Small Bearing Inspection System Based On Machine Vision

For the factory inspection of small needle roller bearings,the main inspection items are missing needle,notch defect detection and size error detection.The detection efficiency and accuracy of the current manual or semi-automatic detection methods cannot be guaranteed.Therefore,this paper has carried out a research on the needle roller bearing detection system based on machine vision.Compared with traditional detection methods,the system has the advantages of non-contact,fast speed,good accuracy,and strong on-site anti-interference ability,which can well meet the detection needs of modern manufacturing.The specific research content is as follows:(1)Built a machine vision hardware system.Through the evaluation and analysis of the detection object,combined with the company's requirements for detection accuracy,the selection of CMOS sensor cameras,high-resolution industrial lenses,backlights,proximity sensors and UP Squared processing equipment,builds a machine vision detection system.This platform is Needle bearing inspection provides hardware support.(2)Use the HALCON toolbox to calibrate the camera,determine the internal and external parameter matrix of the camera.and transform image coordinates to world coordinates.Perform image preprocessing,including using median filtering and gray linear transformation to denoise and enhance the it.Considering the flaw of traditional Ostu method being easily affected by noise,the image segmentation effect is improved by improving the Ostu method.This method adds weight when obtaining the optimal threshold,which can effectively distinguish the target area from the background area,thereby providing subsequent segmentation Work provides a better environment.(3)Design and verification of bearing inspection algorithm.First,use Blob analysis to roughly locate the image,and combine the subtraction method with morphological analysis to complete the extraction of the two areas of interest for the inner and outer roller needles and gap defects of the bearing;secondly,use the template matching algorithm to complete the preliminary needle missing determination.And innovatively add the step of detecting missing needles outside the bearing to further improve the detection accuracy of missing needle defects;further,set the threshold according to the geometric feature description to complete the detection of notch defects;ultimatly,in combination with edge extraction and circle fitting outer diameter measurements,the algorithm was experimentally validated to meet detection requirements,demonstrating the accuracy and feasibility of the algorithm.(4)Developed a needle roller bearing inspection system based on machine vision.Based on a comprehensive investigation of user needs,Combined with the HALCON development kit,based on the MFC framework,using VS2015 to complete the visual development of the needle bearing inspection system.In addition,taking into account the execution speed of the program,etc.multithreading and dynamic link library technology are incorporated.Realized the call of the database and the five main functions of the test result picture display,standard parameter display,current result output,operation,and statistical result display.In the experimental stage,the three detection accuracy of missing needle detection,gap defect detection and size error detection are all higher than 95%,which demonstrates the superiority of machine detection.Through the self-developed software,this paper realizes the automatic detection of needle defects,gap defects,dimensions and other parameters of needle roller bearings,which provides a useful exploration for solving the automatic detection of needle roller bearings,and is a project for machine vision inspection.The application provides ideas and solutions.