Development Of Wafer Positioning And Cutting System Based On Machine Vision

With the rapid development of the semiconductor industry,the scale of the scribing machine market continues to expand,and the demand for wafers continues to grow.As the silicon wafer becomes thinner and the width of the cutting channel becomes narrower,the precision requirements of the wafer cutting equipment are gradually improved.The machining equipment with better performance and the visual positioning and cutting system with high precision,stability and reliability can improve the yield of the finished product of the chip on the wafer.At present,the use of the scribing machine in China is largely dependent on imports,and a set of wafer positioning and cutting system based on machine vision is developed to meet the needs of domestic industrial development.The main contents of this research are as follows:According to the technological requirements and the principle of machine vision positioning,a hardware platform for the vision positioning system is built.The hardware system includes the imaging system and the motion control system.Through the lighting experiment,the appropriate light source,camera,and lens were selected for the system,and the appropriate motion controller was selected according to the requirements of motion accuracy.The rough pendulum forward algorithm based on multiple template matching is studied,the edge detection is performed using Canny operator,the fine pendulum forward algorithm based on Hough transform is studied,the principle of SAD algorithm,NCC algorithm,template matching algorithm based on shape contour and the The execution process and experiments show that the template matching based on shape contours has better robustness.The specific implementation steps of the Hough straight line detection algorithm and the method of extracting the center coordinates of the cutting lane are elaborated in detail,and the detection results are processed.Calibrate the angle to complete the precise positioning of the wafer.At the same time,this paper also studies the wafer image preprocessing process,compares and analyzes different kinds of filter algorithms,and selects the appropriate filter algorithm.Motion control system are introduced in the axis motor and the drive way,detailed elaborated the motion controller and PC software communication mode and all kinds of sports instruction of the controller,realized the operation,the movement of the shaft and the Galil motion control card based trajectory planning,at the same time,this paper introduces the basic principle of serial port communication,communication between spindle and frequency converter,and the speed of the spindle control and status monitoring,and integrate them into the positioning cutting system software.The overall framework of the software is built in a modular way,and the human-computer interaction interface of each module and the realization of main functions are introduced in detail.According to the study of the theory of the above,wafer cutting positioning system based on machine vision has carried on the test,cutting accuracy has reached the expected goals.