Based On Image Processing For The Design And Implementation Of Real-time Monitoring Wafer Curvature

The current commercial real-time measuring instrument cannot be used formeasuring the curvature of Silicon grid graphics substrate, this paper developed anequipment for real-time monitoring curvature of epitaxial wafer in order to meet thedemand that is to obtain real-time stress changes of GaN epitaxial wafer of SiliconGraphics during its development.This paper first describes overseas and domestic research status aboutmonitoring devices of epitaxial wafer curvature and makes analysis about real-timemonitoring devices of epitaxial wafer curvature including the principle of epitaxialwafer curvature monitoring. At the same time, this paper introduce the overallstructure design of the system in details including the designs of high-speed cameraoperation mode, the structure of lower machine as well as the overall framework ofthe upper computer software and so on. In the design of lower computer structure, itintroduces the design of the PCB board and control programs. In design of overallframework of the upper computer software, it describes the design of MFCdevelopment mode and access database. This paper adopted median filter to smoothimage and Canny operator for image segmentation in image processing and analysisand then derived target spot center after connectivity domain processing throughcomparing different ways of algorithm. Finally, this paper describes practicalapplication in epitaxial experiments and production for real-time monitoring systemof epitaxial wafer curvature. By analyzing the result, the measurement resolution ofthe system can reach±1km-1, the measurement accuracy is higher than the similarcommercial equipment’s. By analyzing the test results in Swan CCS MOCVD, thissystem can display the stress changes of Silicon grid graphics substrate duringdevelopment in real-time, it is conducive to optimize the growth process of theepitaxial wafer and improve thickness uniformity and crystals quality of epitaxialwafers.This paper’s completion was based on the author’s participation in research anddevelopment of the real-time detection system on epitaxial wafer curvature. Theproject was supported by the National863Program (2011AA03A101).