Automated Detection and Patching of Erroneous Areas in Shadow Moiré Phase-Shifting Images and System Calibration under Heating Condition

• Main Authors: JieLin, 林 倢
• Other Authors: Yuan-Fang Chen
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/yrwy75

碩士 === 國立成功大學 === 機械工程學系碩博士班 === 101 === Grinding wheel surface has artificial diamonds and the printed circuit board contains many reflective spots and cavities on the surface. When using the shadow moiré system to measure the 3D surface profile of grinding wheel and printed circuit board, these diamonds, reflective spots and cavities lead to wrong gray level on phase-shifting patterns and result in error. When measuring specimen by shadow moiré system under heating condition, the system will produce unpredictable changes because of thermal expansion and cause measurement error. In this paper, using edge detection for averaged image of the four fringe patterns and setting thresholds to distinguish erroneous areas in phase-shifting images. The structure element size of proper-opening and proper-closing to automate detection and patching of erroneous areas in phase-shifting Images is decided by the maximum erroneous area of bounding rectangle width and height. The error of unwrapping phase could be eliminated by using 3x3 median filter in phase-shifting patterns or wrapping phase and make it successfully. To get error of measured value and standard value in different temperature, calibration plate is measured. In this paper, using one order least squares method and cubic spline method to fit error of measured value and standard value to establish program of shadow moiré system calibration under heating condition. By testing, the error using cubic spline is smaller than using one order least squares.