ELA-YOLO: a lightweight detection algorithm for critical components of transmission lines in complex scenarios

Abstract

This study proposes the enhanced line-detection adaptive you only look once (ELA-YOLO), an enhanced YOLOv8-based object detection algorithm, to improve the identification and classification of critical power components. By integrating efficient multi-scale attention (EMA) into redesigned cross stage partial feature fusion (C2f) modules (C2f_EMA), the backbone network achieves dynamic multi-scale feature fusion. The neck network is further optimized through asymmetric padding convolution (APConv) in C2f_AP modules, enhancing spatial feature integration. Additionally, the large selective kernel (LSK) attention mechanism strengthens context-aware feature extraction capabilities. Experimental results demonstrate that ELA-YOLO outperforms YOLOv8s with a 2.8% improvement in mean average precision at intersection over union threshold 0.50 (mAP₅₀) while incurring only a 4.5% computational overhead, establishing an optimal balance between detection accuracy and operational efficiency for real-world power equipment inspection scenarios.

Keywords

A

References

[1] Irizarry J, Gheisari M, Walker B N. Usability assessment of drone technology as safety inspection tools. Journal of information technology in construction, 2012, 17(12): 194-212[J]

[2] Nooralishahi P, Ibarracastanedo C, Deane S, et al.. Drone-based non-destructive inspection of industrial sites: a review and case studies. Drones, 2021, 5(4): 106-135 J]

[3] Girshick R, Donahue J, Darrell T, et al.. Rich feature hierarchies for accurate object detection and semantic segmentation. 2014 IEEE Conference on Computer Vision and Pattern Recognition, June 23–28, 2014, Columbus, OH, USA, 2014, New York, IEEE, 580-587[C]

[4] Ren S Q, He K M, Girshick R, et al.. Faster R-CNN: towards real-time object detection with region proposal networks. IEEE transactions on pattern analysis and machine intelligence, 2016, 39(6): 1137-1149 J]

[5] Lei X S, Sui Z H. Intelligent fault detection of high voltage line based on the faster R-CNN. Measurement, 2019, 138(1): 379-385 J]

[6] Hu H T, Xu J, Huang Y, et al.. Insulator defect detection of transmission tower based on improved faster R-CNN. Information technology and informatization, 2023, 16(7): 63-66[J]

[7] Liu W, Anguelov D, Erhan D, et al.. SSD: single shot multibox detector. 14th European Conference on Computer Vision, October 11–14, 2016, Amsterdam, The Netherlands, 2016, Heidelberg, Springer, 21-37[C]

[8] Li R H, Yang Y, Li N, et al.. Transmission line pin detection based on improved SSD. The 3rd International Conference on Artificial Intelligence, Information Processing and Cloud Computing, June 21–22, 2022, Online, 2022, New York, IEEE, 1-6[C]

[9] Liu M, Li Z, Li Y, et al.. A fast and accurate method of power line intelligent inspection based on edge computing. IEEE transactions on instrumentation and measurement, 2022, 71: 1-12[J]

[10] Cha S K, Huang C R. Insulator self-explosion fault detection method based on ConvNeXt and attention mechanism. Ningxia electric power, 2023, 27(3): 42-50[J]

[11] He M, Qin L, Deng X, et al.. MFI-YOLO: multi-fault insulator detection based on an improved YOLOv8. IEEE transactions on power delivery, 2024, 39(1): 168-179 J]

[12] Lu Q, Lin K H, Yin L F. 3D attention-focused pure convolutional target detection algorithm for insulator defect detection. Expert systems with applications, 2024, 249: 123720 J]

[13] Wu J, Jing R, Bai Y S, et al.. Small insulator defects detection based on multiscale feature interaction transformer for UAV-assisted power IoVT. IEEE internet of things journal, 2024, 11(13): 23410-23427 J]

[14] Hosseini M M, Umunnakwe A, Parvania M, et al.. Intelligent damage classification and estimation in power distribution poles using unmanned aerial vehicles and convolutional neural networks. IEEE transactions on smart grid, 2020, 11(4): 3325-3333 J]

[15] Hao Y, Liang W, Wang X, et al.. Automatic calculation of graphic area change rate for icing overhead power line insulators based on bounding box automatic matching and GrabCut contour automatic segmentation. IEEE transactions on power delivery, 2023, 38(4): 2821-2830 J]

[16] Ouyang D L, He S, Zhan J, et al.. Efficient multi-scale attention module with cross-spatial learning. 2023 IEEE International Conference on Acoustics, Speech and Signal Processing, June 4–8, 2023, Rhodes Island, Greece, 2023, New York, IEEE, 1-5[C]

[17] Yang J N, Liu S L, Wu J J, et al.. Pinwheel-shaped convolution and scale-based dynamic loss for infrared small target detection. Proceedings of the AAAI conference on artificial intelligence, 2025, 39(9): 9202-9210 J]

[18] Li Y X, Hou Q B, Zheng Z H, et al.. Large selective kernel network for remote sensing object detection. 2023 IEEE/CVF International Conference on Computer Vision, October 2–6, 2023, Paris, France, 2023, New York, IEEE, 16748-16759[C]

[19] Woo S, Park J, Lee J Y, et al.. CBAM: convolutional block attention module. 2018 the European Conference on Computer Vision, September 8–14, 2018, Munich, Germany, 2018, Heidelberg, Springer, 3-19[C]

[20] Hu J, Shen L, Sun G. Squeeze-and-excitation networks. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 18–23, 2018, Salt Lake City, UT, USA, 2018, New York, IEEE, 7132-7141[C]

[21] Pang F Q, Jia C H, Li Y, et al.. Pi-Score: an estimation strategy of the class prior in positive-unlabeled learning for electrical insulator defect detection with incomplete annotations. Journal of sensors, 2023, 1: 6696721 J]

[22] Jing H, Yu L, Li X, et al.. DyHeadNet: insulator defect detection with a dynamic detection head combining lightweight and attention mechanism. 2024 China Automation Congress, November 1–3, 2024, Qingdao, China, 2024, New York, IEEE, 5751-5756[C]

[23] Jiao R, Liu J, Li K, et al.. YOLO-DTAD: dynamic task alignment detection model for multicategory power defects image. IEEE transactions on instrumentation and measurement, 2025, 74: 1-14 J]

[24] Liu S H, Zha J L, Sun J, et al.. EdgeYOLO: an edge-real-time object detector. 2023 42nd Chinese Control Conference, July 24–26, 2023, Tianjin, China, 2023, New York, IEEE, 7507-7512[C]

Rights & permissions

Tianjin University of Technology