Automatic Recognition System for Gear Defects Based on Machine Vision

It is particularly necessary to conduct defect detection on the gears during the production process to ensure that the quality of the gear products meets the requirements when they leave the factory. Adopting adaptive median filtering algorithm to denoise gear images, using genetic neural network algorithm for gear defect segmentation, extracting geometric and shape features of gear defect regions based on their characteristics. Given the significant differences in the characteristics of cracks, dents, holes, and tooth breakage defects, different gear defect recognition algorithms are adopted. Based on the research on gear defect recognition algorithms, a machine vision based automatic gear defect recognition system was designed and developed, and gear sample images were collected for simulation experiments of the system. The experimental results show that the system can accurately identify defects in gears and classify them with high recognition accuracy. The system also runs very fast, with an average time of 2 seconds required to detect an image. It has high real-time performance and greatly reduces the time required by manual detection methods.