Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/91425
標題: Application of Zoom FFT in Refining the Orientation of Rotation Axis in Range Image Registration
Zoom FFT 於形貌配準高精度轉軸定向之應用
作者: 林俊良
Chun-Liang Lin
關鍵字: Range image registration
FFT
frequency domain
Zoom FFT
形貌資訊配準
傅立葉轉換
頻率域
Zoom FFT i
引用: [1] J.W.Cooley and J.W. Tukey 'An algorithm for the machine calculation of complex Fourier series' Math. Comput. 19:297-301,1965. [2] H. Elmer and S. Ronald F 'The Zoom FFT Using Complex Modulation' IEEE Proc.1977,ICA S SP,78-81. [3] L .Lucchese, G.M. Cortelazzo, and C. Monti, 'High Resolution Estimation of Planar Rotations Based on Fourier Transform and Radial Projections,' Proc. Int'l Symp. Circuits and Systems 1997,vol. II, pp.1181-1184, June 1997. [4] L. Lucchese, G. Doretto and G. M. Cortelazzo ,'A Frequency Domain Technique for Range Data Registration' IEEE Transaction on Pattern Analysis and Machine Intelligence, vol. 24, NO. 11,November 2002. [5] L .Lucchese, G.M. Cortelazzo, and C. Monti, 'High Resolution Estimation of Planar Rotations Based on Fourier Transform and Radial Projections,' Proc. Int'l Symp. Circuits and Systems 1997,vol. II, pp.1181-1184, June 1997. [6] K. I. Joy, 'Catmull-Rom-spline' On Line Geometric Modeling Notes : 1-4.(2002). [7] D. Pei, S. Yang, H. Yang, Q. Wang, M. Li, 'High Efficient and Real-Time Realization of Zoom FFT Based on FPGA' College of physics and Electronic Engineering Northwest Normal University Lanzhou, Gansu, China. [8] G.M. Cortelazzo, G.. Doretto, and L. Lucchese, 'Free-Form Textured Surfaces Registration by a Frequency Domain Technique,' Proc.IEEE Int'l Conf. Image Processing, vol. 1, pp. 813-817, Oct. 1998. [9] Y. Keller, A. Averbuch, Y. Shkolnisky: 'Algebraically Accurate Volume Registration Using Euler's Theorem and the 3-D Pseudo-Polar FFT, 'CVPR pp.795-800,2005: [10] P. J. Besl and N. D. McKay, 'A method for registration of 3-D shapes, 'IEEE Transactions on Pattern Analysis and Machine Intelligence ,pp. 239-256, 1992.
摘要: The three dimensional (3D) range image registration has excellent application, such as in industrial inspection and computer tomography to achieve low cost and efficiency advantages. This paper refers to the Zoom FFT concept in an algorithm of refining the orientation of rotation axis in range image registration in 3D.The Fourier transform allows the decoupling of the estimate of the rotation parameter from the estimate of the translation parameter, and focusing on the high precision rotation axis. We propose a two-step procedure the first step is to estimate common first-calculated (FC) axis, and second is to estimate high precision axis .The range image in spatial domain based on Gaussian Filter to distribute to 3D grid and changes to the frequency domain. Using the characteristic of two spectrums of common axis has same function value to estimate FC axis. Finally application of Zoom FFT is in refining the orientation of rotation axis. The resolution of rotation axis spectrum after zoom in improves from 0.2 to 0.025. Keyword : Range image registration ;FFT ;frequency domain ;Zoom FFT
本文引用 Zoom FFT 概念於頻率域三維形貌配準高精度轉軸定向。 頻率域配準可利用傅立葉轉換特性將旋轉與平移分開計算,在文中著 重於求取高精度轉軸。演算法流程分為兩大步驟 : 第一步驟,計算 兩形貌資訊之共同初估轉軸,第二步驟,放大初估轉軸附近頻譜且計 算高精度轉軸。 空間中形貌資訊依照高斯擴散至 3D 網格內並轉至頻率域,利用 兩視角型貌資訊的共同旋轉軸有相同的頻譜函數值之特性計算正規 化差異函式得到初估旋轉軸,最後應用 Zoom FFT 於放大初估旋轉軸 附近頻譜而求得高精度轉軸。從結果顯示 32 倍 Zoom FFT 後的可將 原始正規化差異函式之徑向投影頻譜解析度由 0.02 改善至 0.0225, 顯示頻譜放大超過 10 倍以上而得更好的旋轉軸精度。 關鍵字: 形貌資訊配準;傅立葉轉換;頻率域;Zoom FFT
URI: http://hdl.handle.net/11455/91425
文章公開時間: 2018-05-11
Appears in Collections:機械工程學系所

文件中的檔案:

取得全文請前往華藝線上圖書館



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.