Cascaded based Adaptive R-CNN Network for Insulator Defect Detection

Abstract

Target detection is a popular direction in computer vision and digital image processing. It is widely used in robot navigation, intelligent video surveillance, industrial detection, aerospace, and many other fields. It has important practical significance to reduce the consumption of human capital through computer vision. Due to the widespread use of deep learning, target detection algorithms have developed relatively rapidly. In this paper, we propose a method for detecting insulator defects Based on Cascaded Fast R-CNN. We use RoIAlign instead of RoIPooling to get a better positioning effect and add a mask branch on the basis of Faster R-CNN to complete the pixel-level segmentation task, which can detect insulator defects in real-time with high detection accuracy.