改进YOLOv10的天台光伏涉电区域行为人预警算法

doi:10.3778/j.issn.1002-8331.2407-0519

摘要/Abstract

摘要： 随着分布式光伏接入规模的日益增大，安装于私人用户天台区域的光伏电池板及其附属发电系统可能对用户带来触电风险。针对区域监控设备算力受限和天台涉电范围内行为人越界预警问题，讨论采用一种改进YOLOv10、DeepSort和PNPOLY的天台光伏板区域行为人跟踪预警算法。基于YOLOv10重新设计了特征融合模块，模块引入部分卷积思想，解决了特征融合部分运算量大问题。同时针对受光伏电池板遮挡的行为人检测，引入了针对遮挡目标的损失函数（Repulsion-IoU损失函数），并在渐进特征融合结构AFPN（asymptotic feature pyramid network for object detection）基础上设计了更好融合多层次目标的三层特征提取结构AFPN-3。在改进目标检测的基础上，利用DeepSort实现对行为人的持续跟踪定位，使用轻量化分类网络Fasternet替代原本特征提取网络，有效降低了模型的体积，提高了行为人跟踪的质量。为了解决行为人越界判定问题，采用PNPOLY算法，使用检测行为人脚部坐标位置进行精确的越界判别。实验表明，与YOLOv10n相比，在仅损失检测精度0.3个百分点的前提下，改进后行为人检测模型运算量下降了18%，参数量下降了19%，结合改进后的DeepSort模型，平均跟踪精度较之前提升1.4个百分点，模型大小只有8.7 MB。真实场景实验中，所提算法天台光伏涉电区域行为人预警正确率达到94%，具有轻量化、精确度高的特点，能满足天台光伏板区域行为人跟踪预警的实际检测需求。

关键词: 天台光伏电池板涉电区域, YOLOv10, DeepSort, 行为人跟踪, 预警检测

Abstract: With the increasing scale of distributed photovoltaic (PV) installations, PV panels and their associated power generation systems installed on private rooftops may pose an electric shock hazard to users. To address the issues of boundary crossing warnings and the limited computational capacity of regional monitoring equipment, this paper proposes an improved tracking and warning algorithm based on YOLOv10, DeepSort, and PNPOLY. To reduce the high computational complexity in feature fusion, the feature fusion module in YOLOv10 is redesigned using the concept of partial convolution. Furthermore, to enhance the detection of individuals occluded by PV panels, the Repulsion-IoU loss function is introduced, and a three-layer feature extraction structure (AFPN-3) is designed based on the asymptotic feature pyramid network for object detection (AFPN) to better integrate multi-level targets. On this basis, continuous individual tracking is performed using DeepSort, and the lightweight Fasternet is used to replace the original feature extraction network, reducing the model size and improving tracking quality. To accurately determine boundary crossing, the PNPOLY algorithm is used to detect the coordinates of individuals’feet. Experimental results show that, compared to YOLOv10n, the improved model reduces computational complexity by 18%, reduces the number of parameters by 19%, and only loses 0.3 percentage points in detection accuracy. The improved DeepSort model increases the average tracking accuracy by 1.4 percentage points and reduces the model size to 8.7 MB. The proposed algorithm achieves a 94% accuracy rate in warning individuals in electrified rooftop PV areas, demonstrating its lightweight and high-precision characteristics, meeting the practical needs for tracking and warning individuals near rooftop PV panels.

Key words: rooftop solar photovoltaic platform, YOLOv10, DeepSort, pedestrian tracking, early warning and detection

李军, 房志远, 周昊星. 改进YOLOv10的天台光伏涉电区域行为人预警算法[J]. 计算机工程与应用, 2025, 61(5): 211-221.

LI Jun, FANG Zhiyuan, ZHOU Haoxing. Improvement of Early Warning Algorithm of YOLOv10 for Actors in Rooftop Photovoltaic Power-Related Area[J]. Computer Engineering and Applications, 2025, 61(5): 211-221.

参考文献

[1] 尹泽中, 李功权. 三维场叠加视频流的电子围栏越界检测方法[J]. 测绘通报, 2024(6): 103-108.
YIN Z Z, LI G Q. Electronic fence out-of-bounds detection method based on 3D field superimposed video stream[J]. Bulletin of Surveying and Maping, 2024(6): 103-108.
[2] 夏长庚, 金贵红, 石家德. 便携式电子围栏在变电站作业区域周界应用[J]. 电子技术与软件工程, 2019(24): 203-204.
XIA C G, JIN G H, SHI J D, et al. Portable electronic fence is applied in the perimeter of the substation operation area[J].
Electronic Technology & Software Engineering, 2019(24): 203-204.
[3] 李其元. 基于STM32的张力式电子围栏的软硬件开发 [D]. 绵阳: 西南科技大学, 2017.
LI Q Y. The hardware and software development of tension type electronic fence base on chip STM32[D]. Mianyang: Southwest University of Science and Technology, 2017.
[4] REDMON J, FARHADI A. YOLOv3: an incremental improvement[J]. arXiv:1804.02767, 2018.
[5] LYU S, LI R, ZHAO Y, et al. Green citrus detection and counting in orchards based on YOLOv5-CS and AI edge system[J]. Sensors, 2022, 22(2): 576.
[6] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, 2016: 779-788.
[7] REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[8] WOJKE N, BEWLEY A, PAULUS D. Simple online and real-time tracking with & a deep association metric [C]//Proceedings of the 2017 IEEE International Conference on Image Processing, 2017: 3645-3649.
[9] DU Y H, ZHAO Z C, SONG Y, et al. StrongSORT: make deepsort great again[J]. IEEE Transactions on Multimedia, 2023, 25: 8725-8737.
[10] 陈柳, 陈明举, 薛智爽, 等. 轻量化高精度卷积神经网络的安全帽识别方法[J]. 计算机工程与应用, 2021, 57(22): 177-181.
CHEN L, CHEN M J, XUE Z S, et al. Lightweight and high-precision convolutional neural network for helmet recognition method[J]. Computer Engineering and Applications, 2021, 57(22): 177-181.
[11] 何湘杰, 宋晓宁. YOLOv4-Tiny的改进轻量级目标检测算法[J]. 计算机科学与探索, 2024, 18(1): 138-150.
HE X J, SONG X N. Improved YOLOv4-Tiny lightweight target detection algorithm[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(1): 138-150.
[12] 王春梅, 刘欢. YOLOv8-VSC: 一种轻量级的带钢表面缺陷检测算法[J]. 计算机科学与探索, 2024, 18(1): 151-160.
WANG C M, LIU H. YOLOv8-VSC: lightweight algorithm for strip surface defect detection[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(1): 151-160.
[13] ZHOU J J, ZHANG B H, YUAN X L, et al. YOLO-CIR: the network based on YOLO and ConvNeXt for infrared object detection[J]. Infrared Physics and Technology, 2023: 104703.
[14] 郑晓, 王淑琴, 张文聪, 等. 基于深度学习的安全帽监管系统[J]. 计算机系统应用, 2021, 30(11): 118-126.
ZHENG X, WANG S Q, ZHANG W C, et al. Safety helmet supervision system based on deep learning[J]. Computer Systems & Applications, 2021, 30(11): 118-126.
[15] BOCHKOVSKIY A, WANG C, LIAO H M. YOLOv4: optimal speed and accuracy of object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 883-892.
[16] 杨学杰, 李思毛, 李建业, 等. 面向巡检机器人的电力设备状态检测算法研究[J]. 计算机技术与发展, 2021, 31(3): 201-205.
YANG X J, LI S M, LI J Y, et al. Research on algorithm for equipment condition monitoring based on inspection robot[J]. Computer Technology and Development, 2021, 31(3): 201-205.
[17] 丁冰, 杨祖莨, 丁洁, 等. 基于改进YOLOv3的高速公路隧道内停车检测方法[J]. 计算机工程与应用, 2021, 57(23): 234-239.
DING B, YANG Z L, DING J, et al. Improved YOLOv3-based parking detection method in highway tunnels[J]. Computer Engineering and Applications, 2021, 57(23): 234 -239.
[18] 葛雯, 姜添元. 改进YOLO与Deepsort检测跟踪算法的研究[J]. 计算机仿真, 2022, 39(5): 186-190.
GE W, JIANG T Y. Research on improved YOLO and DeepSort detection and tracking algorithm[J]. Computer Simulation, 2022, 39(5): 186-190.
[19] 韩晓冰, 王雨田, 黄综浏, 等. 改进YOLOv5+DeepSort的行人跟踪算法[J]. 现代电子技术, 2023, 46(7): 33-38.
HAN X B, WANG Y T, HUANG Z L, et al. Improved YOLOv5+DeepSort pedestrian tracking algorithm[J]. Modern Electronic Technology, 2023, 46(7): 33-38.
[20] 房凯. 基于深度学习的围栏跨越行为检测方法[J]. 计算机系统应用, 2021 , 30(2): 147-153.
FANG K. Deep learning-based detection method of fence crossing action[J]. Computer Systems & Applications, 2021, 30(2): 147-153.
[21] 丁俊峰, 肖文韬, 李明远, 等. 融合视觉感知与RTK定位的变电站越界违章检测[J]. 计算机技术与发展, 2023, 33(8): 206-213.
DING J F, XIAO W T, LI M Y, et al. Research on out-of-bounds detection integrating visual perception and RTK positioning in substations[J]. Computer Technology and Development, 2023, 33(8): 206-213.
[22] WANG A, CHEN H, LIU L H, et al. YOLOv10: real-time end-to- end object detection[J]. arXiv:2405.14458, 2024.
[23] CHEN J R, KAO S H, HE H, et al. Run, don’t walk: chasing higher FLOPS for faster neural networks[C]//Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: 12021-12031.
[24] YANG G Y, LEI J Z, ZHU Z K, et al. AFPN: asymptotic feature pyramid network for object detection[C]//Proceedings of the 2023 IEEE Conference on Systems, Man, and Cybernetics, 2023: 2184-2189.
[25] WANG X L, XIAO T T, JIANG Y N, et al. Repulsion loss: detecting pedestrians in a crowd[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018: 7774-7783.
[26] WIBOWO A T, HARIADI M, SULISTYNON T M Y, et al. Mapping of tourism destinations by GIS-blockchain RPCA based on PNPOLY algorithm[C]//Proceedings of the 2020 International Seminar on Application for Technology of Information and Communication, 2020: 412-416.
[27] ZHAO Y L, LI Y L, WANG S J, et al. Asymmetric deep hashing for person re-identifications[J]. Tsinghua Science and Techology, 2022, 27(2): 396-411.