this paper proposes an edge-computed cleaning UAV solution based on the Jetson Orin Nano platform. Within the YOLO11n framework, a lightweight photovoltaic panel soiling detection algorithm designated as GOP-YOLO11n is developed. By introducing GroupDC structure, the C3k2 modules in the backbone network undergo lightweight reconstruction, minimizing computational overhead while enhancing the feature representation of small targets against complex backgrounds. Furthermore, an OptimizedDC multi-scale feature extraction module is embedded into the feature fusion network to improve the perception of multi-scale targets, such as tiny bird droppings and elongated water stains. In the detection head, an upsampling and feature aggregation block (UP-Block) is implemented to strengthen multi-pixel pattern aggregation. For the experiments, a visible-light image dataset of PV panels captured by UAVs, yielding 4 190 valid imagespartitioned into training, validation, and test sets at an 8：1：1 ratio, with input resolution standardized to 960 pixels × 960 pixels. Under identical training and testing configurations, GOP-YOLO11n achieved a Precision of 91.1%, a Recall of 85.8%, a mAP@50 of 92.5%, and a mAP@50：0.95 of 76.8%, outperforming the baseline YOLO11n model by 2.9, 3.4, 1.8, and 4.4 percentage points, respectively. The model parameters were reduced from 2.582 M to 2.068 M, with a floating-point operation count of 6.7×10⁹. Moreover, hardware deployment tests using TensorRT FP16 on the Jetson Orin Nano Super 8 G platform showed an inference speed of 114.7 f/s, a single-frame latency of 8.7 ms, and a GPU memory footprint of 1 685 MB. The results indicate that the proposed method effectively reconciles the need for high-accuracy multi-scale soiling detection in complex scenes with the constraints of lightweight edge-side real-time inference, thereby providing robust visual perception support for soiling recognition, localized cleaning strategies, and the intelligent operation and maintenance (O&M) of PV cleaning UAVs.

Key words

photovoltaic stain detection / GOP-YOLO11n / lightweight model / multi-scale feature fusion / edge deployment

引用本文

赵立军, 胡繁森, 代凡程, 李强, 李建, 黎斌. 基于GOP-YOLO11n光伏清洗无人机的轻量化多尺度表面污渍检测方法. 智能化农业装备学报（中英文）, 2026 , 7 (2) : 66 -76 . DOI: 10.12398/j.issn.2096-7217.2026.02.006

Lijun Zhao, Fansen Hu, Fancheng Dai, Qiang Li, Jian Li, Bin Li. GOP-YOLO11n-Based lightweight multi-scale surface soiling detection method for photovoltaic cleaning UAVs[J]. Journal of Intelligent Agricultural Mechanization, 2026 , 7 (2) : 66 -76 . DOI: 10.12398/j.issn.2096-7217.2026.02.006

基金

收起

2025年重庆市研究生科研创新项目(CYS25872)
重庆市永川区“揭榜挂帅”项目(2025yc-jbgs20005)
2025年重庆文理学院校级研究生科研创新项目

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

[1]

Wan L, Zhao L, Xu W, et al. Dust deposition on the photovoltaic panel: a comprehensive survey on mechanisms, effects, mathematical modeling, cleaning methods, and monitoring systems [J]. Solar Energy, 2024, 268: 112300.

[2]

王银，高瑞泽，李茂环，等．基于YOLOv5 LiteX算法的光伏组件缺陷检测[J]．太阳能学报，2023，44（9）：101-108.

Wang Yin, Gao Ruize, Li Maohuan, et al. Defect detection of photovoltaic modules based on YOLOv5 LiteX algorithm[J]. Acta Energiae Solaris Sinica, 2023, 44(9): 101-108.

[3]

艾上美，周剑峰，张必朝，等．基于改进SSD算法的光伏组件缺陷检测研究[J]．智慧电力，2023，51（12）：53-58.

Ai Shangmei, Zhou Jianfeng, Zhang Bichao, et al. Defect detection of photovoltaic modules based on improved SSD algorithm [J]. Smart Power, 2023, 51(12): 53-58.

[4]

彭雪玲，林珊玲，林志贤，等．改进的YOLOv5s太阳能电池片缺陷检测算法[J]．液晶与显示，2024，39（2）：237-247.

Peng Xueling, Lin Shanling, Lin Zhixian, et al. Improved YOLOv5s defect detection algorithm for solar cells [J]. Chinese Journal of Liquid Crystals and Displays, 2024, 39(2): 237-247.

