Research On Circuit Board Defect Detection System Based On Machine Vision

With the implementing of our country's intelligent manufacturing 2025 strategy and the continuous improvement of the level of industrial informatization,the demand for the quality of electronic products is becoming higher and higher.As the core component of electronic products,the quality test of PCB has become the key to meet the increasing quality requirements of the electronic manufacturing industry.The traditional PCB defect detection method can't meet the requirements of high speed and high precision in the future production process.In order to solve this problem,this dissertation studies the PCB defect detection system based on machine vision.The main research contents are as follows:1)The dissertation introduces the development status of circuit board defect detection at home and abroad,and analyzes the research significance of circuit board defect detection system.A circuit board defect detection system is designed,and the overall structure and working principle of the system are introduced.A centralized control platform for defect detection system is designed,which can remote monitor and control of the key information of the system.2)Various factors that affect the quality of image acquisition are analyzed,and the image preprocessing algorithms are studied from image contrast enhancement,image denosing method and image segmentation.The camera imaging model is analyzed,by studying the camera calibration algorithm can gain the calibration parameters and remove the effect of distortion on image quality.3)The dissertation researches on the PCB defect detection algorithm from two aspects:bare board and finished board.For bare board,design a defect detection method based on contour contrast,which can be used for classification and recognition of common bare board defects.For finished board,it design a detection method based on SVM for solder joint defect detection and an intelligent detection method based on neural network for resistance component defect detection,and two methods are verified and the results are analyzed.