Inspection of rail surface defects based on image processing

Abstract

A rail surface defects inspection method based on automated machine vision system is proposed in the paper. Two kinds of defect images including spalling of rail head and cracks in surface are analyzed with this method. Some related algorithms comprising denoising, image segmentation and feature extraction are applied in processing the images of rail surface defect. Then accurate region of defect is extracted and recognized by dynamic thresholding and feature matching. Percentage of wear of rail head and length of cracks in surface are calculated next as an evaluation of flaw on inspected rail head section.