近年來遙測影像技術日益發達，傳統所採用的多光譜影像 (Multispectral Imaging) 系統已逐漸被高光譜影像 (Hyperspectral Imaging) 系統所取代。由於高光譜感測器所獲取的波段影像 (Band Image) 數量龐大，含有大量的資訊冗餘，如何從原始影像中選擇適當數量的波段來進行後續資料分析、壓縮、或資料傳輸為遙測領域長期的熱門議題之一。本篇論文針對資料傳輸與循序式資料分析的角度，提出兩類非監督式的序列式波段選擇 (Sequential Band Selection) 方法。類別一建立在高維空間上，遵循的原則為所欲選擇的波段將與已選擇的波段擁有最低的相似性，或者能涵蓋最大的高維空間涵蓋資訊；類別二則使建立在自稀疏性回歸 (Sparse Self-Representation) 的模型上，假設整幅高光譜影像可由本身的某些少數波段影像給代表，並依據貪婪演算法的正交匹配追蹤 (Orthogonal Matching Pursuit) 的策略，來尋找出最具代表性的波段。值得注意的是，本篇論文所提出的兩種方法所產生的結果均為一個波段序列 (Band Sequence)，意即其結果適合用於資料傳輸與同步分析。透過兩幅真實高光譜影像的實驗結果，證明我們提出的方法在所選擇的波段數目較少時，於影像分類 (Image Classification) 與光譜反混合 (Spectral Umixing) 的應用上效果比傳統的序列式波段選擇方法還要好。

Band selection (BS) is an important topic that has received wide attention for hyperspectral imaging (HSI) in remote sensing community for many years. In this thesis, two types of unsupervised sequential band selection (SQBS) methods for progressive data analysis of hyperspectral imagery are presented. In the first method, we propose two kinds of algorithms based on two individual perspectives: minimum band similarity and maximum band information. In the other method, we adopt sparse self-representation (SSR) model to assume that all the bands can be represented by a set of representative bands. A famous greedy algorithm, called orthogonal matching pursuit (OMP), is used to solve this optimization problem formulated by SSR model. Both two methods have the following properties: 1. The BS result forms a band sequence so that the progressive data analysis can be carried out. 2. The selected bands are highly un-correlated, that is, the information redundancy of the selected bands is minimized. The experiments implemented on two real hyperspectral datasets show that using the proposed SQBS methods can achieve significantly better performance for progressive classification and spectral unmixing over the conventional sequential band selection methods when the number of selected bands is low.

審定書 i
公開授權書 ii
誌謝 iii
摘要 iv
Abstract v
目錄 vii
圖目錄 ix
表目錄 x
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機 2
1.3 論文架構 3
第二章 相關研究與文獻回顧 5
2.1 高光譜影像介紹 5
2.2 波段選擇方法 7
2.2.1 波段選擇概念 7
2.2.2 循序式波段影像傳輸與處理 11
2.3 驗證方法 13
2.3.1 硬式影像分類器 13
2.3.2 混合像素模型與光譜反混合(像元解析) 17
2.4 文獻回顧總結與本論文實驗設定 22
第三章 相關性之序列式波段選擇法 25
3.1 最小相似性波段選擇法 (MSBS) 25
3.1.1 相關性係數 (CORR): 26
3.1.2 光譜角度分類 (SAM): 27
3.1.3 光譜資訊差異 (SID): 27
3.2 最大化資訊波段選擇法 (MIBS) 27
3.2.1 自動目標物產生過程 (ATGP) 28
3.2.2 單形體積生長演算法 (SGA) 28
第四章 稀疏性架構之序列式波段選擇法 30
4.1 稀疏性回歸介紹 30
4.2 正交匹配追蹤法介紹 31
4.3 正交匹配追蹤之序列式波段選擇法 (OMP-BS) 32
4.4 群組正交匹配追蹤序列式波段選擇法 (GOMP-BS) 37
第五章 實驗資料與實驗結果 42
5.1 義大利帕維亞大學高光譜影像 (Pavia) 42
5.1.1 實驗資料 42
5.1.2 前處理過程 44
5.1.3 波段選擇結果 44
5.1.4 循序式資料分析: 影像分類 46
5.1.5 循序式資料分析: 光譜反混合 52
5.1.6 計算複雜度分析 59
5.2 普渡大學松木林高光譜影像 (Purdue) 62
5.2.1 實驗資料 62
5.2.2 前處理過程 64
5.2.3 波段選擇結果 64
5.2.4 循序式資料分析: 影像分類 66
5.2.5 循序式資料分析: 光譜反混合 71
5.2.6 計算複雜度分析 78
5.3 實驗結果總結 81
第六章 結論與未來展望 84
6.1 結論 84
6.2 未來展望 85

[1] NASA,Landsat Science. Available: http://rst.gsfc.nasa.gov/
[2] P. W. Mausel, W. J. Kramber , and J. K. Lee, "Optimum band selection for supervised classification of multispectral data," Photogrammetric Engineering and Remote Sensing, vol. 56, pp. 55-60, 1990.
