近年來，機器視覺量測之應用越來越廣泛，例如：快速量測、高準度量測以及極端環境下的監控都可應用到機器視覺量測。在這些量測最重要的過程就是標記物的定位。
本研究之目的在建立一套伸長量影像量測系統，利用影像處理方法量測拉伸試驗中試件表面兩標記物形心距離之變化。本文利用兩種圓形標記物形心之量測方法：形心法與圓心法，以及三種影像扭曲之校正方法：反向扭曲校正、雙線性形狀函數校正及二次項形狀函數校正。實驗中分析了兩種量測方法對於不同顏色之標記物、不同影像內插次數、不同標記物尺寸與不同量測視野對於量測穩定性之影響，以及探討標記物位移量測之準確性，並比較三種影像校正方法之校正結果，最後建立伸長量影像量測系統，在拉伸試驗中量測試件上兩標記物形心距離之伸長量。
實驗結果顯示，使用形心法量測白點黑底標記物之量測穩定性最好，在相同標記物尺寸下，影像內插次數越多以及量測視野越小，可提高影像解析度，使量測越穩定，而在相同影像解析度下，標記物尺寸越大，在影像中像素點的數量越多，亦可使量測越穩定。三種影像扭曲校正法中，使用反向扭曲校正後之誤差最小，誤差均在0.2% 以下。在拉伸試驗中，對量測結果作曲線擬合修正可降低影像量測與兩點延伸計間之量測誤差至1 % 以內。動態拉伸量測中，以拉伸夾頭之位移量作為量測標準，將影像伸長量測之結果進行校正後，在夾頭位移1.38 μm前絕對誤差之平均為5 μm，最大誤差為16 μm，在夾頭位移1.38 μm後誤差均小於1 %。

The purpose of this paper was to develop a vision extension measuring system, using image processing to measure the centroid distance change of two targets on the specimen of tensile test. This paper applied two methods to measure the centroid of circular targets: the centroid method and the center method, and three image distortion calibration methods: inverse distorted calibration, bilinear shape function calibration and quadratic shape function calibration. The experiment analyzed the measuring stability about two measuring methods with different target colors, different times of image interpolation, different target size and different field of view, and discussed the accuracy about measuring the target displacement and compared three image distortion calibration methods, and finally applied the vision measurement system in the tensile test to measure the extension of the target centroid distance.

1. D. Marr and E. C. Hildreth, “Theory of edge detection,” Proc. R. Soc. London B 207, 187-217, 1980.
2. J. Canny, “A computational approach to edge detection,” IEEE Trans. Pattern Analysis Mach. lntell. 8(6), 679-698, 1986.
3. A.J. Tabatabai and O.R. Mitchell, “Edge location to subpixel values in digital imagery,” IEEE Trans Pattern Anal Mach Intell. 6(2), 188-201, 1984.
4. S. Ghosal, R. Ehrotra, “Orthogonal moment operators for subpixel edge detection,” Pattern Recognition, 26 (2), 295–306, 1993.
5. K. Ohtani and M. Baba, “A fast edge location measurement with subpixel accuracy using a CCD image,” Conference Record - IEEE Instrumentation and Measurement Technology Conference, v 3, p 2087-2092, 2001.
6. H. Zhou, Z.H. Liu and J.G. Yang, “An Improved Sub-Pixel Location Method for Image Measurement,” Communications in Computer and Information Science, v 214 CCIS, n PART 1, p 83-92, 2011.
7. J.G. Wu, K.F. He and B. Qin “Subpixel Edge Detection of Autofocus for Micro-machine Vision System,” Advanced Materials Research, v 216, p 228-232, 2011.
8. Y. Yin, M. Wang, X. Liu and X. Peng, “Center location of circular targets with surface fitting method,” Proceedings of SPIE - The International Society for Optical Engineering, v 8499, 2012.
9. 繆紹綱，“數位影像處理”，台灣培生教育出版股份有限公司，1999。
10. Thomas Klinger, “Image Processing with LabVIEW and IMAQ Vision”, Pretice Hall PTR,2003.
11. H. Farid and A.C. Popescu, “Blind removal of lens distortion,” Journal of the Optical Society of America A: Optics and Image Science, and Vision, v 18, n 9, p 2072-2078, September 2001.
12. 陳延昌，“應用幾何校正於條紋反射法作物體外貌量測”，國立成功大學機械工程學系碩士論文，2008。