Shape Detection And Accuracy Analysis Of Sheet Parts Based On Machine Vision

In industry,detection refers to the inspection and measurement of the object to be detected to obtain the parameters of the object to verify the manufacturing effect.With the continuous progress of industrial development,detection technology.The widespread application of various components in modern manufacturing has put forward higher requirements for inspection efficiency,accuracy,and automation.Visual inspection technology has the advantages of non-contact,high detection efficiency,low cost,and good flexibility,making it shine brightly in today’s manufacturing industry.The research and application of visual inspection technology is of great significance and practical value.Based on the existing theoretical knowledge of machine vision,combined with the common part detection environment,this thesis completes the detection requirements of target parts,and explores and studies the methods to enhance the real-time detection and measurement accuracy of relevant parts from the software aspect.The main research contents are as follows:(1)According to the functions and design principles of the measuring system,the hardware acquisition part of the visual inspection system is designed,the industrial camera,lens,light source and other equipment for image acquisition are selected and built;the software framework design of the detection system has been completed,the human-computer interaction interface of the detection system is completed through python language,Open CV library and Py Qt.(2)According to the image processing process,the preprocessing module composed of image graying,image filtering and image enhancement is determined.The functions of the preprocessing part are realized by various algorithms.By comparing and analyzing the processing effect,the median filter is selected as the filtering algorithm of the system,and the adaptive contrast histogram enhancement is selected as the enhancement algorithm of the system.(3)In the parameter processing module,four detection operators are studied,and Canny operator is selected as the detection operator of the system,and the operator is improved by replacing the filter algorithm and detection method,and determining the upper and lower thresholds by using the maximum between-class variance method.The processing effect of the improved algorithm is verified by the detection results;the contour extraction function and contour matching function of the image are completed;according to the characteristics of complex parts,the circumscribed rectangle is selected,and the circumscribed circle is used as the detection parameter to complete the minimum circumscribed circle extraction function.The minimum circumscribed rectangle is extracted by the convex hull rotation chuck method.(4)Analyze the error source of the detection system;the camera parameters are obtained through calibration,and the image distortion correction is completed;the pixel equivalent is calculated by detecting the standard gauge block several times,and the error compensation method is proposed for the pixel equivalent error to reduce the detection error;detect three parts with different shapes,and verify the detection accuracy of the detection system by comparing with the manual detection results;verify the system accuracy analysis function through the circlip parts.