Research On PCBA Defect Detection System Based On Machine Vision

PCBA plays a crucial role in the field of electronic products.With the continuous rise of scientific and technological power and the material level of People's Daily life,the structure of circuit boards is complex and diverse,which makes it more difficult to control the quality of circuit boards.At present,the domestic level of automatic online detection technology is limited,and most of them rely on manual detection and semi-automatic detection.Traditional quality detection methods are time-consuming,have high requirements on professional level of staff and low detection accuracy,which cannot meet the requirements of the new industrial situation.With the improvement of the manufacturing technology level of electronic devices,the specifications of components are becoming smaller,the number is increasing,and the layout is becoming more and more compact.The traditional detection method relying on probe contact may change the original parameters of components,which increases the difficulty of fault detection.Therefore,it is urgent to develop a fast,efficient and accurate method to detect defects automatically.Based on the research of PCBA defect detection principle and technology,this paper designs and builds a complete modern detection system to acquire,identify and detect PCBA images.The detection system consists of three parts:motion module,image acquisition module and image processing module.The motion module is composed of PLC,stepper motor driver and stepper motor.It is used to transport the circuit board to and from the starting point to the detection position,and fuzzy PID control is added to ensure the accuracy of transmission on the detection platform.The image acquisition module includes a CCD industrial camera and LED light source,which are used to collect PCBA images to be detected from a preset Angle.The function of the image processing module is to match the detected image with the template image by using the template matching algorithm based on gray value after preprocessing the collected image to be detected.In addition to the above parts,this paper also proposes an image denoising algorithm combining improved wavelet threshold and improved bilateral filter for the template matching algorithm based on gray value which is susceptible to noise.According to the overall design of the above system,the physical object of the PCBA defect detection system was successfully built,and the equipment could well achieve the expected goal.Matlab is used to carry out simulation experiments on the image denoising optimization algorithm proposed in this paper.By comparing the results with the traditional mean filtering algorithm,Gaussian filtering algorithm,hard threshold denoising algorithm,soft threshold denoising algorithm,wavelet threshold denoising algorithm's mean square error and peak signal-to-noise ratio,the algorithm has better denoising effect.A refrigerator PCBA database was selected for the test of defect detection program,and the number of detection targets was 18,600.The detection results of the test program showed that the number of missed judgments was 53,the number of misjudgments was 10,that is,the rate of missed judgments was 0.28%,the rate of misjudgments was 0.054%,the accuracy was 99.666%,and the detection time was 9.1s.Experimental results show that the system has the advantages of fast detection speed and high success rate.