Academic Journal of Engineering and Technology Science, 2025, 8(1); doi: 10.25236/AJETS.2025.080102.
Guangyao Wang
College of Information Engineering, Nanjing University of Finance and Economics, Nanjing, 210023, China
Power transmission lines are a critical component of the power system, and due to their widespread presence in outdoor environments, they are highly susceptible to the attachment of foreign objects. These foreign objects pose a risk of causing accidents in power transmission lines. To ensure the stability and safety of the power supply, regular inspections of transmission lines are necessary. Aiming at the precise and efficient detection of foreign objects along power transmission lines, this research puts forward a new FMH-YOLO model which is founded on the YOLOv8 network. Firstly, we introduce Partial Convolution and the EMA to design a new Faster-EMA Block (FEB) structure. Based on this, we construct the FE-C2F module to enhance the model's feature extraction capabilities. Secondly, we propose a multi-scale lightweight convolution module based on GhostConv. Building on this, we design the ML-C2F module to improve the model's ability of feature integration. Additionally, a detection head specifically for small objects is added. The experimental results demonstrate that FMH-YOLO outperforms the baseline YOLOv8 in multiple evaluation metrics, achieving improvements of 2.7% in Precision, 2.2% in Recall, 2.9% in AP50, and 1.8% in [email protected]. These results evidently signify its enhanced efficacy and outstanding capabilities in the mission of identifying foreign objects on power transmission lines. Moreover, the graphical representations also attest to the remarkable proficiency of FMH - YOLO. Therefore, this method provides technical support for foreign object detection on power transmission lines, and when deployed on drones, it can effectively carry out inspection tasks.
YOLOv8, object detection, transmission line, Faster-EMA Block, multi-scale lightweight convolution
Guangyao Wang. FMH-YOLO: Detecting Foreign Objects on Transmission Lines via Enhanced Yolov8. Academic Journal of Engineering and Technology Science (2025) Vol. 8, Issue 1: 8-17. https://doi.org/10.25236/AJETS.2025.080102.
[1] Zhang W, Liu X, Yuan J, et al. RCNN-based foreign object detection for securing power transmission lines (RCNN4SPTL)[J]. Procedia computer science, 2019, 147: 331-337.
[2] Wang G, Chen Y, An P, et al. UAV-YOLOv8: A small-object-detection model based on improved YOLOv8 for UAV aerial photography scenarios[J]. Sensors, 2023, 23(16): 7190.
[3] Hao J ,Yan G ,Wang L , et al.A Lightweight Transmission Line Foreign Object Detection Algorithm Incorporating Adaptive Weight Pooling[J].Electronics,2024,13(23):4645-4645.
[4] Yang G, Wang J, Nie Z, et al. A lightweight YOLOv8 tomato detection algorithm combining feature enhancement and attention[J]. Agronomy, 2023, 13(7): 1824.
[5] Liu G, Reda F A, Shih K J, et al. Image inpainting for irregular holes using partial convolutions[C]//Proceedings of the European conference on computer vision (ECCV). 2018: 85-100.
[6] Ouyang D, He S, Zhang G, et al. Efficient multi-scale attention module with cross-spatial learning[C]//ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023: 1-5.
[7] Chen J, Kao S, He H, et al. Run, don't walk: chasing higher FLOPS for faster neural networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2023: 12021-12031.
[8] Terven J, Córdova-Esparza D M, Romero-González J A. A comprehensive review of yolo architectures in computer vision: From yolov1 to yolov8 and yolo-nas[J]. Machine Learning and Knowledge Extraction, 2023, 5(4): 1680-1716.