智能化农业装备学报（中英文）

基于激光散斑STCD算法的种子活力分类方法研究

预览PDF | 下载PDF

门森 ¹ , 李潇鑫 ¹ , 张君豪 ¹ , 刘玥 ¹ , 刘冰冰 ¹^,²^,^*^,^*

智能化农业装备学报（中英文） | 2026,7(1): 75-85

收起

智能化农业装备学报（中英文） | 2026 , 7 (1) : 75 -85

基于激光散斑STCD算法的种子活力分类方法研究

全屏

门森¹, 李潇鑫¹, 张君豪¹, 刘玥¹, 刘冰冰¹^,²^,^*^,^*

作者信息

^1.北京联合大学机器人学院，北京市，100020

^2.北京市信息服务工程重点实验室，北京市，100020

通讯作者:

^* 刘冰冰，女，1985年生，吉林梨树人，副教授，博士；研究方向为机器人技术，机器视觉识别。E-mail：lynnbit@126.com

作者简介:

门森，男，1989年生，黑龙江牡丹江人，讲师，博士；研究方向为计算机视觉、模式识别。E-mail：mensen1989@163.com

Research on seed viability classification method based on laser speckle STCD algorithm

Sen MEN¹, Xiaoxin LI¹, Junhao ZHANG¹, Yue LIU¹, Bingbing LIU¹^,²

Affiliations

^1.Robotics College, Beijing Union University, Beijing 100020, China

^2.Beijing Key Laboratory of Information Service Engineering, Beijing 100020, China

Corresponding Author:

^* LIU Bingbing, E-mail: lynnbit@126.com

MEN Sen, E-mail: mensen1989@163.com

出版时间: 2026-02-15 doi: 10.12398/j.issn.2096-7217.2026.01.008

文章导航

摘要

收起

为解决传统种子活力检测方法速度慢，耗时长，成本高的问题，提出了一种将激光散斑技术、STCD和深度学习模型相结合的种子活力无损检测方法，构建了“图像采集-特征处理-模型训练-活力判定”一体化种子活力检测系统。采用632.8 nm He-Ne激光器、CCD相机搭建成像系统，获取玉米种子激光散斑图像，经过图像预处理，将单个玉米种子区域提取出来，后采用二维小波变换进行散斑图像去噪，之后利用STCD进行图像数据处理，将STCD与图像差分改进策略深度融合，将玉米种子激光散斑图像的特征信息凸显出来，以前4帧加权平均为参考帧，引入3×3邻域局部方差加权机制与遗忘因子α=0.7构建累积差分图，最终得到1 600张差分图像，活种子与失活种子各800张，按7：2：1划分数据集。模型方面，采用MobileNetV3为骨干网络，构建轻量化Faster R-CNN模型，调整参数，优化训练策略至验证集准确率稳定。后用最优训练模型进行玉米种子散斑图像目标检测、特征提取与深度学习，最终完成种子分类回归，实现对种子活力的高效准确检测。结果表明，种子激光散斑图像经STCD处理后特征信息明显，差分图像信噪比最高达35.8 dB，经过轻量化Faster R-CNN模型训练后，活种子分类准确率达92.25%、失活种子分类准确率达93.25%，模型ROC曲线中AUC值98%，F1分数为92.63%，假阳性率为4.75%，单张图像识别时间为0.031 5 s，比原模型单张检测时间缩短了0.049 3 s，检测速度比原模型提高2.5倍，收敛轮数25，比原模型减少一半以上，训练效率高。证明该方法可以实现种子活力的快速无损检测。

