Research On Dimension Measurement Method Of Non Telecentric System Based On Machine Vision

With the development of social economy,my country's manufacturing industry has been greatly improved,and part size measurement is an indispensable part of the manufacturing industry.Traditional size measurement relies on manual labor,which has problems such as large errors,low efficiency and high cost.Measurement methods based on machine vision are widely used in various industrial fields due to their advantages of high precision,low time consumption,and real-time data acquisition.In order to pursue distortion-free image quality,some vision measurement projects use telecentric lenses.However,telecentric lenses have problems such as high cost,large size,and heavy weight,which directly lead to high production costs.Aiming at this problem,this paper proposes a non-telecentric system measurement method based on machine vision.This method uses Huarui MK1228 M non-telecentric lens to correct the distorted image through the image distortion correction algorithm,and proposes a sub-pixel edge detection algorithm based on Sobel combined with the integral median method to achieve high-precision and high-efficiency measurement.cost,providing a new idea for industrial measurement.Since the standard gauge block has a specific standard value,therefore,standard gauge blocks are used in this paper for experimental measurements,which is beneficial to verify the feasibility of the measurement method in this paper.The specific research contents are as follows:(1)A set of measurement hardware system is designed.The system uses a backlight and a front spherical integral light source.The backlight makes the contour image of the target gauge block clearer,and the image contrast is large.The spherical integral light source can make the frontal illumination uniform,so as to obtain a good calibration image for camera calibration.In addition,on this basis,the camera and lens parameters were selected.(2)The causes of image distortion are introduced,and the calibration algorithm is used to correct the image distortion.Due to the use of ordinary industrial lenses,the image pixels will be centered on the origin of the pixel coordinate system,and deviations will occur along the radial position,resulting in radial distortion.In this paper,the calibrated internal parameters of the camera will be used to correct the radial distortion of the image;not only that,because the optical axis of the camera cannot be absolutely perpendicular to the surface of the object to be measured,so the measured image also has tangential distortion.This paper uses a calibration plate.The method of fitting on the target gauge block calculates the posture of the gauge block and maps it to the measured standard plane to realize the tangential distortion correction of the image.(3)After analyzing the traditional sub-pixel edge detection algorithm,this paper studies and implements an efficient sub-pixel edge detection algorithm.The algorithm is based on Sobel pixel-level edge detection.After obtaining the pixel-level edge position,it continues to integrate the edge gray gradient direction of the image,and takes the position of the coordinate point corresponding to the median value of the integration as its sub-pixel coordinate,and then connects all the The sub-pixel points form a sub-pixel contour,which is fitted using the least squares method,and finally the size is measured according to the obtained sub-pixel contour.(4)The proposed sub-pixel algorithm is analyzed in terms of accuracy and timeliness,compared with other two traditional sub-pixel algorithms,the accuracy and time efficiency of the algorithm are evaluated,and a set of software system interface is designed.The experimental results show that the accuracy of the algorithm can reach 0.02 mm,and the time consumption is only 498 ms,which has good application value.