Research On Automatic Detection Technology Of The Surface Of Microwave Cavity Based On Machine Vision

With the gradual transformation of my country's manufacturing industry from automation to intelligence,3C electronic product automatic manufacturing production lines have increasingly higher requirements for the automatic detection of surface defects of various parts and final products.Traditional manual detection methods and the naked eye with the help of electron microscopes Semi-automatic detection methods have the disadvantages of low efficiency and large errors,which cannot meet the needs of automated production lines.This paper analyzes the research status of machine vision and defect detection technology at home and abroad.Based on machine vision technology,an automatic defect detection system for microwave resonator with complex characteristics of a certain type is designed,and automatic defect detection of microwave resonator is completed.The main research contents are as follows:(1)According to the surface characteristics of different materials,the system composition and lighting scheme were designed and verified.The selection of camera,lens and light source was completed under laboratory conditions,as well as the construction and experimental verification of the principle prototype.(2)A detailed analysis of the defects on the silver-plated ceramic side of the microwave cavity was carried out.Based on the comparison and analysis of different image preprocessing methods,an image processing algorithm for obtaining the Region of Interest(ROI)was designed,and then the area to be inspected was determined.Later,the method of morphological analysis was used to detect defects and analyze the types of defects through the shape and spatial distribution of the defects,so as to achieve reliable detection of defects,and the detection accuracy rate reached more than 98.9%.(3)For the PCB(Printed Circuit Board)surface of the microwave resonator,a method based on template matching is adopted,which greatly reduces the difference in gray value between different areas due to the difference in light reflectivity,and then analyzes the image according to the defect characteristics.The detection of defects is realized.This method has good detection effects for single-area and cross-area defects,and the detection accuracy rate can reach 93.9%.(4)In view of the extreme unevenness of illumination caused by the complex environment of the workshop site,the evaluation criteria and gray-level equalization algorithm are designed.Through the verification of the actual detection results,the unevenness of the image after equalization is reduced by more than 50%.(5)After more than a year of production site operation,the algorithm was continuously adjusted and optimized according to the needs of the site conditions.In the final operation test of more than three months and more than 1,000 samples per day,the side detection accuracy rate can reach more than 98.9%,The detection accuracy rate of PCB surface can reach more than 93.9%,and the comprehensive detection accuracy rate can reach more than 92.9%.