Railway clearance intrusion detection method with monocular vision

摘要

Foreign objects intruding railway clearance significantly threaten the safety of train operation.Therefore,real-time intrusion detection is very important and has to be settled urgently.In this paper,an intrusion detection method with monocular stereo vision is proposed to judge the clearance intrusion and object classification.Firstly,the detection area is divided into danger and safe region by rail detection with Hough transformation.Then image dithering are detected and eliminated in the image sequence to generate a stabile background.After background updating,the foreground is extracted by background subtraction method and classified by several regulars of train moving.Experiments show that it is an effective way to detect the foreign object intruding railway clearance.