[3] N. Keshava, "Distance metrics and band selection in hyperspectral processing with applications to material identification and spectral libraries," IEEE Transactions on Geoscience and Remote Sensing, vol. 42, pp. 1552-1565, 2004.
[4] D. Qian and Y. He, "Unsupervised band selection for hyperspectral image analysis," in 2007 IEEE International Geoscience and Remote Sensing Symposium, 2007, pp. 282-285.
[5] D. Qian and Y. He, "Similarity-Based Unsupervised Band Selection for Hyperspectral Image Analysis," Geoscience and Remote Sensing Letters, IEEE, vol. 5, pp. 564-568, 2008.
[6] H. Su and Q. Du, "Hyperspectral band clustering and band selection for urban land cover classification," Geocarto International, vol. 27, pp. 395-411, 2012/08/01 2012.
[7] I. C. Chein and W. Su, "Constrained band selection for hyperspectral imagery," Geoscience and Remote Sensing, IEEE Transactions on, vol. 44, pp. 1575-1585, 2006.
[8] R. Hsuan, D. Qian, I. C. Chein, and J. O. Jensen, "Comparison between constrained energy minimization based approaches for hyperspectral imagery," in Advances in Techniques for Analysis of Remotely Sensed Data, 2003 IEEE Workshop on, 2003, pp. 244-248.
[9] I. C. Chein, D. Qian, S. Tzu-Lung, and M. L. G. Althouse, "A joint band prioritization and band-decorrelation approach to band selection for hyperspectral image classification," Geoscience and Remote Sensing, IEEE Transactions on, vol. 37, pp. 2631-2641, 1999.
[10] S. Li and H. Qi, "Sparse representation based band selection for hyperspectral images," in 2011 18th IEEE International Conference on Image Processing, 2011, pp. 2693-2696.
[11] W. Sun, L. Zhang, and B. Du, "A Sparse Self-Representation Method for Band Selection in Hyperspectral Imagery Classification," presented at the 7th Workshop on Hyperspectral Image and Signal Processing 2015.
[12] K. Fisher and C.-I. Chang, "Progressive band selection for satellite hyperspectral data compression and transmission," Journal of Applied Remote Sensing, vol. 4, pp. 041770-041770-17, 2010.
[13] C. I. Chang and K. H. Liu, "Progressive Band Selection of Spectral Unmixing for Hyperspectral Imagery," IEEE Transactions on Geoscience and Remote Sensing, vol. 52, pp. 2002-2017, 2014.
[14] R. Schultz, S.-Y. Chen, Y. Wang, C. Liu, and C.-I. Chang, "Progressive band processing of anomaly detection," 2013, pp. 88710P-88710P-11.
[15] R. C. Schultz, M. Hobbs, and C.-I. Chang, "Progressive band processing of simplex growing algorithm for finding endmembers in hyperspectral imagery," 2014, pp. 91240L-91240L-8.
