Identification Method of Railway Insulator Based on Edge Features

Abstract

Automatic detection and identification of insulators from the image is an important prerequisite for their state detection. In this paper, an insulator recognition algorithm based on Sobel edge feature and SURF algorithm is proposed. The algorithm firstly extracts image edge features, then uses SURF algorithm to extract feature points on edge, and uses Haar wavelet to describe feature points, and uses Euclidean distance to match the detected feature points. Finally, the MSAC algorithm is used to eliminate the error matching caused by noise and other disturbances, so as to realize the intelligent identification of the insulators in the catenary. The experimental results show that the proposed algorithm can accurately identify insulators from target images, which provides a feasible reference for visual identification and positioning of insulators in the intelligent detection system of catenary insulator of electrified railway.