A Rail Abrasion Detection System Based on Image Processing

Abstract

Quality of steel rail affects directly the running state of locomotive. According to detection problem of rail head’s abrasion loss, a feasibility study was done that collecting image through CCD camera. Computing models of abrasion detection were established by MATLAB soft. The key question about image location was studied using the method of edge detection and image filling. Results indicate that the image detection system could obtain the abrasion loss accurately in non-contact condition. This method could provide a reference for the development of loss prevention and new detection technology.