Research On PCB Component Detection Method Based On Machine Vision

With the rapid development of science and technology,electronic products tend to be miniaturized and high-density,and the output increases year by year.Many companies have introduced assembly-line patch machines and welding machines to replace the traditional manual patch components and welding work.However,component patch errors still occur from time to time such as component errors,patch orientation error,location error etc.This paper analyzes the problems that may occur in the process of PCB component mounting,and focuses on the PCB component detection algorithm based on machine vision with the premise of low cost.This paper combines digital image processing technology,uses Python programming,takes raspberry pie as hardware platform,collects PCB component images,obtains the correctness of the PCB components through detection algorithm,and finally the detection of the components is realized automatically.The main work is as follows:(1)Hardware platform: Analyse the principle of image acquisition and acquisition device,determine the image acquisition device and the main controller platform;(2)Make of ROI(Region of Interest)frame for components:The PCB coordinate file is read,and the physical coordinates,dimensions and rotation angles of components are obtained by regular expressions.Hough transform is used to detect the positioning circle,and according to the center of the positioning circle,the image is corrected by projective transform,then the transformation formula of physical coordinates and pixel coordinates is found out,and the ROI frame can be automatically made according to the pixel information of components.(3)Component detection:extracting the color information of components to be tested and comparing with the corresponding template to determine whether the components are missing;extracting the corner information of components and matching the corner of the template to determine the correctness and error of components;for the polarity detection of patchelectrolytic capacitors,this paper proposes a method to determine the polarity of capacitors by image quadrant segmentation and calculating the area;it also proposes a method to extract and recognize the characters on the surface of capacitors by image segmentation;in this paper,affine transformation is used to detect patch resistance characters in four directions,and the resistance's value is detected.(4)Testing and verification : The verification results of this paper are as follows:PCB board image positioning and component ROI frame automatic production,PCB component existence detection;template matching experiment of components;patch polarity capacitance polarity detection and character recognition,and the surface character detection of patch resistance.