The Research And Realization On Precision Measurement System Of Mobile End Plug Based On Machine Vision

Part geometry measurement technology is an indispensable part of manufacturing technology.In industrial production,in order to ensure the quality of the product,it is necessary to measure its processing size efficiently and accurately.However,the traditional manual measurement methods have the disadvantages of high labor intensity,low detection efficiency,high production cost and can not guarantee the overall quality of products.With the advent of the fourth industrial revolution,industrial production is moving towards a more intelligent and automated direction.Machine vision inspection technology has the advantages of real-time on-line,non-contact,fast speed,high sensitivity and high accuracy.It not only saves time and cost,but also avoids man-made errors in the measurement process,and improves the level of industrial automation.At present,precision measurement technology based on machine vision has become a hot topic in industry.In this thesis,the precision measurement technology of small irregular shape workpiece based on machine vision is studied and applied.Taking the mobile phone tail plug-in workpiece as the detection object,firstly,the image of the workpiece is collected by CCD industrial camera.Through image enhancement,image registration,image edge detection and target line extraction,the pixel size of each part of the mobile phone tail plug-in is calculated.Finally,the physical size of the workpiece is determined by camera calibration technology.In image registration,the problem of low matching accuracy faced by current SIFT feature matching algorithm is analyzed.Aiming at the problem that SIFT matching algorithm completely ignores the geometric relationship between different feature points and only pays attention to local optimization,and is prone to more mismatches when searching for matching feature points in workpiece images with smoother gray changes,an improved image registration algorithm with constrained SIFT feature matching is proposed.Firstly,the object external contour is extracted by image topological structure analysis method based on boundary tracking,and minimum circumscribed circle of the contour is fitted.The rough geometric transformation information between two images is obtained by ellipse matching.Secondly,the SIFT feature point matching is constrained by geometric information to improve the matching accuracy.Finally,the RANSAC algorithm is used to accurately match the feature points,determine the transformation matrix and transform the image registration.Experiments show that the proposed algorithm improves the accuracy of feature point matching by 12% compared with that before the improvement,and effectively improves the accuracy of registration.In line detection,a new voting mechanism is proposed to suppress false peaks in Hough space by changing the voting weights of Hough space in order to accurately detect the edge lines of images.Experiments show that the proposed algorithm improves the accuracy of line detection by 22% compared with the improved algorithm.In this thesis,a large number of experiments have been carried out on the tail plug-in parts of mobile phones.The experimental results show that the proposed algorithm has higher feature point matching accuracy and line detection accuracy than before.The final measurement result of the visual measurement system has high robustness,and the measurement accuracy reaches 0.0158 mm.It can meet the requirements of industrial production detection and has a certain degree of versatility.