Defect detection of micromotor friction plate based on machine vision

The friction plate of micromotor is detected to judge whether it is complete by means of machine vision technology. Firstly, grayscale image of the friction plate is binarized according to the selected threshold, and then the connected region of the binary image is marked. Based on major semi axis of the minimum bounding rectangle of the target region, the region of the friction plate boundary is extracted. Secondly, gaps in the region of the friction plate boundary are eliminated through closing operation, then the region of boundary is transformed into sub-pixel contour. Finally, an improved least square method is used to fit the contour to obtain the center and radius of the friction plate. Distances between pixel points and the center of the friction plate are calculated in order to find the minimum distance. The minimum distance is compared with the radius to estimate whether the friction plate is incomplete. The experiment results show that the accuracy of the detection system reaches 98%. Therefore, the system can replace manual detection, effectively reduce labor cost and improve detection efficiency.