[5]

刘枭雄，郑茜颖．基于深度学习的光伏板缺陷分类定位算法研究[J]．光电子技术，2024，44（1）：54-60.

Liu Xiaoxiong, Zheng Qianying. Research on photovoltaic panel defect classification and localization algorithm based on deep learning [J]. Optoelectronic Technology, 2024, 44(1): 54-60.

[6]

Han H, Suh J. Detection of photovoltaic panel dust using drone shooting imagery and deep learning techniques [J]. Journal of the Korean Society of Mineral and Energy Resources Engineers, 2024, 61(1): 23-32.

[7]

崔建伟，王月明．基于改进YOLOv8的光伏组件缺陷检测研究[J]．太阳能学报，2025，46（10）：189-196.

Cui Jianwei, Wang Yueming. Research on defect detection of photovoltaic modules based on improved YOLOv8 [J]. Acta Energiae Solaris Sinica, 2025, 46(10): 189-196.

[8]

Michail A, Livera A, Tziolis G, et al. A comprehensive review of unmanned aerial vehicle-based approaches to support photovoltaic plant diagnosis [J]. Heliyon, 2024, 10(1).

[9]

Xie C, Li Q, Yang Y, et al. Research on a photovoltaic panel dust detection algorithm based on 3D data generation[J]. Energies, 2024, 17(20): 5222.

[10]

Liu L, Li Q, Liao X, et al. Bird droppings defects detection in photovoltaic modules based on CA-YOLOv5[J]. Processes, 2024, 12(6): 1248.

[11]

Alatwi A M, Albalawi H, Wadood A, et al. Deep learning-based dust detection on solar panels: a low-cost sustainable solution for increased solar power generation[J]. Sustainability, 2024, 16(19): 8664.

[12]

Naeem U, Chadda K, Vahaji S, et al. Aerial imaging-based soiling detection system for solar photovoltaic panel cleanliness inspection [J]. Sensors, 2025, 25(3): 738.

[13]

Khanam R, Asghar T, Hussain M. Comparative performance evaluation of YOLOv5, YOLOv8, and YOLOv11 for solar panel defect detection [C]// Solar, MDPI, 2025, 5(1): 6.

[14]

Evstatiev B I, Trifonov D T, Gabrovska-Evstatieva K G, et al. PV module soiling detection using visible spectrum imaging and machine learning [J]. Energies, 2024, 17(20): 5238.

[15]

Masita K, Hasan A, Shongwe T, et al. Deep learning in defects detection of PV modules: a review [J]. Solar Energy Advances, 2025, 5: 100090.

[16]

Mittal P. A comprehensive survey of deep learning-based lightweight object detection models for edge devices [J]. Artificial Intelligence Review, 2024, 57(9): 242.

[17]

Zhang J, Chen X, Wei H, et al. A lightweight network for photovoltaic cell defect detection in electroluminescence images based on neural architecture search and knowledge distillation [J]. Applied Energy, 2024, 355: 122184.

[18]

Chen C, Chen Z, Li H, et al. Research on defect detection in lightweight photovoltaic cells using YOLOv8-FSD [J]. Sensors, 2025, 25(3): 843.

[19]

Ge Y, Li Z, Meng L. YOLO-MSD: a robust industrial surface defect detection model via multi-scale feature fusion[J]. Applied Intelligence, 2025, 55: 840.

[20]

Yuan J, Zhao L, Ye W, et al. MSFF-YOLO: A multi-scale feature fusion network for aero-engine blades surface defects detection [J]. Journal of Nondestructive Evaluation, 2026, 45(1): 30.

[21]

Brydon Z, Lee K, Hassani A. An automated framework for drone-based solar panel soiling detection [C]// 2023 8th International Conference on Robotics and Automation Engineering (ICRAE), IEEE, 2023: 203-210.

[22]

Ahmed S, Rashid H, Qadir Z, et al. Deep learning-based recognition and classification of soiled photovoltaic modules using HALCON software for solar cleaning robots [J]. Sensors, 2025, 25(5): 1295.

[23]

Rahman A. Solar panel surface defect and dust detection: Deep learning approach [J]. Journal of Imaging, 2025, 11(9): 287.

[24]

Fang M, Qian W, Qian T, et al. DGImNet: a deep learning model for photovoltaic soiling loss estimation [J]. Applied Energy, 2024, 376: 124335.

