臺灣博碩士論文加值系統

本文提出運用圖形處理單元(Graphic Processing Unit, GPU)來實踐數位全像顯微術(Digital Holographic Microscopy, DHM)中，雜訊消除的窗口傅立葉轉換(Windowed Fourier transform, WFT)技術。
使用數位全像顯微術系統量測生物細胞之型態時，擷取之相位資訊通常會潛藏著許多的雜訊干擾，需在相位展開前將雜訊濾除，才能夠獲得正確的細胞型態。窗口傅立葉轉換之技術能夠有效的對雜訊進行濾除，且參數的判斷對於濾波之效果影響甚大，因此，本文對窗口傅立葉轉換的參數設定進行分析，能夠選取出較佳的參數以獲得不錯的濾波效果。最後使用最小平方相位展開法(Least Square phase unwrapping)的技術重建物體的三維資訊。
由於窗口傅立葉轉換的技術計算複雜度相當高，本文應用圖形處理單元(Graphic Processing Unit, GPU)平行處理的能力，來提升數值計算的時間，並對提升速度之效果進行評估。經由圖形處理單元的計算提升，可將運算速度提高至10 倍以上。

Window Fourier transform (WFT) is used to reduce the noise effect for the phase
image of the digital holographic system and apply it to measure the morphology of
the living cells. In advance, the Graphic processing unit (GPU) is developed to
increase the computation speed to achieve real-time processing.
In this thesis, the Digital Holographic Microscopy (DHM) system is apply to
measure the cell morphology, however, the phase image is highly perturbed by noise. The noise effect should be reduced before the phase unwrapping to obtain the correct morphology of cells. The Window Fourier transform is develop to filtering the noise and the parameter determination is crucial for the de-noise performance. Thus, filtering performance with related to the Window Fourier transform parameters is analyzed. The best selection of the Window Fourier transform filter parameters is then achieved.
Finally, the weakness of the Window Fourier transform filter is high
computational complexity. Thus, we apply the Graphic processing unit to improve the computation speech and the computation efficiency is evaluated. As a result, one can improve the computation speed up to 10 times for most of the cases.

第1章 緒論.....1
1.1 研究動機與目的.....1
1.2 文獻回顧.....1
1.3 研究貢獻.....3
1.4 本文架構.....3
第2章 光學系統簡介.....4
2.1 數位全像術.....4
2.2 數位全像顯微術.....5
2.2.1 數位全像顯微術記錄與重建.....6
2.2.2 Fresnel Transform.....8
2.2.3 角頻譜法(Angular spectrum approach).....11
2.2.4 數位重建.....13
第3章 研究方法.....14
3.1 窗口傅立葉轉換原理.....15
3.1.1 Windowed Fourier transform.....15
3.1.2 Windowed Fourier filtering.....16
3.2 圖形處理單元.....20
3.2.1 CUDA 程式架構.....21
3.2.2 CUDA 內建常數配置.....21
3.2.3 CUDA 記憶體模型.....23
3.3 相位展開.....26
3.3.1 相位展開技術之基本概念.....26
3.3.2 相位展開技術之基本分類.....27
3.3.3 最小平方相位展開法.....30
第4章 實驗與模擬.....33
4.1 Windowed Fourier filtering 參數分析.....33
4.1.1 參數分析(1).....36
4.1.2 參數分析(2).....40
4.2 使用圖形處理單元執行WFT.....45
4.3 GPU 用於DHM 之細胞影像處理.....48
第5章 結論與未來展望.....61
參考文獻.....62

