Research On Alignment System Of High Precitsion Based On Dual Microscopic Vision

In the field of intelligent manufacturing and precision machining,the process to realize the precise positioning and align of components is an indispensable step.The high-precision automatic alignment system plays an important role in improving the production efficiency and machining accuracy,and is the key of efficiency and quality.With the development of micron and sub-micron technology in precision electronics industry and micro-nano manufacturing industry,the pin size of electronic components and the width of etched lines are reducing the micro-nano level,meanwhile the size of the corresponding basal workpieces such as circuit board and photolithography mask are increasing.Therefore,it has become a critical problem in the field of precision machining and inspection site about how to achieve the cross-scale high-precision positioning and alignment of the basal workpiece.Combined with the requirements for the cross-scale automatic alignment of photolithography mask,a high-precision alignment system based on dual micro-vision was designed in this paper.The pose detection algorithm for workpiece based on local information,the method of system calibration and integration,the mechanical structure of precision motion platform and the alignment strategy were researched.The experiment with the performance of the visual alignment system were completed and the cross-scale precise alignment of the photolithography mask was accomplished.The main contents of this dissertation areFirstly,considering the requirements of accuracy and the difficulties of cross-scale alignment in the automatic alignment of photolithography mask,a scheme based on dual micro-vision and precision motion control technology was proposed.The basic process of automatic alignment and the functional requirements of vision module and motion control module are analyzed.A novel three degrees of freedom(DOF)motion platform based on the generalized parallel decoupling mechanism is designed.Parameter selection of the main hardware in the system was studied,and the experiment platform of high-precision alignment system based on dual micro-vision was builtNext,in the research of pose detection algorithm based on dual micro-vision,this paper presents a novel approach to derive the overall pose deviation of the photolithography mask,which combines the left and right local pose information with the geometric dimensions of the workpiece.The algorithm mainly includes three steps:image preprocessing,subpixel edge detection and pose solving.In the image preprocessing,the minimum ROI extraction algorithm is designed to reduce the time of feature detection,and the adaptive smoothing filter algorithm is adopted to solve the edge details fuzzy problem.An improved Canny algorithm was proposed and the sub-pixel edge profile extraction was completed based on polynomial interpolation in order to improve the accuracy of edge positioning.The discrete edge points are fitted by the RANSAC algorithm to obtain the center point and the local position deviation.The mathematical relationship between the two local position information and the pose deviation of microchannel was derived according to the geometric dimensions of photolithography mask.Accuracy experiment of visual inspection show that the algorithm has a distance detection accuracy reaching 1.93?m and an angle extraction accuracy reaching 0.018°for parallel lines.Then,the calibration and integration of the vision system and the motion control system were studied.Internal-parameters of the left camera and the right camera were calibrated thereby the image distortion correction can be completed.Relationship between the two pixel coordinate system were derived and then calculated by designing high-precision plane target.The integration method of vision system and motion control system was analyzed,and the coordinate transformation relationship between the pixel coordinate system of left camera and the motion coordinate system was obtained using nine-point calibration method.A platform movement strategy based on micro-vision and an alignment accuracy discrimination method based on local deviation were presented.The experimental results show that the average calibration error of the left camera is 0.147913 pixel,the average calibration error of the right camera is 0.137758 pixel,the standard deviation of coordinate transformation between visual coordinate and motion coordinate is 0.185 pixel in X direction and 0.168 pixel in Y directionFinally,the performance of the alignment system based on dual micro-vision was verified.Experimental results show that:the repetitive positioning accuracy of the novel 3-DOF platform based on generalized parallel decoupling mechanism is 3.21 ?m in X direction and 2.98?m in Y direction,and the coupling degree of X-axis and Y-axis is minimal,which can be neglected.For the cross-scale alignment experiment of photolithography mask,results show that the rotation alignment error of the designed visual alignment system is 0.01547°,the lateral offset alignment accuracy of the left and right ends is 3.61?m and 3.89?m,the accuracy of end-surface clearance is 3.76?m,the syetem can meet the precision requirement of basal workpiece alignment.