[25]

Munira S, Ahshan R, Abid M S, et al. Solar panel dust detection using deep learning model [C]// 2024 IEEE Sustainable Power and Energy Conference, IEEE, 2024: 379-384.

[26]

孙守鑫，孙一鸣，张峰，等．基于无人机图像和MSA-TCN模型的山区烟株计数[J]．农业工程学报，2025，41（21）：146-154.

Sun Shouxin, Sun Yiming, Zhang Feng, et al. Counting tobacco plants in mountainous areas based on UAV images and MSA-TCN model [J]. Transactions of the Chinese Society of Agricultural Engineering, 2025, 41 (21): 146-154.

[27]

Kotthapalli M, Ravipati D, Bhatia R. Yolov1 to Yolov11: a comprehensive survey of real-time object detection innovations and challenges [J]. arXiv preprint arXiv: 2508.02067, 2025.

[28]

Lin T Y, Dollár P, Girshick R, et al. Feature pyramid networks for object detection [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 2117-2125.

[29]

Liu S, Huang D, Wang Y. Learning spatial fusion for single-shot object detection [J]. arXiv preprint arXiv: 1911.09516, 2019..

[30]

Zhang S, Chi C, Yao Y, et al. Bridging the gap between anchor-based and anchor-free detection via adaptive training sample selection[C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 9759-9768.

[31]

Ronneberger O, Fischer P, Brox T. U-net: convolutional networks for biomedical image segmentation [C]//International Conference on Medical image computing and computer-assisted intervention, Cham: Springer international publishing, 2015: 234-241.

[32]

Wang J, Sun K, Cheng T, et al. Deep high-resolution representation learning for visual recognition [J]. IEEE transactions on pattern analysis and machine intelligence, 2020, 43(10): 3349-3364.

[33]

肖张娜，罗陆锋，陈明猷，等．基于改进YOLO-v4的果园环境下葡萄检测[J]．智能化农业装备学报（中英文），2023，4（2）：35-43.

Xiao Zhangna, Luo Lufeng, Chen Mingyou, et al. Grape detection in orchard environment based on improved YOLO-v4 [J]. Journal of Intelligent Agricultural Equipment, 2023, 4(2): 35-43.

[34]

马盼，杨子恒，万虎，等．基于YOLOv8网络的棉蚜图像识别算法及软件系统设计[J]．智能化农业装备学报（中英文），2023，4（3）：42-49.

Ma Pan, Yang Ziheng, Wan Hu, et al. Design of cotton aphid image recognition algorithm and software system based on YOLOv8 network [J]. Journal of Intelligent Agricultural Equipment, 2023, 4(3): 42-49.

[35]

魏堂伟，张津诚，王晶，等．基于改进YOLOv7的茶叶嫩芽识别模型研究[J]．智能化农业装备学报（中英文），2024，5（2）：42-50.

Wei Tangwei, Zhang Jincheng, Wang Jing, et al. Research on tea bud recognition model based on improved YOLOv7 [J]. Journal of Intelligent Agricultural Equipment, 2024, 5(2): 42-50.

2026年第7卷第2期

预览PDF | 下载PDF

引用本文

BibTeX

文章信息

doi: 10.12398/j.issn.2096-7217.2026.02.006

接收时间：2026-04-20
首发时间：2026-11-05
出版时间：2026-05-15

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2026-04-20
修回日期：2026-05-08

基金

2025 Chongqing Graduate Scientific Research and Innovation Project(CYS25872)

2025年重庆市研究生科研创新项目(CYS25872)

Yongchuan District “Unveiling the Commander” Project of Chongqing(2025yc-jbgs20005)

重庆市永川区“揭榜挂帅”项目(2025yc-jbgs20005)

2025 Graduate Scientific Research and Innovation Project of Chongqing University of Arts and Sciences

2025年重庆文理学院校级研究生科研创新项目

作者信息

^1.重庆文理学院，重庆市，402160

^2.丘陵山地智能农机装备重庆市高校工程研究中心，重庆市，402160

^3.重疆无人机科技（重庆）有限公司，重庆市，402160

通讯作者:

^* 黎斌，男，1991年生，重庆綦江人，博士，副教授；研究方向为山地智能农业装备及机器人。E-mail：cncqlibin@163.com

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/znhnyzbxb/CN/10.12398/j.issn.2096-7217.2026.02.006

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

BibTeX
EndNote
RefWorks
TxT