Design And Implementation Of HCI-A3 PCB Board Defect Detection System

With the rapid development of the PCB industry,higher requirements have been placed on PCB inspection technology.The PCB Defect Automated Optical Inspection System(AOI)is based on image processing to detect and process defects encountered during production.Practice shows that this method has high accuracy and practicability.Many enterprises at home and abroad have developed products and put them into use.However,in the face of the expensive foreign testing equipment and the demand for domestic PCB automatic testing equipment,it is of great economic value to study the economical and practical PCB defect detection system.In this paper,PCB bare board is taken as the research object,and its typical defects are identified and classified to improve PCB production efficiency.This paper draws on the domestic and international related industry products,designs and implements a set of PCB defect detection system,analyzes the key technologies of the system,and completes the realization of the main functional modules.The system mainly consists of two parts: the hardware system consists of three parts: lighting unit,image acquisition unit and motion control unit.The software system includes several modules: system initialization,standard image acquisition,image acquisition and preprocessing,defect identification and classification.On the one hand,the software system implements control of the lower computer,and on the other hand,image processing and defect detection.The specific research contents and results of this topic are as follows:1.Design a set of PCB defect detection system,complete the selection of main components,build a system hardware work platform,and determine the image acquisition scheme in combination with the hardware environment.2.On the basis of analyzing the grammatical structure of Gerber file,firstly extract the information through regular expression,and design the corresponding class to save the extracted primitive information and path information,and then use GDI+ to complete the standard image drawing,and give the drawing process.The solution to the pixel coordinate mismatch problem that occurs in it.3.The core problem of autofocus technology,namely the selection of image sharpness evaluation function,is studied.The resolution analysis function of common spatial image is analyzed from the perspective of amplitude-frequency response.The Brenner algorithm is better verified by experiments.4.According to the characteristics of PCB image,the image preprocessing scheme is de-noising,distortion correction and image enhancement.Based on the experiments of common algorithms in each step,the image denoising field has been used.Widely applied adaptive total variation model for PCB image denoising;for the barrel distortion of the image,using bilinear interpolation for grayscale reconstruction,the distortion becomes effective correction;An image enhancement method that combines redundancy with local mean square error to stretch the grayscale range while enhancing contrast.5.The registration algorithm based on concentric circles is used to analyze the principle of traditional Hough detection,and then the improved algorithm is tested.The running target is marked by the run length coding algorithm.Finally,the characteristic parameters such as the connected area number,the binary image area and the out-of-line pixel value of the defect target are statistically analyzed,and the defect types are classified by multiple defect features.6.Developed a PCB defect detection software system.Through testing the main functions of the system,it shows that the main module functions of the system are completed.The hardware cost of the system is relatively low,the detection result is relatively stable,the software interface is simple and easy to operate,and meets the actual needs of the Chinese market.