Image detection method for power line warning ball installation robot based on improved YOLOv8

Aiming at the problems such as insufficient detection accuracy and high false detection rate in the target detection process due to the use of insulated traction soft rope on the line during the operation of the transmission line warning ball installation robot, this paper proposes an image processing method based on the improved YOLOv8 network. Firstly, the feature extraction capability is optimized by introducing DCNv4 deformable convolution and CBAM attention mechanism in the YOLOv8 network; secondly, the CIoU loss function is replaced with Wise-IoU loss function, which further improves the detection accuracy and robustness; and finally, the reliability of the gap identification is checked using DeepLabv3+ semantic segmentation algorithm. The experimental results show that the method exhibits comprehensive performance improvement in four types of target detection tasks, namely, bolt, wire clip, wire and wire clip closure, with the average precision, average recall and average F1-score reaching 94.91%, 95.43% and 95.16%, respectively, which effectively improves the operational efficiency of the warning ball installation robot.