[1] William W. Macy, Jr., "Two-dimensional fringe-pattern analysis,"Appl. Opt. 22,
3898-3901 (1983)
[2] “Introduction to Opto-Mechatronic Systems,” 國家實驗研究院, (2005).
[3] A. Capanni, L. Pezzati, D. Bertani, M. Cetica, and F. Francini, “Phase-shifting
speckle interferometry: a noise reduction filter for phase unwrapping,” Opt.
Eng., 36, pp.2466-2472 , (1997).
[4] J. Meneses, T. Gharbi, and P. Humbert, “Phase-unwrapping algorithm for
images with highnoise content based on a local histogram,” Appl. Opt., 44,
pp.1207-1215, (2005).
[5] F. Qian, X. Wang, X. Wang, and Y. Bu, “Adaptive filter for unwrapping noisy
phase image in phase-stepping interferometry,” Opt. & Laser Tech., 33,
pp.479-486, (2001).
[6] C.K. Hong, H.S. Ryu, and H.C. Lim, “Least-square fitting of the phase map
obtained in phase-shifting electronic speckle pattern interferometry,” Opt. Lett.,
20, pp.931-933, (1995).
[7] G. Nico, “Noise-residue filtering of interferometric phase images,” J. Opt. Soc.
Am. A , 17, pp.1962-1974, (2000).
[8] J. Bioucas-Dias, V. katkovnik, J. Astola, and k. Egiazarian, “Absolute phase
estimation: adaptive local denoising and global unwrapping,” Appl. Opt., 47,
pp.5358-5368, (2008).
[9] Q. Kemao, H. Wang, and W. Gao, “Windowed fourier transform for frimge
pattern analysis: theoretical analyses,” Appl. Opt., 47, pp.5408-5418, (2008).
[10] Tomoyoshi Shimobaba, Yoshikuni Sato, Junya Miura, Mai Takenouchi, and
Tomoyoshi Ito, "Real-time digital holographic microscopy using the graphic
processing unit," Opt. Express 16, 11776-11781 (2008)
[11] Wenjing Gao, Nguyen Thi Thanh Huyen, Ho Sy Loi, and Qian Kemao,
"Real-time 2D parallel windowed Fourier transform for fringe pattern analysis
using Graphics Processing Unit," Opt. Express 17, 23147-23152(2009)
[12] T. Zhang and I. Yamaguchi , “Three-dimensional microscopy with
phase-shifting digital holography ”Opt. Lett., 23, pp.1221–1223, (1998).
[13] P. Guo and A. J. Devaney, “Digital microscopy using phase-shifting digital
holography with two reference waves”, Opt. Lett., 29, pp.857-859, (2004).
[14] 廖華炫,“數位全像資訊於雜訊通道之傳輸,”逢甲大學通訊工程學系碩士論
文, (2006).
[15] M. K. Kim, L. Yu, and C. J. Mann “Interference techniques in digital
holography,” J. Opt. A: Pure Appl. Opt. 8, S518-S523,(2006).
[16] F. charriere, J. Kuhn, C. Depeursinge, “Influence of shot noise on phase
measurement accuracy in digital holographic microscopy, ” Appl. Opt., 15,
(2007).
[17] F. charriere, T. Colomb, C. Depeursinge, “Shot-noise influence on the
reconstructed phase image signal-to-noise ratio in digital holographic
microscopy,” Appl. Opt., 45, pp.7667-7673 (2006).
[18] Q. Kemao, “Two-dimensional windowed Fourier transform for fringe pattern
analysis: Principles, applications and implementations,” Opt. and Lasers Eng.,
45, pp.304-317, (2007).
[19] D. C. ghiglia, L. A. Romero, “Robust two-dimensional weighted and
unweighted phase unwrapping that uses fast transforms and iterative methods,”
Opt. Soc., 11, pp.107-117, (1994).
[20] CUDA Programming Guide 3.2,
http://developer.download.nvidia.com/compute/cuda/3_0/toolkit/docs/NVIDIA_
CUDA_ProgrammingGuide.pdf
[21] 蘇禹龍, “The study for Minimum LP-Norm on Unwrapping Phase,” 中興大學
機械工程學系碩士論文, (2006)
[22] R.Cusack,J. M. Huntley and H. T. Goldgrein, “Improved noise immune phase
unwrapping algorithm,” Appl. Opt., 34, pp.781-789 (1995).
[23] FFTW, “FFTW library 3.2.2”
http://www.fftw.org/
[24] NVIDIA, “CUDA CUFFT library 2.3” (2009),
http://www.nvidia.com/object/cuda_develop.html
[25] SUSS MicroOptic, “SMO TechInfo (Introduction To MicroLenses)”(2008)
http://www.suss-microoptics.com/downloads/SMO_TechInfo_Sheet_08.pdf