Track Surface Defect Detection Based on Image Processing

摘要

In this paper, computer vision-based methods are presented to detect the rail track surface defects automatically. The detection is the key foundation to inspect and assess railways, and for the operation safety and rail maintenance, railways inspection is the critical task. To achieve this goal, the rail surface edge's likelihood is investigated, and the Canny edge detector for defects extraction is introduced to guarantee the detection of the rail surface damage accurately. The analysis performed on some image data captured on the field has demonstrated encouraging detection performance on rail track surface defect detection.