Development Of Multi-parameter Detector Of Electrical Connector Shell Based On Machine Vision

The electrical connector shell has excellent advantages such as sealing,bending,air tightness,which plays a pivotal role in the military and civil areas.The quality defects and inefficiencies of the electrical connector shell seriously affect the output of the electrical connector shell.In the past,the artificial sampling method was adopted,which has low efficiency and is easy to miss detection,is unable to meet the needs of modern society on product quality.Thus,an effective method to achieve efficient and rapid detection of electrical connector shell is in need.In order to achieve this gorl,in this paper,development of multi-parameter detector of electrical connector shell based on the machine vision theory.In order to meet the practical application,a series of theoretical research and experimental analysis are carried out from the aspects of hardware design,image preprocessing,edge detection,software development and testing.The main contents of this paper are as follows:1.The hardware design of the multi-parameter electrical connector shell detector,which includes the optimal selection of the visual hardware and the overall mechanical structure of the detector.In order to obtain a clear picture,the parameters of the light source,lighting system,camera and lens are analyzed.On this basis,the finite element analysis and appearance design of the detector are carried out.2.Image preprocessing of multi-parameter electrical connector shell detector.In order to obtain the size measurement and the high-quality picture of the part surface,in the aspect of the algorithm,the image filtering and image enhancement algorithm were adopted to improves the quality of the image.The template matching algorithm reduces the time of searching for the target object and improves the detection efficiency of the detector.3.Research and analysis of edge extraction algorithm for electrical connector shells.Based on the classical edge extraction algorithm theory,the edge extraction algorithm of the electrical connector shell is analyzed.Considering the advantages and disadvantages of the classical edge extraction algorithm,Canny edge extraction algorithm are selected to conduct the rough positioning.The focuses are on the sub-pixel edge detection algorithm and the Zernike moments theory based on the local area gray-scale effect,its advantages and disadvantages are analyzed respectively;Through the eigenvalues of the covariance matrix,the linear and circular discriminates,respectively,through the sub-pixel detection algorithm experimental verification.4.Software design of multi-parameter electrical connector shell detector.Based on VS2010 software platform,a multi-parameter electrical connector shell detector detection algorithm application program can be developed to detect many types of electrical connector shell.This application can achieve the host computer communication,pictures collection,size measurement and surface defect detection.At the same time,the camera imaging model,camera calibration method and calibration are completed.The experimental results showed that the detector can meet our requirements,with detection speed of 2 seconds per object,dimensional tolerance of 0.02 mm and 100% detection rate of surface defect.