关键词

激光散斑 / 时空耦合差分 / Faster R-CNN / 种子活力 / 无损检测

Abstract

收起

To overcome the challenges of slow speed, prolonged time consumption, and high costs in traditional seed vigor detection methods, we propose a non-destructive seed vigor detection approach that integrates laser speckle technology, a STCD algorithm, and a deep learning model, which facilitate the construction of a comprehensive seed vigor detection system encompassing “image acquisition, feature processing, model training, and vigor determination.” An imaging system was developed by utilizing a 632.8 nm He-Ne laser and a CCD camera to capture laser speckle images of pea seeds. After image preprocessing, the individual pea seed area was extracted. Then, two-dimensional wavelet transform was applied to denoise the speckle images. Subsequently, the STCD method was utilized for image data processing. The STCD method is deeply integrated with the image difference improvement strategy to highlight the feature information of the pea seed laser spot image. The weighted average of the first 4 frames is taken as the reference frame. The 3×3 neighborhood local variance weighting mechanism and the forgetting factor α=0.7 are introduced to construct the cumulative difference map. Finally, 1 600 difference images are obtained, with 800 viable seeds and 800 non-viable seeds. Divide the dataset in a ratio of 7∶2∶1. In terms of the model, MobileNetV3 was adopted as the backbone network to construct the lightweight Faster R-CNN model. The parameters were adjusted and the training strategy was optimized until the Accuracy of the validation set was stable. Subsequently, the optimally trained model was employed for target detection, feature extraction, and deep learning on the speckle images of maize seeds. Ultimately, seed classification and regression were accomplished, enabling efficient and accurate detection of seed viability. The results indicate that the laser speckle images of seeds processed by STCD exhibit distinct feature information, with a maximum signal-to-noise ratio of 35.8 dB for the differential images. After training with the lightweight Faster R-CNN model, the classification accuracy for viable seeds is 92.25% and the accuracy for non-viable seeds is 93.25%. The model achieves an AUC value of 98% in the ROC curve, an F1-score of 92.63%, a false positive rate (FPR) of 4.75%. The recognition time for a single image is 0.031 5 seconds, which is 0.049 3 seconds shorter than that of the original model. The detection speed is 2.5 times faster than the original model, and the convergence epoch is 25, reduced by more than half compared to the original model, demonstrating high training efficiency. These findings confirm that the proposed method enables rapid and non-destructive detection of seed viability.

Key words

laser speckle / spatiotemporal coupled difference / Faster R-CNN / seed vigor / non-destructive testing

引用本文

门森, 李潇鑫, 张君豪, 刘玥, 刘冰冰. 基于激光散斑STCD算法的种子活力分类方法研究. 智能化农业装备学报（中英文）, 2026 , 7 (1) : 75 -85 . DOI: 10.12398/j.issn.2096-7217.2026.01.008

Sen MEN, Xiaoxin LI, Junhao ZHANG, Yue LIU, Bingbing LIU. Research on seed viability classification method based on laser speckle STCD algorithm[J]. Journal of Intelligent Agricultural Mechanization, 2026 , 7 (1) : 75 -85 . DOI: 10.12398/j.issn.2096-7217.2026.01.008

基金

收起

国家自然科学基金(31770769)
北京市教育委员会科研计划项目(KM201911417008)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

[1]

王成，郝楠，李月明．种子活力的分子调控机制研究进展[J]．园艺与种苗，2022，42（4）：83-85，92.

WANG Cheng, HAO Nan, LI Yueming, et al. Research progress on the molecular regulation of seed vigor [J]. Horticulture & Seed, 2022, 42(4): 83-85, 92.

[2]

尹田振，彭彦昆，李永玉，等．便携式多品种花生种子活力无损检测装置研究[J]．农业机械学报，2024，55（S2）：340-347.

YIN Tianzhen, PENG Yankun, LI Yongyu, et al. Design of portable non-destructive device for viability assessment of multiple peanut seed varieties [J]. Transactions of the Chinese Society for Agricultural Machinery, 2024, 55(S2): 340-347.

[3]

杨冬风，李爱传，刘金明，等．耦合平均影响值-连续投影算法优化种子活力近红外检测模型[J]．光谱学与光谱分析，2022，42（10）：3135-3142.

