Wire defect visual detection algorithm based on Faster R-CNN

The transmission system plays an important role in the power transmission process. Line defect detection is the key link to ensure the safe operation of the transmission system. Its detection effect directly determines the inspection effect. However, image detection algorithms of line defects have always relied on traditional image processing methods, which often require a large number of presuppositions and lack universality and effectiveness under complex real conditions. In this paper, the deep learning method is introduced into the broken strand detection algorithm, which provides a new idea for broken strand detection. In this paper, a method for detecting broken strands based on Faster R-CNN deep learning is proposed. The background, theory and experiment of the two algorithms are described respectively. They overcome the shortcomings of the poor universality of traditional image algorithms. In addition, they can improve the unsatisfactory effect of general deep learning methods in detecting broken strands. The operation efficiency of the robot is greatly improved, with a broad application prospect.