Visual Inspection and Grading System for Ceramic Glass Material Production Based on Computer Vision Technology

Traditional manual detection methods have many drawbacks in detecting surface defects of ceramic glass materials. At present, a method is needed to replace manual electronic automatic detection to avoid the problems of time-consuming, laborious and large errors. The purpose of this paper is to study the design of a visual inspection and classification system for the production of ceramic glass materials based on computational vision techniques. A computer vision-based quality control system for ceramic glass materials is designed. The mechanical structure is equipped with a rotating mechanism for adjusting the angle of the ceramic glass material, and the servo motor controls the working distance of the camera to adjust the image capture of the ceramic glass material of different sizes. In the image preprocessing task, in order to avoid the influence of the background image, an improved media filtering method is used to solve it. The design of the program and the realization of the main operation of the system are introduced. The software is written in Matlab language under Windows7 operating system. The main functions of the system are realized by software one by one. After testing 100 glassware, the correct detection rate of scratch defects is 99%, which meets the test requirements.