Research On Defect Inspection System Of Fastening In Railway Based On Computer Vision

Line maintenance for the subway and rail systems work and has an important role in the safe operation, and it is periodically needed for preventing dangerous situations. Inspection is operated manually by trained human operator walking along the track searching for visual anomalies. This monitoring is unacceptable for slowness and lack of objectivity, as the results are related to the ability of the observer to recognize critical situations. Currently, the line maintenance rely mainly on highly trained patrol Road railway maintenance workers to labor inspection and visual methods to determine the rail road damage pillow, rail wear and the lack of nuts and other fasteners line conditions. This manual inspection pillow line road, rail fastening systems and methods required a long time, lack of objectivity, low efficiency, missd high. To solve the traditional visual inspection of bare eyes can no longer meet the railway maintenance requirements, this paper focus on the Computer-vision-based System for detecting rail fastening automatically.Based on the analysis of the rail fastening nut missing detection technology on the current status of research, take the this paper focus on the Computer-vision-based System for detecting rail fastening. Design of a computer vision-based detection system rail fastening nut missing, the establishment of a rail fastening nut missing detection system framework. The research including image256gray、filtering、 binarization、Image Enhancement、Fastener positioning、Feature extraction、 design the Classifier、Missing nut fastener positioning. The important research points are designing Fastener location algorithm、image feature extraction algorithm、image classification algorithm, missing nut fastener positioning method. Proposed an algorithm of positioning the rail fastening based on scanning pixels and statistics of specific areas, will be based on K2DPCA-2DPCA feature extraction algorithms used in image feature extraction in the rail fastening and used to support a small samples of the support vector machine algorithm as a final classifier, using GPS locator determine the final positioning of the missing nut fastener location.This paper develop the framework of this system, as demonstrated by the experiment, this system can effectively identify missing nut of rail fastening, which, could server as an alternative of the visual inspection and reduce the recognition time effectively.