YANG Dongfeng, LI Aichuan, LIU Jinming, et al. Optimization of seed vigor near-infrared detection by coupling mean impact value with successive projection algorithm [J]. Spectroscopy and Spectral Analysis, 2022, 42(10): 3135-3142.

[4]

任亚举，王瑞敏，薛冬，等．基于深度学习结合高光谱技术的大豆种子活力检测方法[J]．江苏农业学报，2025，41（5）：927-936.

REN Yaju, WANG Ruimin, XUE Dong, et al. Soybean seed vigor detection based on deep learning combined with hyper-spectral technology [J]. Jiangsu Journal of Agricultural Sciences, 2025, 41(5): 927-936.

[5]

李文君，雷荣，曾艳，等．基于近红外光谱技术的不同种皮色大豆种子活力预测模型的构建[J]．中国油料作物学报，2025，47（4）：1051-1064.

LI Wenjun, LEI Rong, ZENG Yan, et al. Construction of seed vigour prediction models for different seed coat soybean based on near-infrared spectroscopic techniques [J]. Chinese Journal of Oil Crop Sciences, 2025, 47(4): 1051-1064.

[6]

乔菊香，杨晓洪，姚宗泽，等．基于多光谱成像的颜色特征在水稻种子活力检测上的应用[J]．杂交水稻，2025，40（2）：15-23.

QIAO Juxiang, YANG Xiaohong, YAO Zongze, et al. Application of color features based on multispectral imaging in rice seedvigour detection [J]. Hybrid Rice, 2025, 40(2): 15-23.

[7]

吴静珠，李晓琪，刘翠玲，等．基于ATR的玉米种子活力太赫兹特征波段筛选方法[J]．农业机械学报，2020，51（4）：382-386，393.

WU Jingzhu, LI Xiaoqi, LIU Cuiling, et al. Screening method of characteristic thz region to corn seed vigor based on ATR [J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(4): 382-386, 393.

[8]

石睿，张晗，王成，等．高光谱图谱结合策略检测小麦单粒种子活力[J]．光谱学与光谱分析，2024，44（11）：3206-3212.

SHI Rui, ZHANG Han, WANG Cheng, et al. Detection of wheat single seed vigor using hyperspectral imaging and spectrum fusion strategy [J]. Spectroscopy and Spectral Analysis, 2024, 44(11): 3206-3212.

[9]

郭振宇，樊亚杉，翟保杰，等．光声光谱二氧化碳传感的种子活力检测[J]．光学精密工程，2025，33（3）：367-376.

GUO Zhenyu, FAN Yashan, ZHAI Baojie, et al. Detection of seed viability by photoacoustic carbon dioxide sensing [J]. Optics and Precision Engineering, 2025, 33(3): 367-376.

[10]

杜文玲，郭鹏，刘笑．基于高光谱小波特征的脱绒棉种活力检测[J]．农业工程学报，2024，40（20）：174-186.

DU Wenling, GUO Peng, LIU Xiao. Detection of delinted cotton seed vigor based on hyperspectralwavelet features [J]. Transactions of the Chinese Society of Agricultural Engineering, 2024, 40(20): 174-186.

[11]

李阳，李晓琪，杨佳滢，等．基于THz时域反射成像技术的玉米种子淀粉分布可视化研究[J]．光谱学与光谱分析，2023，43（9）：2722-2728.

LI Yang, LI Xiaoqi, YANG Jiaying, et al. Visualisation of starch distribution in corrn seeds based on terahertz time-domain spectral reflection imaging technology [J]. Spectroscopy and Spectral Analysis, 2023, 43(9): 2722-2728.

[12]

贺锋涛，刘思琪，张建磊，等．基于卷积神经网络的激光散斑水下湍流强度估计研究[J/OL]．中国激光，1-22[2025-08-25].

