| In industrial manufacturing,it is necessary to measure the position and size of industrial products.Industrial development makes the target to be detected refined,so the accuracy of these parts detection is also higher and higher.In high-precision detection,the most direct way to detect traditional industrial parts is to observe them through an industrial microscope.But this method will increase labor costs,and different people may have different results,which lacks stability.More importantly,it is necessary to have enough detection speed while meeting the detection accuracy.The human eye can only observe one part of the part at a time,and it is difficult to complete the detection work efficiently.Therefore,this paper uses the machine vision system based on image processing technology to measure the specific position of industrial parts.In this paper,I mainly study the sub-pixel edge detection technology and center positioning technology,and standard circle and the spinneret of the parts actually produced by the factory are taken as test objects.The main research contents of this paper include the following aspects:(1)In this paper,I propose a subpixel edge detection method based on curvelet transform and local effect.In this algorithm,the edge of the curve is detected.Firstly,edge analysis is performed through Wrapping curvelet transformation.After the original image is transformed,a set of curve coefficients is obtained.A threshold is set to set the smaller coefficients to zero,and then the coefficients are inverted to obtain a new image;Then the local effect edge detection is applied to the new image,and the subpixel coordinates of the edge points are calculated according to the pixel points in the edge neighborhood;Finally,the edge detection results in standard circle and micro hole prove the superiority of the algorithm in edge detection.(2)Aiming at the problem of inaccurate center location caused by noise point interference,this paper proposes an edge point filtering method,sub-pixel edge detection is combined with center location.Firstly,the subpixel edge detection algorithm is used to locate the position of edge points in the image;Then judge from two aspects: the distance between two points on the same circle is less than the diameter,and the gradient angle between adjacent points is acute,edge points of the same circle is selected;Finally,the least squares method is used to determine the position of the center in the circle.Through experiments,we compared the current more accurate methods of locating the center of a circle,and the experimental results showed that this algorithm has high detection accuracy and anti-noise ability.(3)This article establishes a visual measurement platform and a visual detection system.In actual micro hole measurement,the center positioning method based on subpixel edge detection in this paper can achieve an accuracy of about 0.02 mm. |
