Research On The Positioning System Of Wafer Based On Machine Vision

With the progress of semiconductor manufacturing technology andburgeoning demand for tiny chips in the electronic market, the size of thechip has been decreasing as well as pad diameter and the distance betweenchip leads. The traditional encapsulation equipment has already beenunable to meet the increasingly severe requirements. As a result, visualpositioning technology gradually becomes key technology to guarantee thehigh precision of encapsulation equipment. A positioning system of waferbased on machine vision is described in this paper, which aims at thereal-time location and alignment of the chip. The paper mainly includesfollowing four parts.1. Visual system and motion control system were designed to meetthe specific requirements of the positioning system of wafer.2. The paper done some research in the camera calibration technique.The imaging model was analysed in detail as well as the coordinatetransforming relations. A linear model was applied to calibrate the visualsystem with telecentric lens3. Positioning algorithm was researched in this paper. In the algorithm,background subtraction method was used to finish image preprocessing,and then Canny operator was used to detect the edges in the image. Theconnected regions in the edge-extracted image, which corresponded todifferent positioning marks, were labeled according to the number ofpixels. A new algorithm, which combined the Hough Transform and theleast square fitting method, was proposed to detect lines in the separatedimages. Then the position of the mark in the image could be calculated.Finally, the displacement and angle needed in the alignment could also be calculated based on the positioning information.4. The interactive interface of motion control system was designed.The positioning system of wafer was able to implement its function byadopting the positioning algorithm to guide the motion control system.