Abstract: For the situation that conveyor belt is easily deviated during operation, the belt deviation monitoring method based on the computer vision is proposed. Firstly, the Region of Interest (ROI) in video surveillance is set to reduce the amount of calculation, and the image preprocessing is performed on the ROI. Then the improved Canny edge detection algorithm is used to get the ROI edge binary image, and the linear feature of conveyor belt edge is extracted by using Progressive Probabilistic Hough Transform (PPHT). Finally, we determine whether the conveyor belt is deviated or not according to the linear features.