HE Fengtao, LIU Siqi, ZHANG Jianlei, et al. Estimation of laser speckle intensity in underwater turbulence based on convolutional neural network [J]. Chinese Journal of Lasers, 1-22[2025-08-25].

[13]

EVANS L E, GRAY A L, WALSH K R, et al. Combining in vivo two-photon and laser speckle microscopy with the ex vivo capillary-parenchymal arteriole preparation as a novel approach to study neurovascular coupling [J]. Microcirculation, 2025, 32(1): e70001.

[14]

GOLDBERG J, MILLER D R, DIMANCHE A, et al. Intraoperative laser speckle contrast imaging to assess vessel flow in neurosurgery: A pilot study [J]. Neurosurgery, 2024, 94(5): 983-992.

[15]

GOLUBOVA N, RYZHKOV I, LAPIN K, et al. Effect of thinned-skull cranial window on monitoring cerebral blood flow using laser speckle contrast imaging [J]. IEEE Journal of Selected Topics in Quantum Electronics, 2025(4): 31.

[16]

李鹤群，郑予菲，杨涵夕，等．基于激光散斑成像的煤矸特征提取与识别方法[J]．光学学报，2024，44（21）：214-223.

LI Hequn, ZHENG Yufei, YANG Hanxi, et al. Feature extraction and recognition method of coal and gangue based onlaser speckle imaging [J]. Acta Optica Sinica, 2024, 44(21): 214-223.

[17]

SONG J, LIANG R, YUAN B, et al. Dimo-CNN: Deep learning toolkit-accelerated analytical modeling and optimization of cnn hardware and dataflow [J]. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2024, 44(1): 251-265.

[18]

ZHANG X, LIU Q, GUI D, et al. Enhanced river connectivity assessment across larger areas through deep learning with dam detection [J]. Hydrological Processes, 2025, 39(1): e70063.

[19]

QUAN W, CHEN J, LIU Y, et al. Deep learning-based image and video inpainting: A survey [J]. International Journal of Computer Vision, 2024, 132(7): 2367-2400.

[20]

杜艳秋，李鑫，康辉，等．基于深度学习的木板激光散斑图像分类研究[J/OL]．重庆邮电大学学报（自然科学版），1-9[2025-08-21].

DU Yanqiu, LI Xin, KANG Hui, et al. Classification of wood laser speckle images based on deep learning [J]. Journal of Chongqing University of Posts and Telecommunications (Natural Science Edition), 1-9[2025-08-21].

[21]

杨孟杰，毛鹏军，王俊，等．基于改进YOLOv5s的椰子轻量化检测方法[J/OL]．农机化研究，1-9[2025-08-21].

YANG Mengjie, MAO Pengjun, WANG Jun, et al. Coconut lightweight detection method based on improved YOLOv5s [J]. Journal of Agricultural Mechanization Research, 1-9[2025-08-21].

[22]

王德成，黄晨，尤泳，等．基于等离子体处理的苜蓿种子活力综合评价[J]．农业工程学报，2024，40（22）：234-242.

WANG Decheng, HUANG Chen, YOU Yong, et al. Comprehensive evaluation of alfalfa (Medicago sativa L.) seed vitalitybased on plasma treatment [J]. Transactions of the Chinese Society of Agricultural Engineering, 2024, 40(22): 234-242.

[23]

吴静珠，李晓琪，刘翠玲，等．基于THz-TDS反射成像技术的玉米种子活力无损检测研究[J]．光谱学与光谱分析，2020，40（9）：2840-2844.

WU Jingzhu, LI Xiaoqi, LIU Cuiling, et al. Research on nondestructive testing of corn seed vigor based on THz-TDS reflection imaging [J]. Spectroscopy and Spectral Analysis, 2020, 40(9): 2840-2844.

[24]

王元红，杨志明，王琪，等．基于YOLOv5s-SPD的茶芽识别方法及识别系统光源设计与试验[J]．智能化农业装备学报（中英文），2024，5（3）：33-43.

