基于改进RT-DETR的光伏板缺陷检测

doi:10.16088/j.issn.1001-6600.2025060302

摘要/Abstract

摘要： 为了解决现有传统的光伏板缺陷检测精度低、模型参数量大以及复杂背景检测时出现漏检和误检的问题,本文基于RT-DETR模型提出一种高效的光伏板缺陷检测算法。首先,为了提升检测精度,采用FREBlock主干增强特征提取能力同时还能提高检测效率。其次,设计CRDFP多尺度特征融合结构进一步增强特征融合能力。最后,引入可变形注意力机制DAttention,使模型能专注于相关区域的信息特征。实验结果表明,改进后的模型平均类别精度(η_mAP)效果达到79.2%,较传统模型提高3.6个百分点,参数量减少22.6%,运算量降低25.9%,表现出较高的实时检测能力。

关键词: 深度学习, RT-DETR, 光伏板, 缺陷检测, 多尺度特征融合

Abstract: In order to address the issues of low accuracy, large model parameters, and the occurrences of missed detections and false detections under complex backgrounds in the existing traditional photovoltaic panel defect detection, this paper proposes an efficient photovoltaic panel defect detection algorithm based on the RT-DETR model. Firstly, to boost detection accuracy, the FREBlock architecture is developed, which not only improves feature extraction but also enhances detection efficiency. Secondly, the CRDFP multi-scale feature fusion structure is designed to strengthen the integration of features across different scales. Lastly, the deformable attention mechanism is incorporated, which enables the model to focus on the information features of the region of interest. The experimental results indicate that the improved model achieves an average mean Average Precision (mAP) of 79.2%, an increase of 3.6 percentage points over the traditional model. Additionally, the model's parameters reduces 22.6%, and the computational load decreases 25.9%, demonstrating a high capacity for real-time detection.

Key words: deep learning, RT-DETR, photovoltaic panels, defect detection, multi-scale feature fusion

中图分类号: TM615;TP391.41

吕辉, 司可. 基于改进RT-DETR的光伏板缺陷检测[J]. 广西师范大学学报（自然科学版）, 2026, 44(2): 52-64.

LÜ Hui, SI Ke. Photovoltaic Panel Defect Detection Based on Improved RT-DETR[J]. Journal of Guangxi Normal University(Natural Science Edition), 2026, 44(2): 52-64.

