Research On Algorithm Of Safety Distance Detection Of Plug-in Component Of PCB Based On Machine Vision

The safety distance between electronic components of the printed circuit board(PCB)not only affects the quality of electronic equipment,but also concerns the safety of consumers.With the increasing integration and miniaturization of electronic products,the traditional detection methods cannot meet the real-time and reliability requirements of industrial production,and there are few studies on the application of AOI(Automated Optical Inspection)to the safety distance detection.Therefore,setting up the safety distance detection environment of plug-in components,researching of image location algorithm and the security distance detection algorithm of PCB in this paper is of great significance.The paper aims to research on image location of PCB and safety distance detection of the component.First of all,the basic working principle of AOI detection system and its research status at home and abroad are introduced,the hardware system is built according to the specific situation of this paper,and the lighting module and imaging module are analyzed in detail and the parameters of lighting,camera and other components are selected.At the same time,the preprocessing algorithm of image is studied deeply,and the advantages and disadvantages of various preprocessing algorithms are compared and analyzed.The appropriate algorithm is selected finally for enhancing the image,filtering,threshold segmentation and edge detection to facilitate the subsequent registration and electronic component detection.Aiming at the problem that the PCB image cannot be positioned accurately and quickly,an image localization algorithm based on improved Harris and edge detection is proposed in this paper.Firstly,the judgment of the Maximum difference between the neighbor pixels is used to eliminate part of the non-corner points and edge detection is performed with the LOG operator;then,an adaptive threshold corner extraction algorithm is used to extract the corners of the edge,and then the sub-pixel location method is adopted to reduce the accumulation error;finally,the NCC(Normalized cross correlation)algorithm and RANSAC(Random Sample Consensus)algorithm are used to perform rough matching and purify mismatching of the sub-pixel corner points to achieve accurate registration of PCB images.In the end,the paper analyzes and categorizes the characteristics of electronic components,and uses different algorithms to detect different electronic components.On this basis,the paper also proposes a K quadrant division method to obtain the nearest neighbor of the target element,establish the coordinate system,obtain the vertex coordinates of the external rectangle of the component,so as to calculate the distance of the electronic component which needs to carry on the safety specification judgment and compare with the standard safety distance,statistics related information of the electronic components which do not meet the safety standards.In this paper,the software system is developed in combination with OpenCV2.4.11 in VS2013.The experimental results show that the PCB positioning algorithm is effective to locate the PCB which has some of rotation and translation fast and accurately with the positioning error of less than 0.5pix,and the average detection time is less than 32.4%of the traditional Harris positioning algorithm.Meanwhile,the K quadrant division method proposed in this paper can eliminate the non-neighbor components effectively and improve the system operating efficiency greatly.Through the analysis of the relative position of the components,the distance detection and comparison of the components are realized and the results are displayed in real time.