Improved Detection of Transmission Line Defects Based on YOLOv7

Transmission lines, as carriers of high-voltage electric energy, are crucial lifelines. The fatigue and corrosion damage that occur over the long service life of engineering transmission and transformation projects pose significant threats to the safe operation of transmission lines[1]. The tower body and fittings form the main structure of the power system's transmission segment. When defects occur in the tower body, transmission line cables, or fittings, it can often lead to the dropping of overhead ground wires. More seriously, it can result in “tower collapse” accidents, which can cause power outages and tripping faults, severely affecting the safe and stable operation of the power system. After a fault occurs, repair efforts may face challenges due to geographical constraints, leading to delays in restoration and potentially triggering widespread power outages. Therefore, conducting regular inspections of transmission lines and towers to identify and address potential issues is an important measure to prevent economic losses. This paper utilizes a deep learning model based on an improved YOLOv7 with the Focal EloU loss function, which significantly enhances the model's performance in object detection tasks, particularly in terms of detection accuracy and generalization ability. By extracting the tower area images from the top of distribution towers and performing feature extraction, the model aims to detect tower corrosion and transmission line defects.