Infrared Image State Evaluation of Power Cables Based on Mask R-CNN and BP Joint Algorithm

In order to solve the problems of traditional image processing algorithms in handling precise target detection and condition monitoring of power cable, this paper proposes an improved infrared image-based condition diagnosis scenario of power cable based on Mask R-CNN and BP joint algorithm. Mask R-CNN is used to solve the problem of refined image segmentation when the background of cable infrared image is complex. And BP neural network algorithm is used to classify and identify the key features of power cable. The results show that the proposed method has a good detection effect on the operation status of cables in infrared images with an average accuracy rate of 87.63%, which presents a good solution to the problem of infrared image identification of substation equipment and its status assessment.