Real-time Detection Technology Based on Dynamic Line-edge for Conveyor Belt Longitudinal Tear

We propose a method of conveyor belt longitudinal tear dynamic line edge detection based on CCD digital image processing. CCD camera captured video frames of conveyor belt at a certain frequency, we first applied median noise filtering and binarization to preprocessing the motion images of conveyor belt on the condition that they were not degenerative. Then Canny edge detection operator detected the edge point of the conveyor belt longitudinal tear and Hough transformation acquired the number of straight lines of longitudinal edge. Our approach was to calculate the length of each straight line separately and add these figures up. Finally, after calculating the length of the belt tear by comparing the proportion of the real size of conveyor belt and that in the image, the computer made warning or stop. Comprehensive experiment shows that this method can extract conveyer belt longitudinal tear edge line accurately and stably and make a quick intelligent response.