Research On The Key Technology Of PCB Board Inspection Based On Machine Vision

In recent years,the development of electronic manufacturing industry has already entered the fast lane.Printed circuit board(PCB)plays an increasingly important role in industrial production,and product defect testing has become a part of the electronic industry.At present,the integration and precision of circuit board products are becoming higher and higher.Traditional manual inspection means that the accuracy and speed of product inspection are no longer consistent.Automatic and intelligent(AOI)system arises at the historic moment.This paper discusses the main technology of detecting PCB based on machine vision,and designs an optical detection system.which can detect circuit defects of PCB board and polarity direction of capacitor,diode and other components in real time.Based on the purpose to achieve,this paper analyzes the functions of the system and builds a set of effective hardware and software system.Through the analysis of the experimental results,it shows that the system has a short detection time,high accuracy and low false alarm rate,which can well meet the needs of detecting PCB board.The main research contents of this document are as follows:1.This paper briefly introduces the domestic and foreign researches.Through the functional analysis of the system,suitable industrial cameras,light sources,lenses and shadowless lamp boxes are selected to build the external hardware system.The software system is programmed by Lab VIEW and MATLAB.The first is the image acquisition and splicing module,which ensures the uniform light of the external environment and the small shadowing area of components.The camera parameters are configured with Lab VIEW for image acquisition,and the collected images can be saved in a folder and called at any time.2.Three correlative algorithms of image preprocessing are studied.On this basis,for PCB with large area,two image Mosaic algorithms based on gray template matching and Harris corner feature are studied and compared.Finally,the optimized Image Mosaic algorithm based on Harris corner feature is selected as the Mosaic method of the system.3.The image sharpening and morphological processing related algorithms are studied,Proposal for a method for detecting defects in PCB circuits;With regard to the objectives and the context,the method of dividing up by thresholds has been used,and then the binary image template matching method is used to determine the capacitance polarity.The hole and coordinate of the chip are located and the welding direction of the chip is judged by Hough transform.The polarity of the diode is judged by an algorithm based on quadrant segmentation.4.Finally,a certain number of circuit boards with different types of defects were repeatedly tested for defects,and the experimental results were statistically analyzed to summarize the causes of false positives and the direction of future improvement.The final experiment proves that the detection speed of the system is about 900 ms and the accuracy is above 95%,which can meet the relevant needs.