Research on Deviation Detection of Belt Conveyor Based on Inspection Robot and Deep Learning

Figure 1

Schematic of the deep-learning-based belt conveyor deviation detection method. The inspection robot captures an image and divides it into imgO and imgI; then, they are input into the ROI detector (OM-SSD) to extract the ROI_O and ROI_I, respectively. The ROIs are converted into gray images, and then, the conveyor belt edge detection algorithm and the idler outer edge detection algorithm are implemented on them to get the edges. The points are the midpoints of the detected lines representing the conveyor belt edges and the vertices of the idler outer edges, respectively. The coordinates of the points are corrected by the geometric correction algorithm and then used to estimate the DD. The distances between the two points in ROI_O and ROI_I represent the lengths of the exposed outer and inner idler, and the difference between the lengths indicates the DD. (a) Phase 1: ROI extraction. (b) Phase 2: DD estimation.