Insulators are crucial insulation components and structural supports in power grids, playing a vital role in the transmission lines. Due to temperature fluctuations, internal stress, or damage from hail, insulators are prone to injury. Automatic detection of damaged insulators faces challenges such as diverse types, small defect targets, and complex backgrounds and shapes. Most research for detecting insulator defects has focused on a single defect type or a specific material. However, the insulators in the grid's transmission lines have different colors and materials. Various insulator defects coexist, and the existing methods have difficulty meeting the practical application requirements. Current methods suffer from low detection accuracy and mAP0.5 cannot meet application requirements. This paper proposes an improved you only look once version 7 (YOLOv7) model for multi-type insulator defect detection. First, our model replaces the spatial pyramid pooling cross stage partial network (SPPCSPC) module with the receptive filed block (RFB) module to enhance the network's feature extraction capability. Second, a coordinate attention (CA) mechanism is introduced into the head part to enhance the network's feature representation ability and to improve detection accuracy. Third, a wise intersection over union (WIoU) loss function is employed to address the low-quality samples hindering model generalization during training, thereby improving the model's overall performance. The experimental results indicate that the proposed model exhibits enhancements across various performance metrics. Specifically, there is a 1.6% advancement in mAP_0.5, a corresponding 1.6% enhancement in mAP_0.5:0.95, a 1.3% elevation in precision, and a 1% increase in recall. Moreover, the model achieves parameter reduction by 3.2 million, leading to a decrease of 2.5 GFLOPS in computational cost. Notably, there is also an improvement of 2.81 milliseconds in single-image detection speed. This improved model can detect insulator defects for diverse materials, color insulators, and partial damage shapes in complex backgrounds.
Citation: Zhenyue Wang, Guowu Yuan, Hao Zhou, Yi Ma, Yutang Ma, Dong Chen. Improved YOLOv7 model for insulator defect detection[J]. Electronic Research Archive, 2024, 32(4): 2880-2896. doi: 10.3934/era.2024131
Insulators are crucial insulation components and structural supports in power grids, playing a vital role in the transmission lines. Due to temperature fluctuations, internal stress, or damage from hail, insulators are prone to injury. Automatic detection of damaged insulators faces challenges such as diverse types, small defect targets, and complex backgrounds and shapes. Most research for detecting insulator defects has focused on a single defect type or a specific material. However, the insulators in the grid's transmission lines have different colors and materials. Various insulator defects coexist, and the existing methods have difficulty meeting the practical application requirements. Current methods suffer from low detection accuracy and mAP0.5 cannot meet application requirements. This paper proposes an improved you only look once version 7 (YOLOv7) model for multi-type insulator defect detection. First, our model replaces the spatial pyramid pooling cross stage partial network (SPPCSPC) module with the receptive filed block (RFB) module to enhance the network's feature extraction capability. Second, a coordinate attention (CA) mechanism is introduced into the head part to enhance the network's feature representation ability and to improve detection accuracy. Third, a wise intersection over union (WIoU) loss function is employed to address the low-quality samples hindering model generalization during training, thereby improving the model's overall performance. The experimental results indicate that the proposed model exhibits enhancements across various performance metrics. Specifically, there is a 1.6% advancement in mAP_0.5, a corresponding 1.6% enhancement in mAP_0.5:0.95, a 1.3% elevation in precision, and a 1% increase in recall. Moreover, the model achieves parameter reduction by 3.2 million, leading to a decrease of 2.5 GFLOPS in computational cost. Notably, there is also an improvement of 2.81 milliseconds in single-image detection speed. This improved model can detect insulator defects for diverse materials, color insulators, and partial damage shapes in complex backgrounds.