参考文献

[1] 王文静, 王斯成.我国分布式光伏发电的现状与展望[J].中国科学院院刊, 2016, 31(2):165-172.DOI:10.16418/j.issn.1000-3045.2016.02.002.
[2] SU B Y, CHEN H Y, ZHOU Z.BAF-detector:an efficient CNN-based detector for photovoltaic cell defect detection[J].IEEE Transactions on Industrial Electronics, 2022, 69(3):3161-3171.DOI:10.1109/TIE.2021.3070507.
[3] 吴涛, 赖菲.基于LeNet-5模型的太阳能电池板缺陷识别分类[J].热力发电, 2019, 48(3):120-125.DOI:10.19666/j.rlfd.201807147.
[4] GUERRIERO P, DI NAPOLI F, VALLONE G, et al.Monitoring and diagnostics of PV plants by a wireless self-powered sensor for individual panels[J].IEEE Journal of Photovoltaics, 2015, 6(1):286-294.DOI:10.1109/JPHOTOV.2015.2484961.
[5] REDMON J, FARHADI A.YOLOv3:an incremental improvement[EB/OL].(2018-04-08)[2025-09-08].https://ui.adsabs.harvard.edu/abs/2018arXiv180402767R.DOI:10.48550/arXiv.1804.02767.
[6] LI C, LI L, JIANG H, et al.YOLOv6:a single-stage object detection framework for industrial applications[EB/OL].(2022-09-07)[2025-09-08].https://ui.adsabs.harvard.edu/abs/2022arXiv220902976L.DOI:10.48550/arXiv.2209.02976.
[7] WANG C Y, BOCHKOVSKIY A, LIAO H M.YOLOv7:trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Piscataway,NJ:IEEE, 2023:7464-7475.DOI:10.1109/CVPR52729.2023.00721.
[8] REIS D, KUPEC J, HONG J, et al.Real-time flying object detection with YOLOv8[EB/OL].(2024-05-22)[2025-09-08].https://ui.adsabs.harvard.edu/abs/2023arXiv230509972R.DOI:10.48550/arXiv.2305.09972.
[9] 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.DOI:10.1109/TPAMI.2016.2577031.
[10] ZHANG T Y, CHEN C Y, LIU Y, et al.PSRR-MaxpoolNMS++:fast non-maximum suppression with discretization and pooling[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025, 47(2):978-993.DOI:10.1109/TPAMI.2024.3485898.
[11] ALSHEHRI M, OUADOU A, SCOTT G J.Deep transformer-based network deforestation detection in the Brazilian Amazon using sentinel-2 imagery[J].IEEE Geoscience and Remote Sensing Letters, 2024, 21:2502705.
[12] CARION N, MASSA F, SYNNAEVE G, et al.End-to-end object detection with transformers[C]//Computer Vision- ECCV 2020.Cham:Springer, 2020:213-229.DOI:10.1007/978-3-030-58452-8_13.
[13] ZHU X, SU W, LU L, et al.Deformable DETR:deformable transformers for end-to-end object Detection[EB/OL].(2021-03-18)[2025-09-08].https://ui.adsabs.harvard.edu/abs/2020arXiv201004159Z.DOI:10.48550/arXiv.2010.04159.
[14] SHEHZADI T, HASHMI K A, STRICKER D, et al.Sparse semi-DETR:sparse learnable queries for semi-supervised object detection[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Piscataway,NJ:IEEE, 2024:5840-5850.DOI:10.1109/CVPR52733.2024.00558.
[15] ZHANG H, LI F, LIU S, et al.DINO:DETR with improved denoising anchor boxes for end-to-end object detection[EB/OL].(2022-07-11)[2025-09-08].https://ui.adsabs.harvard.edu/abs/2022arXiv220303605Z.DOI:10.48550/arXiv.2203.03605.
[16] ZHAO Y A, LV W Y, XU S L, et al.DETRs beat YOLOs on real-time object detection[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Piscataway,NJ:IEEE, 2024:16965-16974.DOI:10.1109/CVPR52733.2024.01605.
[17] 苏俊, 唐潮龙, 刘智权, 等.基于全局与局部特征提取增强的光伏板缺陷检测算法[J].计算机工程与应用, 2025, 61(12):299-310.DOI:10.3778/j.issn.1002-8331.2412-0035.
[18] CHEN J R, KAO S H, HE H, et al.Run, don’t walk:chasing higher FLOPS for faster neural networks[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Piscataway,NJ:IEEE, 2023:12021-12031.DOI:10.1109/CVPR52729.2023.01157.
[19] DING X H, ZHANG X Y, MA N N, et al.RepVGG:making VGG-style ConvNets great again[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Piscataway,NJ:IEEE, 2021:13728-13737.DOI:10.1109/CVPR46437.2021.01352.
[20] CHINTHAKA D, FERNANDO S.Multi-scale efficient attention net for person re-identification[C]//2024 International Conference on Advances in Technology and Computing (ICATC).Piscataway,NJ:IEEE, 2024:1-6.DOI:10.1109/ICATC 64549.2024.11025307.
[21] XIA Z F, PAN X R, SONG S J, et al.Vision transformer with deformable attention[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Piscataway,NJ:IEEE, 2022:4784-4793..DOI:10.1109/cvpr52688.2022.00475.
[22] NI Z L, CHEN X H, ZHAI Y J, et al.Context-guided spatial feature reconstruction for efficient semantic segmentation[C]//Computer Vision-ECCV 2024.Cham:Springer, 2025:239-255.DOI:10.1007/978-3-031-72943-0_14.
[23] 刘沅林, 金建辉, 李盛洪, 等.PV-HSD-2025:面向光伏热斑检测的红外基准数据集与YOLOv8-P1增强架构[J/OL].计算机工程与应用:1-19.[2025-09-08].https://kns.cnki.net/KCMS/detail/detail.aspx?filename=JSGG2025041100E& dbname=CJFD&dbcode=CJFQ.

相关文章 15

[1]	田晟, 冯帅涛, 李嘉. 一种基于复合框架的城市道路场景车辆轨迹提取方法[J]. 广西师范大学学报（自然科学版）, 2026, 44(2): 31-51.
[2]	王旭阳, 梁宇航. 多尺度非对称注意力遥感去雾Transformer[J]. 广西师范大学学报（自然科学版）, 2026, 44(2): 77-89.
[3]	张胜伟, 曹洁. 融合傅里叶卷积与差异感知的钢材表面微小缺陷检测算法[J]. 广西师范大学学报（自然科学版）, 2026, 44(2): 90-102.
[4]	罗缘, 朱文忠, 王文, 吴宇浩. 基于改进PatchTST的多步水质预测模型[J]. 广西师范大学学报（自然科学版）, 2026, 44(2): 115-131.
[5]	韩华彬, 高丙朋, 蔡鑫, 孙凯. 基于HO-CNN-BiLSTM-Transformer模型的风机叶片结冰故障诊断[J]. 广西师范大学学报（自然科学版）, 2025, 43(6): 13-28.
[6]	魏梓书, 陈志刚, 王衍学, 哈斯铁尔·马德提汗. 基于SBSI-YOLO11的轻量化轴承外观缺陷检测算法[J]. 广西师范大学学报（自然科学版）, 2025, 43(6): 80-91.
[7]	刘廷汉, 梁艳, 黄鹏升, 闭金杰, 黄守麟, 李廷会. 基于改进YOLOv8s的人脸痤疮小目标检测[J]. 广西师范大学学报（自然科学版）, 2025, 43(5): 114-129.
[8]	黎宗孝, 张健, 罗鑫悦, 赵嶷飞, 卢飞. 基于K-means和Adam-LSTM的机场进场航迹预测研究[J]. 广西师范大学学报（自然科学版）, 2025, 43(4): 15-23.
[9]	石天怡, 南新元, 郭翔羽, 赵濮, 蔡鑫. 基于改进ConvNeXt的苹果叶片病害分类算法[J]. 广西师范大学学报（自然科学版）, 2025, 43(4): 83-96.
[10]	梁胤杰, 南新元, 蔡鑫, 李云鹏, 勾海光. 基于数据增广与改进YOLOv8的桥梁缺陷检测[J]. 广西师范大学学报（自然科学版）, 2025, 43(3): 84-97.
[11]	苏春海, 夏海英. 抗噪声双约束网络的面部表情识别[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 70-82.
[12]	戴林华, 黎远松, 石睿, 何忠良, 李雷. HSED-YOLO:一种轻量化的带钢表面缺陷检测模型[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 95-106.
[13]	李欣, 宁静. 基于时空特征融合的电力系统暂态稳定评估[J]. 广西师范大学学报（自然科学版）, 2024, 42(6): 89-100.
[14]	侯海燕, 谭玉枚, 宋树祥, 夏海英. 头部姿态鲁棒的面部表情识别[J]. 广西师范大学学报（自然科学版）, 2024, 42(6): 126-137.
[15]	卢家辉, 陈庆锋, 王文广, 余谦, 何乃旭, 韩宗钊. 基于多尺度注意力的器官图像分割方法[J]. 广西师范大学学报（自然科学版）, 2024, 42(6): 138-148.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed