本文探討使用數位相機影像攫取邊緣線之精確度。研究方法為透過影像處理之邊緣線檢測技術求取邊緣線，與實際精密量測值比較，以求得誤差。數位相機雖有快速且大量擷取影像之能力，但應用於精密量測時，會受到拍攝環境、視角誤差、物體表面材質、有限像素與有限灰階解析等之限制。為求得低誤差的邊緣線，本文對影像像素點上之灰階，求其一階導數法求得物體邊緣線圖之粗略草圖，再試著利用疊代式濾波器，與修正型單調性三次方曲線內插法，提升有限灰階之解析並得到次像素之精密度。本文也在不同亮度之光源、不同角度與不同材質下拍攝物體，試著找出拍攝環境與物體表面材質對於影像邊緣線之影響與關係，對於影像邊緣線不清晰，初步測試灰階轉換與直方圖等化法顯現模糊之邊緣線。由於物件之大小和座標不易精密定義，本文使用數種疊合法，以求得實際量測值與影像邊緣線誤差之最小值。測試後發現使用疊代式最小平方誤差法無法確實疊合數據，而使用列搜尋法則需花費較長的計算時間。本文採用綜合陣列搜尋法與最小平方誤差法，加快計算速度與達到更確實之疊合結果。本文並發展VB.net使用者介面，提供使用者方便選取欲計算邊緣線的工具。

A complete procedure of examining the precise edge detection of an image taking from a digital camera is developed in this study. Although the information of a digital image can be extracted via suitable post processing, the accuracy is restricted by the following factors: the environment quality, vision aberration induced by the location between object and camera, shape and reflectivity of object surface, and finite resolution of pixel and gray level. In order to define the precision of the edge detection, precise measurements of an object had been done before the digital images are taken. The error of the edge detection is defined as the difference between the extracted edge locations and the precisely measured edge data. The edge detection procedure involves: convert the image data to be gray level, calculate the first order derivative of the gray level. The local maximum point of the first order derivative is considered as the first estimation of the edge. The improvement of the edge location employs the iterative filter and monotonic cubic spline interpolation. To relate the environment to the edge detection, the effect of strength and incident angle of the light source and two objects with different surface material are also examined. For convenience, the extracted data is fixed and the measured edge data is properly transformed via the translation, rotation and enlargement processes. The best transformation is examined by the three procedures: least squares difference search, direct difference comparison via matrix search, and a search combining the least squares difference and direct comparison. It is found that the combined search attains the best and fast search. It seems that the proposed procedure can give an error of the edge detection about 0.25 pixel. This study also develops an user friendly Visual Basic.net interface.

參考文獻

1.R. C. Gonzalez, and R. E. Woods , “Digital Image Processing” Addison-Wesley, 1992.
2.S. E. Umbaugh, “Computer Vision and Image Processing : a Practical Approach Using CVIPtools”, Prentice Hall, 1998.
3.Y. N. Jeng, “The Moving Least Squares and Least p-Power Methods for Random Data,” The 7-th National Computational Fluid Dynamics Conference, pp.9-14, Aug. 2000.
4.G. Fasshauer, “Approximate Moving Least-Squares Approximations: A Fast and Accurate Multivariate Approximation Method,” Curve and Surfaces Fitting, vol.1, no.1, pp.1-10, 2002.
5.K. M. Liew, Y. Q. Huang, and J. N. Reddy, “A Hybrid Moving Least Squares and Differential Quadrature (MLSDQ) Meshfree Method,” Int. J. of Computational Engineering Science,” vol.3, No.1, pp.1-12. 2002.
6.Y. N. Jeng, P. G. Huang, and H. Chen, “Wave Decomposition in Physical Space Using Iterative Moving Least Squares Methods,” Proceedings of 11-th National Computational Fluid Dynamics Conference, Tai-Tung, Paper No. CFD11-0107, Aug. 2004.
7.Y. N. Jeng, Y. C. Cheng, and T. M. Yang, “Wave Decomposition Across Discontinuity Using Iterative Moving Least Squares Methods,” Proceedings of 11-th National Computational Fluid Dynamics Conference, Tai-Tung, Paper No. CFD11-0110, Aug. 2004.
8.Y. N. Jeng and P. G. Huang, “Evaluation of Derivatives from Composite Wave via Iterative Filter and A Numerical Method of Solving Partial Differential Equations,” Proceedings of 11-th National Computational Fluid Dynamics Conference, Tai-Tung, Paper No. CFD11-0111, Aug. 2004.
9.Y. N. Jeng and Y. C. Cheng, “A Simple Strategy to Evaluate the Frequency Spectrum of a Time Series Data with Non-Uniform Intervals,” 10-th National Computational Fluid Dynamics Conference of Taiwan, Hua-Lien, 2003, Paper No.A-9, also Transactions of the Aeronautical and Astronautic Society of the Republic of China in press, Sept. 2004.
10.楊宗明“應用離散移動式最小平方法誤差法於疊代式高/低通濾波器“ 國立成功大學航空太空工程學系碩士論文，July 2004。
11.A. C. Sanderson, and L. E. Weiss, “Image-base Visual Servo Control Using Relational Graph Error Signals,” Proceeding of IEEE International Conference on Cybernetics and Society, pp. 1074-1077, 1980.
12.F. N. Fritsch, and R. E. Carlson, “Monotone Piecewise Cubic Interpolation, “SIAM Numer. Analy. Vol.17, pp.238-246, 1980.
13.C. de Boor and B. Swartz, “Piecewise Monotone Interpolation,” J. Approx. Theory, vol.21,pp.411-416,1977.
14.H. T. Huynh, “Accurate Monotone Cubic Interpolation,” SIAM J. Number. Anal. vol.30, no.1, pp57-100, Feb.1993.
15.蔡維文,“影像數據誤差研究”國立成功大學航空太空工程學系碩士論文,June 2005