Research On The Method Of Mark Point Positioning Based On Image Matching

From silicon wafer positioning to PCB chip welding alignment,it is necessary to use visual alignment system to locate the mark point of devices to ensure the correct production process in the production of electronic products.With the more and more functions of electronic products,more and more chip devices on the circuit board put forward higher requirements for the performance of the visual alignment system,and the performance of the visual alignment system largely depends on the performance of the mark point positioning algorithm.At present,the method of image matching is commonly used to locate the mark point in industrial production.In this method,a mark point image is used as a template image,and an image captured by an industrial camera in real time is used as a target image.The position of the template image is located at the target image,and the coordinate of the mark point in the target image is calculated based on the coordinate of the mark point in template image.According to the industrial production process,this dissertation studies the algorithm of the mark point positioning under standard and abnormal conditions.Under the standard condition,the mark point graphic in the target image is complete.In this scenario,this dissertation focuses on improving the efficiency of the image matching algorithm.The LINE2D algorithm based on similarity measure calculates the gradient direction of each pixel,and quantizes and codes them to get response maps.Finally,it stores response maps linearly to realize fast image matching.However,this algorithm only uses the gradient direction for image matching,which is easily affected by background interference.In order to achieve fast mark point positioning,this dissertation proposes the LINEG algorithm based on the LINE2D algorithm.The image gradient information and the relative gray are used to describe the edge contours and the difference between the inside and outside of the mark point,so as to achieve accurate matching while ensuring the efficiency of the algorithm.Through experimental tests,LINEG algorithm execution time is reduced by 52%compared with the current visual alignment system algorithm under the standard condition.Under the abnormal condition,the incomplete mark point in the target image will affect the calculation of the similarity measure of the LINEG algorithm and lead to the failure of matching.In view of the above problems,this dissertation combines the characteristics of the mark point and uses the method of line feature matching to construct an image matching algorithm.The line feature descriptor is the core of line feature matching.Aiming at the problem that the BOLD of different line segments of the centrally symmetric image are the same,it is proposed to use the absolute position of the line segments to improve the calculation method of the geometric primitives of the BOLD.Aiming at the problem that the accuracy of the BOLD is reduced when the mark point is partially missing,the BLBD is constructed by integrating the LBD,and the line characteristics are matched by describing the relationship between the line segment and the surrounding line segments and the characteristics of the line segment itself.Considering the stability of the slope of the line segment extracted from the image,it is proposed to use the intersection of the line where the line segment is located as the matching point of the template image and the target image,so that the homography matrix of the template image to the target image can be calculated to achieve image matching.Through experimental tests,the effectiveness of the algorithm for mark point positioning under the abnormal condition is verified.