Research On High-precision Machine Vision Measurement

Dimensional measurement has been the critical process in industrial manufacture.With the development of industry measuring technique in our country,machine vision technology is being more and more widely used in industry,which specifies that it is a new prospect in industrial manufacture nowadays.High-precision machine vision measurement system is mainly composed of illumination system,imaging system and data processing system,using computer to deal with the image and then process the accurate dimension calculate of the test object.The dimension data can be the criteria to decide whether the test object will satisfy the industrial requirement.Applying machine vision technique to process dimensional measurement has the advantages of easy-operation,high-efficiency and economic.However,the precision of machine vision measurement is always restricted by the imaging,illumination and algorithm.This paper focus on high-precision dimensional measurement,designed and built an machine vision dimensional measurement system,and a high-precision edge detection algorithm is proposed.Experiments are designed for discuss the efficiency and accuracy of the machine vision system and the algorithm.Firstly,the paper focus on increasing the imaging quality,discuss the advantages and disadvantages of different illumination system and imaging system,and then designed the illumination system and imaging system which can satisfy the high-precision demand of dimensional measurement.After deciding the total hardware scheme,this paper studied several edge detection and dimensional measurement algorithm and proposed a sub-pixel accuracy measurement algorithm called intensity integration threshold.This measurement scheme firstly searching for the edges by Sobel operator,and then extract the intensity nearby the edge pixels,calculate the integration and figure out the accurate sub-pixel edge by the threshold.After calculating the sub-pixel edges,by least-square fitting the geometrical edges,the sub-pixel accuracy dimension of the test object will be obtained.Finally,used the machine vision system to experiment,took the picture of the standard gauge and inner diameter gauge,processed the images in Matlab,and obtained the edges and the dimensions of each test object in each picture.Evaluated the accuracy and efficiency of the whole machine vision system and elaborated the advantages of the machine vision dimensional measurement system and the proposed algorithm.Experimental results shows that the proposed intensity integration threshold algorithm achieves higher efficiency and accuracy which can meet the application of high-accuracy dimensional measurement.