臺灣博碩士論文加值系統

全極化合成孔徑雷達（Polarimetric Synthetic Aperture Radar, PolSAR）圖像的顏色編碼或分配對於極化資訊的視覺顯示和解釋至關重要。通常的做法是使用RGB色彩空間來顯示偏極編碼的PolSAR圖像，基於模型的目標分解的色彩訊息。然而，為了表示多偏極SAR資料的色度，基本RGB色彩空間的區域並不完全覆蓋人類的感知系統。
　　在本文中，基於CIE-Lab顏色空間是均勻的，我們使用了一個顏色分配框架，為了能有更好的視覺感知和資訊解釋。通過增加各種參數，從而使色彩增強視覺效果的有效解釋性。顯示出新的顏色分配方案不僅保留了極化特徵的色調，而且嵌入目標資訊的總功率。由於從PolSAR圖像資訊中實現色彩空間的轉換與計算將會耗費大量時間。因此本論文使用圖形處理器(Graphics Processing Units, GPU)的平行架構之CUDA（Compute Unified Device Architecture）技術進行加速，使整體更進一步提升效率。
　　由實驗結果證明，本論文提出的平行架構做運算圖資資訊空間的轉換，在程式執行時間上具有更快的速度，比起只有使用 CPU 做運算皆有顯著的效能提升，並且可以獲得正確的色彩空間顯現。

　　Color encoding or assignment of multi-polarization or fully polarimetric synthetic aperture radar (PolSAR) image is vital for visual display and interpretation of the polarimetric information. It is a common practice to use RGB or HIS color space to display the chromatic information for polarization-encoded model-based target　decomposition of PolSAR images. However, to express the multi partial Pol SAR’s chroma, the basic RGB area is not cover all the perception system of human.
Our research used the frame of color distribution which based on the evenly CIE-Lab color space, for better visual aware and information explanation. Let color have strongly visual explanation which through every kinds of parameter. The new color distribution not only keep the pole feature’s color but also enhance the total power of target’s information. But it spends much time to converse the color space from PolSAR image. Thus, we adopt the Compute Unified Device Architecture (CUDA) which is parallel structure of Graphics Processing Units (GPU) to speed up the transferring-time.
The experimental result shows that we do the data’s conversion with parallel structure not only can be more efficient on the running time than using CPU but get the correct color space.

摘 要 i
ABSTRACT iii
誌謝 v
目 錄 vi
表目錄 viii
圖目錄 ix
第一章 緒論 1
1.1 研究背景介紹 1
1.2 研究動機與目的 2
第二章 相關文獻回顧 4
2.1 合成孔徑雷達 4
2.2極化合成孔徑雷達 6
2.3 RGB色彩空間研究 6
2.4 CIE-Lab色彩空間研究 7
2.5 Yamaguchi’s Four Components Decomposition with Rotation 8
2.6 平行架構CUDA 11
2.6.1 GPGPU 11
2.6.2 CUDA 12
2.6.3 記憶體階層存取差異 14
第三章 研究方法 16
3.1 Yamaguchi’s New Four-component Decomposition Algorithm 16
3.2 Lab色度圖與分析 18
3.3 CIE-Lab之色彩空間應用於PolSAR資料分量問題之處理 18
3.3.1 資料前處理 18
3.3.2 Lab顏色分配方法 20
3.3.3針對感興趣散射做調整 22
3.4應用GPU針對運算式子加速 22
3.4.1 Kernel 1平行架構 22
3.4.2 Kernel 2平行架構 23
第四章 實驗結果 25
4.1 實驗圖資介紹 25
4.2 實驗環境 26
4.3 實驗流程 26
4.4 實驗結果與討論 27
4.4.1不同大小圖資之運算時間 27
4.4.2 以CIE-Lab所呈現的結果圖 30
第五章 結論與未來發展 31
5.1 結論 31
5.2 未來展望 31
參考文獻 32

[1]http://www.kc1718.com/pdlistone/tech/7145613.html
[2]張中白，「合成孔徑雷達」，科學發展390期，2005。
[3]陳錦瑰，「利用雷達干涉生成數值地形及其應用-監測九份地區之變形趨勢」，碩士論文，2008
[4]Lillesand et al., 2004
[5]Armin W. Doerry and Fred M. Dickey, “Synthetic Aperture Radar,” Optics & Photonics News, pp. 28-33, November 2004.
[6]Yang-Lang Chang, and Jyh-Perng Fang, Chapter of "A simulated annealing band selection approach for high-dimensional remote sensing images", Simulated Annealing, Theory with Applications, ISBN 978-953-307-134-3, Sciyo International, 2010.
[7]Lena Chang, Z. S. Tang, S. H. Chang, Yang-Lang Chang, "A region-based GLRT detection of oil spills in SAR images," Pattern Recognition Letters, Vol. 29, Issue 14, pp: 1915–1923, October 2008, Impact Factor: 1.034.
[8]Cheng-Yen Chiang; Kun-Shan Chen; Chih-Yuan Chu,” Lab Color Space Assignment for Decomposed Fully Polarization SAR Data,” EUSAR 2014; 10th European Conference on Synthetic Aperture Radar; Proceedings of, 3-5 June 2014, ISBN: 978-3-8007-3607-2
[9]Steve Wright. Digital Compositing for Film and Video. Focal Press. 2006. ISBN 024080760X.
[10]Richard S Hunter, abstract, Journal of the Optical Society of America, 38:1094 (1948).
[11]Richard S Hunter, abstract, Journal of the Optical Society of America, 38:661 (1948).
[12]Y. Yamaguchi, A. Sato, W.-M. Boerner, R. Sato, and H. Yamada, “Four- component scattering power decomposition with rotation of coherency matrix,” IEEE Trans. Geosci. Remote Sensing, vol. 49, no. 6, pp. 2251–2258, Jan. 2011.
[13]Wadell, J. M., “Scattering From Rough Surfaces,” Master’s Thesis, Naval
Postgraduate School, Monterey, California, September 1995.
[14]NVIDIA(2008). “CUDA Programming Guide v2.0”, NVIDIA.
[15]Yang-Lang Chang, A Novel Approach to Hyperspectral Image Classification,
Ph.D. Thesis, National Central University, Taiwan, 2003.
[16]Usai, S.,2001,“A New Approach For Long Term Monitoring Of Deformations
By Differential SAR Interferometry” 學位論文
[17]徐斌峰，一個維度優先權方法應用於粒子群優法在高維度影像特徵抽取，碩士論文，國立台北科技大學電機工程研究所，台北，2011。
[18]馮智勇、洪國展、蔡雅婷 多采科技有限公司,”引領高速平行運算風潮 打造專屬個人超級電腦 - 淺談 NVIDIA GPU 與 CUDA 技術,” 臺灣災害管理學會電子報,第十六期
[19]Wei-Sheng Liao, Tung-Ju Hsieh and Yang-Lang Chang , "GPU Parallel Computing of Spherical Panorama Video Stitching," The Second IEEE International Workshop on Parallel and Distributed Computing in Remote Sensing (IEEE PDCRS 2012), in conjunction with ICPADS 2012 : IEEE 18th International Conference on Parallel and Distributed Systems, December 17-19, 2012 , Singapore, ISSN: 1521-9097, (EI).
[20]Michael Garland, Scott Le Grand, John Nickolls, "Parallel Computing Experiences with CUDA", IEEE, 2008
[21]Enhua Wu, Youquan Liu, "Emerging Technology about GPGPU ", IEEE, 978-1-4244-2342-2, 2008