Research On Edge Alignment System Of AOI Defection Equiment Based On Image Processing

This paper is based on the vision system,motion system and pneumatic system in the AOI wafer inspection equipment for edge alignment research.This paper designs the edge alignment scheme and completes the overall structure design.A recipe model is designed for wafer preprocessing,and image processing algorithm is used to locate the wafer center and notch.Finally,an experimental scheme for wafer edge alignment is designed to verify the accuracy of the system.The work of the thesis is as follows:1.Research on wafer edge alignment system.The analysis compares the currently commonly used edge alignment schemes and processes,and adds a wafer preprocessing process to handle edge alignment,thereby saving time in the workflow.The vision system is the key to wafer imaging.According to the system design index,2X lens,area CCD camera and coaxial light source were selected,and the structure of the vision system was determined.The motion table system consists of two parts,the drive system and the pneumatic system,and adopts the motion table scheme with three degrees of freedom of XY-(?).The movement in the X and Y axis directions is completed by the combination of the stepping motor and the subdivision driver,and the wafer is adsorbed on the splicing column by the pneumatic system.2.The process scheme based on the edge alignment of the first wafer in the cassette.The recipe model of edge alignment process is designed,which standardizes the parameters involved in edge alignment.During edge alignment,the first wafer in the cassette is transferred to the motion table for edge alignment.After the edge alignment is completed,the alignment results,algorithm parameters and vision system parameters are written into the Recipe model,which completes the class.The recipe model of the wafer is established.The other 24 wafers in the cassette can directly use the recipe model,saving a lot of time for edge alignment.Subsequent alignment,defect detection and other processes can directly obtain the required parameter values from the recipe model.Through experimental comparison and analysis,the yield is increased by about 12.5% compared with the current process.3.Image algorithm and mathematical model design.The algorithm design of the two parts of the circle center positioning and the gap positioning is completed.The image algorithm is used to obtain the pixel coordinates of the edge points,and the pixel coordinates are converted into physical coordinates on the motion stage through the mathematical model,that is,the positioning of the center of the circle is completed.The image algorithm is used to fill up the wafer gap image,and make a complement with the initially identified gap image,and obtain a smooth image of the gap through opening and closing operations.The gap center can be obtained by the maximum value method and the mean value method.The processing time of the maximum value method is 10 ms,and the processing time of the optimized mean value method is 2ms.After building the experimental platform,the debugging experiment was carried out,and the 200 mm and 300 mm wafer edge alignment experiments were designed.It is verified by experiments that the positioning accuracy of the edge alignment algorithm reaches±0.01 mm,the center positioning accuracy reaches ±0.01°,and the edge alignment time is2.3s.