Non-contact dimension measurement of mechanical parts based on image processing

Aiming at the problems that mechanical parts are mass produced in industrial production, and the workload of traditional manual measurement is heavy, the measurement efficiency is low, an online non-contact measurement system based on image processing and machine vision technology is presented. The hardware of the measurement system includes light source, lens, camera and computer, the measurement system adopts transmitted illumination to highlight the contour feature of mechanical parts. Preprocessing is firstly performed to the captured image by the software of the measurement system, then threshold segmentation and the edge contour extraction are conducted; finally, in order to improve the detection precision of the geometric elements, a method to locate the center of circles accurately is adopted, which based on Hough transform algorithm. The results show the validity of the method and the feasibility of the algorithm in the system, it improves the efficiency of the measurement system and realizes the non-contact online measurement of mechanical parts.