WANG Yuanhong, YANG Zhiming, WANG Qi, et al. Design and experiment of tea bud recognition methodand light source of recognition system based on YOLOv5-SPD[J]. Journal of Intelligent Agricultural Mechanization, 2024, 5(3): 33-43.

[25]

吴擎，韦润轩，周乐，等．基于改进YOLOv5s的轻量化鲜茶叶识别方法[J]．智能化农业装备学报（中英文），2025，6（1）：1-14.

WU Qing, WEI Runxuan, ZHOU Le, et al. Lightweight fresh tea leaf recognition method based on improved YOLOv5s [J]. Journal of Intelligent Agricultural Mechanization, 2025, 6(1): 1-14.

[26]

陈俊霖，赵鹏，曹先林，等．基于通道剪枝的轻量化YOLOv8s草莓穴盘苗分级检测与定位方法[J]．智慧农业（中英文），2024，6（6）：132-143.

CHEN Junlin, ZHAO Peng, CAO Xianlin, et al. Lightweight YOLOv8s-Based strawberry plug seedling grading detection and localization via channel pruning [J]. Smart Agriculture, 2024, 6(6): 132-143.

[27]

肖孟宇，马云明，肖亦雄，等．基于改进VGG-16的烟叶成熟度识别[J]．智能化农业装备学报（中英文），2025，6（2）：79-87.

XIAO Mengyu, MA Yunming, XIAO Yixiong, et al. Fresh tobacco maturity recognition based on theimproved VGG-16 model [J]. Journal of Intelligent Agricultural Mechanization, 2025, 6(2): 79-87.

[28]

于合龙，周雷进雨，徐兴梅，等．激光农业研究进展和展望[J]．智能化农业装备学报（中英文），2024，5（3）：1-13.

YU Helong, ZHOU Leijinyu, XU Xingmei, et al. Advancements and prospects in laser agriculture [J]. Journal of Intelligent Agricultural Mechanization, 2024, 5(3): 1-13.

[29]

ABOOKASIS D, MOLCHO D, SHEMESH D, et al. Statistical assessment of the inner shape of a channel based on laser speckle contrast imaging [J]. Journal of Modern Optics, 2024, 71(7-8): 266-274.

[30]

邓林宵，杨雨桦，姚昞晖，等．多基色激光显示系统彩色散斑分布特性的研究[J]．激光技术，2023，47（6）：736-741.

DENG Linxiao, YANG Yuhua, YAO Binghui, et al. Distribution characteristics of color speckle in multi-primary laser display systems [J]. Laser Technology, 2023, 47(6): 736-741.

[31]

MANJUNATHAN A, KHABBA A, AMADID J, et al. Demand-oriented framework for image denoising using CNN and wavelet transform [C]// International Conference on Sensors and Microsystems. Singapore: Springer Nature Singapore, 2024: 1-15.

2026年第7卷第1期

预览PDF | 下载PDF

引用本文

BibTeX

文章信息

doi: 10.12398/j.issn.2096-7217.2026.01.008

接收时间：2025-11-10
首发时间：2026-11-03
出版时间：2026-02-15

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2025-11-10
修回日期：2025-12-29

基金

National Natural Science Foundation of China(31770769)

国家自然科学基金(31770769)

Scientific Research Plan Project of the Beijing Municipal Education Commission(KM201911417008)

北京市教育委员会科研计划项目(KM201911417008)

作者信息

^1.北京联合大学机器人学院，北京市，100020

^2.北京市信息服务工程重点实验室，北京市，100020

通讯作者:

^* 刘冰冰，女，1985年生，吉林梨树人，副教授，博士；研究方向为机器人技术，机器视觉识别。E-mail：lynnbit@126.com

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/znhnyzbxb/CN/10.12398/j.issn.2096-7217.2026.01.008

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