A defect detection method of leaning pins for multiple electrical components based on machine vision

This paper proposes a defect detection method of leaning pins for multiple electrical components based on machine vision. According to the distribution of the plugs on the electrical component surface, a fixture is firstly designed which allows photograph and detection for multiple electrical components, thus avoids replacement of fixture when replacing the electrical components. Then, the high-resolution industrial camera and imaging system is used to take photos of the electrical component surface. After that, the defect detection of leaning pins is realized through the steps of electrical component surface detection, ROI segmentation, pin location extraction and leaning pins detection. The iterative closest point (ICP) algorithm is introduced for point set registration to detect the leaning pins in which the point set consists of the pins of all plugs on a surface. Finally, experiments are carried out on three differential electrical components, and the results show that the proposed method is able to detect the leaning pins automatically.