Abstract: In the face of the problems of low intelligence and low accuracy in the detection of belts by existing detection technologies,a belt tear detection system based on laser visual analysis was proposed. Firstly,according to the production conditions of the open-pit coal mine,an image acquisition system was designed to collect the laser fringe images,and the equipment model and its layout scheme were determined. Secondly,an image preprocessing scheme integrating image ROI extraction,image grayscale,image enhancement and image stitching was proposed to improve the quality of the acquired laser fringe images. Secondly,a laser fringe centerline extraction scheme based on Steger algorithm was proposed,and a clear laser fringe centerline was obtained. Finally,a tear analysis scheme based on the curvature change of the center curve of the laser fringe was proposed,and the tear condition of the belt was determined by the curvature change. In order to verify the superiority of the experiment,an experimental platform was built and field experiments were carried out,and the experimental comparison was carried out with the algorithm proposed by LI X G. The experimental results show that the accuracy of the tear detection system is 94. 36% under laboratory conditions. Under field conditions,the accuracy of the tear detection system is 93%, which proves the practicability of the designed belt tear detection system.