Research On PCB Defect Detection Technology Based On Machine Vision

Printed Circuit Board(PCB)is the support and electrical connection carrier of electronic devices,which directly affects product quality.With the rapid development of modern household appliances,automobiles,aviation,national defense and other industries,the traditional artificial PCB defect detection methods can no longer meet the demand.This paper has carried out the research of PCB board defect detection based on machine vision,and realized the detection of several typical defects.The main work of the paper includes:(1)A PCB defect detection platform is built.According to the project requirements,a visual inspection hardware platform including camera and lens.is built.In this paper,image acquisition and processing software is designed to realize the functions of PCB image preprocessing,positioning correction,component ROI frame generation,and component defect detection.The research platform of the thesis consists of software and hardware platforms.(2)This paper studies the image preprocessing method and analyzes the basic principles of median filtering,piecewise linear transformation enhancement,Otsu,Canny edge detection and other methods.These methods are used to denoise,enhance,binarize,and edge detection on the acquired PCB board images.Finally,the results of image processing are compared and analyzed.(3)This paper studies the correction and positioning methods of PCB board images.In order to solve the problem of crooked PCB image,the area moment method is used to detect the PCB angle and offset.According to the detected angle and offset,affine transformation is used for image correction.In this paper,firstly,the circular contour detection is performed on the PCB board image,the circumference and area of the circular contour are calculated,and feature screening is performed to determine the positioning circle area of the PCB board.Then,the least square method is used to perform circle fitting to obtain circle parameters,so as to achieve precise positioning of the PCB board positioning circle.(4)This paper studies the detection methods of PCB board defects.According to the obtained positioning circle information and the PCB component information file,an image of the ROI area of the PCB board component is obtained to realize the component positioning.This paper extracts the central area of ROI and calculates its average gray value to realize the detection of SMD capacitor missing soldering,analyzes the similarity of the standard deviation of the gray value in the central area of the ROI to realize the detection of SMD resistance missing soldering.Character recognition is performed by dividing the surface characters of the resistor.For the SMD resistors that are not placed horizontally,the image is rotated using affine transformation,and then the SMD resistors are recognized for characters to realize the detection of wrong welding of the SMD resistors.According to the color characteristic H component of the central area of the SMD capacitor,the SMD capacitor wrong welding detection is realized;the area moment judgment is used to realize the SMD capacitor and resistance offset detection.