[16] C. I. Chang, R. C. Schultz, M. C. Hobbs, S. Y. Chen, Y. Wang, and C. Liu, "Progressive Band Processing of Constrained Energy Minimization for Subpixel Detection," IEEE Transactions on Geoscience and Remote Sensing, vol. 53, pp. 1626-1637, 2015.
[17] R. A. Schowengerdt, Remote Sensing, Third Edition: Models and Methods for Image Processing: Academic Press, Inc., 2006.
[18] C. Cortes and V. Vapnik, "Support-vector networks," Machine learning, vol. 20, pp. 273-297, 1995.
[19] C. Cortes and V. Vapnik, Support-Vector Networks, 1995.
[20] N. Keshava and J. F. Mustard, "Spectral unmixing," IEEE Signal Processing Magazine, vol. 19, pp. 44-57, 2002.
[21] J. C. Harsanyi and C. I. Chang, "Hyperspectral image classification and dimensionality reduction: an orthogonal subspace projection approach," IEEE Transactions on Geoscience and Remote Sensing, vol. 32, pp. 779-785, 1994.
[22] H. Ren and I. C. Chein, "A generalized orthogonal subspace projection approach to unsupervised multispectral image classification," IEEE Transactions on Geoscience and Remote Sensing, vol. 38, pp. 2515-2528, 2000.
[23] C.-I. Chang, HYPERSPECTRAL IMAGING, 2003.
[24] I. C. Chein, "Orthogonal subspace projection (OSP) revisited: a comprehensive study and analysis," IEEE Transactions on Geoscience and Remote Sensing, vol. 43, pp. 502-518, 2005.
[25] C.-I. Chang, Hyperspectral imaging: techniques for spectral detection and classification vol. 1: Springer Science & Business Media, 2003.
[26] D. C. Heinz and I. C. Chein, "Fully constrained least squares linear spectral mixture analysis method for material quantification in hyperspectral imagery," IEEE Transactions on Geoscience and Remote Sensing, vol. 39, pp. 529-545, 2001.
[27] K. H. Liu, E. Wong, E. Y. Du, C. C. C. Chen, and C. I. Chang, "Kernel-Based Linear Spectral Mixture Analysis," IEEE Geoscience and Remote Sensing Letters, vol. 9, pp. 129-133, 2012.
[28] C.-I. Chang, Hyperspectral data processing: algorithm design and analysis: John Wiley & Sons, 2013.
[29] R. Hsuan and I. C. Chein, "Automatic spectral target recognition in hyperspectral imagery," IEEE Transactions on Aerospace and Electronic Systems, vol. 39, pp. 1232-1249, 2003.
[30] C. I. Chang, C. C. Wu, W. Liu, and Y. C. Ouyang, "A New Growing Method for Simplex-Based Endmember Extraction Algorithm," IEEE Transactions on Geoscience and Remote Sensing, vol. 44, pp. 2804-2819, 2006.
[31] M. D. Iordache, J. M. Bioucas-Dias, and A. Plaza, "Sparse Unmixing of Hyperspectral Data," IEEE Transactions on Geoscience and Remote Sensing, vol. 49, pp. 2014-2039, 2011.
[32] M. D. Iordache, J. M. Bioucas-Dias, and A. Plaza, "Collaborative Sparse Regression for Hyperspectral Unmixing," IEEE Transactions on Geoscience and Remote Sensing, vol. 52, pp. 341-354, 2014.
[33] M. Ehler and M. Hirn, "Sparse endmember extraction and demixing," in 2012 IEEE International Geoscience and Remote Sensing Symposium, 2012, pp. 1385-1388.
[34] C. h. Zhao, S. l. Cui, and B. Qi, "A sparse multiple endmember spectral mixture analysis algorithm of hyperspectral image," in 2014 12th International Conference on Signal Processing (ICSP), 2014, pp. 687-692.
[35] