Research On Defect Detection System Of A Part Based On Machine Vision

It is essential to detect the part defect in the production and processing of precision instruments and equipment.But nowadays,some traditional detection methods can no longer meet the needs of modern automatic detection.With the continuous development of automatic detection technology,the detection technology based on machine vision has been widely used.Taking a missile part as the research object,the part defect detection system based on machine vision is studied in this paper,aiming at resolving defects of a part in its whole diameter,inner bore quantity,position and area.Aiming at this type of part inspection,another improved fast template matching algorithm based on regular edges is designed.Finally,a part defect inspection system based on machine vision is established to realize automatic inspection of part defects.The details of the work of this article are as follows:Firstly,the hardware platform is set up and the whole process of part detection is designed.According to the needs of part defect inspection,the overall program of the part defect detection system is researched and designed,the system platform is modeled in three dimensions,and the hardware lectotype,design and construction of the system platform are completed.In addition,the related algorithm flow is also designed in detail.Secondly,the related image processing methods for this part are studied,including preprocessing,edge detection,morphological processing,ROI extraction,etc.After the analysis of the related image processing techniques,the related processing effect diagram of the part image is obtained,which provides preparation for the part size measurement and inner hole defect detection.Thirdly,the diameter of the part is measured.Through the research on the calibration coordinate system,linear model and distortion model of lens distortion,this paper completes the camera calibration,obtains the calibration results of camera initialization parameters,and corrects the original image.Then,this paper calibrates the measuring system to obtain the calibration coefficient k and studies the measuring algorithm and program,thereby completing the measurement of the part diameter.Fourthly,the number of inner holes,position and area of the part are detected.The connected domain of the part inner holes is marked by studying the connected domain in the image,and the number of the part inner holes is identified.Segmenting the irregular images of inner holes in the part inner holes,the positions of all inner holes in the images arc detected.In addition,an improved fast template matching method based on regular edges is applied to the research of such parts,which completes the rotation of part image and the production of matching template,and detects the area of inner hole images of the part.Finally,the analysis software of the part defect detection is established.With the GUI module of MATLAB software,the structure of four modules of the part defect detection software is designed,which includes user management,part image import and processing,part size measurement as well as part inner hole detection.Besides,each module is explained and its working principle is analyzed.Finally,the relevant experimental verification of the part defect inspection system designed in this paper is carried out.The actual results show that the system has high accuracy and fast detection speed for the defect inspection of such parts,and meets the defect detection requirements of such parts.It also has certain engineering application value.