Research on application of corner detection in thread vision measurement

Abstract

Abstract With the rapid development of the automation field, automatic inspection of fasteners in the aviation field has become a top priority. It is precious to realize the fast measurement of important information in fastener production by non-contact. This paper aims to detect the optical image detection of the external thread angle of fasteners. We analyze the traditional Ramer algorithm’s contour segmentation steps in separating thread edge from thread bottom arc to obtain accurate thread angle size and thread bottom arc radius. And Harris algorithm is proposed to find corners again by comparison. Finally, using a discrete point curvature to verify the Harris algorithm’s accuracy in corner detection. Meanwhile, by comparing the edge fitted thread angle with the measured thread angle to verify the Harris algorithm’s reliability in corner detection.