Abstract :
Mechanical parts defect detection is an important problem in machinery manufacturing. Therefore, in this paper, we propose a novel mechanical parts defect detection method based on computer vision technology. Firstly, structure of the proposed mechanical parts defect detection system is described, which is made up of five modules: 1) Image collecting, 2) Image processing, 3) Image analyzing, 4) Image understanding, 5) Industrial control. Secondly, the SVM based mechanical parts defect detection, in which the parameters are optimized by the quantum behaved particle swarm optimization algorithm. Thirdly, to testify the effectiveness of the proposed algorithm, the MV-1300UM industrial digital camera is used to collect mechanical part images, and experimental results demonstrate that the proposed method can achieve high recognition rate.