| [1] |
X. Y. Peng, F. X. Liang, J. J. Qian, B. S. Yang, C. Chen, X. G. Zheng, Automatic localization of transmission line insulators based on airborne infrared image texture features, High Volt. Technol., 45 (2019), 922–928. https://doi.org/10.13336/j.1003-6520.hve.20190226033 (in Chinese) doi: 10.13336/j.1003-6520.hve.20190226033
|
| [2] |
Z. Y. Liu, X. R. Liao, J. Chen, H. Jiang, Review on intelligent processing of visible light images for inspection of power overhead line, Power Syst. Technol., 44 (2020), 1057–1069. https://doi.org/10.13335/j.1000-3673.pst.2019.0349 (in Chinese) doi: 10.13335/j.1000-3673.pst.2019.0349
|
| [3] |
S. Huang, Z. S. Wu, Z. G. Ren, H. J. Liu, Y. Gui, Review of research on intelligent inspection robots for electric power, Electr. Meas. Instru., 57 (2020), 26–38, https://doi.org/10.19753/j.issn1001-1390.2020.002.005 (in Chinese) doi: 10.19753/j.issn1001-1390.2020.002.005
|
| [4] |
G. Liu, L. H. Wu, H. Zhang, Research on key technologies for intelligent inspection of helicopters on transmission lines, J. Three Gorges Univ. (Natural Science), 36 (2014), 46-49+62. https://doi.org/10.13393/j.cnki.issn.1672-948x.2014.02.011 (in Chinese) doi: 10.13393/j.cnki.issn.1672-948x.2014.02.011
|
| [5] |
F. M. Chen, Y. H. Du, H. Chen, Application of image processing technology in intelligent inspection of helicopters on transmission lines, Zhejiang Electric Power, 31 (2012), 63–66. https://doi.org/10.19585/j.zjdl.2012.09.018 (in Chinese) doi: 10.19585/j.zjdl.2012.09.018
|
| [6] |
H. Yan, Measurement model for ice thickness of transmission lines based on image technology, Electr. Eng. Mate, (2021), 66-69+72. https://doi.org/10.16786/j.cnki.1671-8887.eem.2021.05.018 (in Chinese) doi: 10.16786/j.cnki.1671-8887.eem.2021.05.018
|
| [7] |
K. Yan, F. C. Wang, C. Y. Zhang, Edge detection of composite insulator hydrophobicity images based on Canny operator, J. Electric Power Sci. Technol., 28 (2013), 45–49+56. https://doi.org/10.3969/j.issn.1673-9140.2013.03.006 (in Chinese) doi: 10.3969/j.issn.1673-9140.2013.03.006
|
| [8] |
P. F. Felzenszwalb, R. B. Girshick, D. McAllester, D. Ramanan, Object Detection with Discriminatively Trained Part-Based Models, IEEE T. Pattern Anal., 32 (2010), 1627–1645, https://doi.org/10.1109/TPAMI.2009.167 doi: 10.1109/TPAMI.2009.167
|
| [9] |
K. P. Liu, B. Q. Li, L. Qin, Q. Li, F. Zhao, Q. L. Wang, et al., Review on the application of deep learning target detection algorithm in insulator defect detection of overhead transmission line, High Volt. Technol., 49 (2023), 3584–3595. https://doi.org/10.13336/j.1003-6520.hve.20220273 (in Chinese) doi: 10.13336/j.1003-6520.hve.20220273
|
| [10] |
H. Jiang, X. Qiu, J. Chen, X. Liu, X. Miao, S. Zhuang, Insulator fault detection in aerial images based on ensemble learning with multi-level perception, IEEE Access, 7 (2019), 61797–61810. https://doi.org/10.1109/ACCESS.2019.2915985 doi: 10.1109/ACCESS.2019.2915985
|
| [11] |
D. Sadykova, D. Pernebayeva, M. Bagheri, A. James, IN-YOLO: Real-time detection of outdoor high voltage insulators using UAV imaging, IEEE T. Power Deliver., 35 (2020), 1599–1601. https://doi.org/10.1109/TPWRD.2019.2944741 doi: 10.1109/TPWRD.2019.2944741
|
| [12] |
S. Q. Wang, Y. F. Liu, Y. H. Qing, C. X. Wang, T. Z. Lan, R. T. Yao, Detection of insulator defects with improved ResNeSt and region proposal network, IEEE Access, 8 (2020), 184841–184850. https://doi.org/10.1109/ACCESS.2020.3029857 doi: 10.1109/ACCESS.2020.3029857
|
| [13] |
C. X. Shi, Y. P. Huang, Cap-count guided weakly supervised insulator cap missing detection in aerial images, IEEE Sens. J., 21 (2021), 685–691. https://doi.org/10.1109/JSEN.2020.3012780 doi: 10.1109/JSEN.2020.3012780
|
| [14] |
W. Q. Zhao, M. F. Xu, X. F. Cheng, Z. B. Zhao, An insulator in transmission lines recognition and fault detection model based on improved faster RCNN, IEEE T. Instrum. Meas., 70 (2021), 1–8. https://doi.org/10.1109/TIM.2021.3112227 doi: 10.1109/TIM.2021.3112227
|
| [15] |
P. Luo, B. Wang, H. R. Ma, F. Q. Ma, H. X. Wang, D. H. Zhu, Low miss rate defect identification method based on combined target detection framework, High Volt. Technol., 47 (2021), 454–464. https://doi.org/10.13336/j.1003-6520.hve.20200701 (in Chinese) doi: 10.13336/j.1003-6520.hve.20200701
|
| [16] |
X. T. Zhang, Y. Y. Zhang, J. F. Liu, C. H. Zhang, X. Y. Xue, H. Zhang, et al., InsuDet: A fault detection method for insulators of overhead transmission lines using convolutional neural networks, IEEE T. Instrum. Meas., 70 (2021), 1–12. https://doi.org/10.1109/TIM.2021.3120796 doi: 10.1109/TIM.2021.3120796
|
| [17] |
Y. Liu, X. B. Huang, Research on insulator burst detection and localization based on YOLOv4 and improved watershed algorithm, Power Syst. Clean Energy, 37 (2021), 51–57. https://doi.org/10.3969/j.issn.1674-3814.2021.07.007 (in Chinese) doi: 10.3969/j.issn.1674-3814.2021.07.007
|
| [18] |
X. Y. Liu, X. R. Liao, S. B. Zhuang, H. Jiang, J. Chen, Insulator detection based on lightweight deep convolutional neural network, J. Fuzhou Univ. (Natural Science Edition), 49 (2021), 196–202. https://doi.org/10.7631/issn.1000-2243.20345 (in Chinese) doi: 10.7631/issn.1000-2243.20345
|
| [19] |
D. L. Wang, J. J. Sun, T. Y. Zhuang, M. S. Li, R. Zhu, Detection method for self-explosion defects of glass insulators based on improved generative adversarial networks, High Volt. Technol., 48 (2022), 1096–1103. https://doi.org/10.13336/j.1003-6520.hve.20210236 (in Chinese) doi: 10.13336/j.1003-6520.hve.20210236
|
| [20] |
J. Kang, Q. Wang, W. B. Liu, Y. Xia, Multi defect detection network for aerial insulators integrating CAT-BiFPN and attention mechanism, High Volt. Technol., 49 (2023), 3361–3376. https://doi.org/10.13336/j.1003-6520.hve.20221803 (in Chinese) doi: 10.13336/j.1003-6520.hve.20221803
|
| [21] |
G. Singh, S. F. Stefenon, K. C. Yow, Interpretable visual transmission lines inspections using pseudo-prototypical part network, Mach. Vision Appl., 34 (2023), 41. https://doi.org/10.1007/s00138-023-01390-6 doi: 10.1007/s00138-023-01390-6
|
| [22] |
B. J. Souza, S. F. Serfenon, G. Singh, R. Z. Freire, Hybrid-YOLO for classification of insulators defects in transmission lines based on UAV, Int. J. Elect. Powe, 148 (2023), 108982. https://doi.org/10.1016/j.ijepes.2023.108982 doi: 10.1016/j.ijepes.2023.108982
|
| [23] |
S. F. Serfenon, G. Singh, B. J. Souza, R. Z. Freire, K. C. Yow, Optimized hybrid YOLOu-Quasi-ProtoPNet for insulators classification, IET Gener. Transm. Dis., 17 (2023), 3501–3511. https://doi.org/10.1049/gtd2.12886 doi: 10.1049/gtd2.12886
|
| [24] | C. Y. Wang, A. Bochkovskiy, H. Y. M. Liao, YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors, arXiv e-prints, (2022), arXiv: 2207.02696. https://doi.org/10.48550/arXiv.2207.02696 |
| [25] | S. Liu, D. Huang, Y. H. Wang, Receptive field block net for accurate and fast object detection, In: Ferrari, V., Hebert, M., Sminchisescu, C., Weiss, Y. (eds) Computer Vision – ECCV 2018. ECCV 2018. Lecture Notes in Computer Science(), vol 11215,404–419. Springer, Cham. https://doi.org/10.1007/978-3-030-01252-6_24 |
| [26] | Q. B. Hou, D. Q. Zhou, J. S. Feng, Coordinate attention for efficient mobile network design, arXiv preprint, (2023), arXiv: 2301.02907. https://doi.org/10.48550/arXiv.2103.02907 |
| [27] | Z. J. Tong, Y. H. Chen, Z. W. Xu, R. YU, Wise-IoU: Bounding box regression loss with dynamic focusing mechanism, arXiv preprint, (2023), arXiv: 2301.10051. https://doi.org/10.48550/arXiv.2301.10051 |
| [28] | C. Szegedy, W. Liu, Y. Q. Jia, P. Sermanet, S. Reed, D. Anguelov, et al., Going deeper with convolution, in 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), (2015), 1–9. https://doi.org/10.1109/CVPR.2015.7298594 |
| [29] | F. Yu, V. Koltun, Multi-Scale context aggregation by dilated convolutions, arXiv preprint, (2015), arXiv: 1511.07122. https://doi.org/10.48550/arXiv.1511.07122 |
| [30] |
Z. Y. Wang, G. W. Yuan, H. Zhou, Y. Ma, Y. T. Ma. Foreign-Object Detection in High-Voltage Transmission Line Based on Improved YOLOv8m, Appl. Sci., 13 (2023), 12775. https://doi.org/10.3390/app132312775 doi: 10.3390/app132312775
|
Figures(13) / Tables(4)
Zhenyue Wang, Guowu Yuan, Hao Zhou, Yi Ma, Yutang Ma, Dong Chen. Improved YOLOv7 model for insulator defect detection[J]. Electronic Research Archive, 2024, 32(4): 2880-2896. doi: 10.3934/era.2024131
DownLoad: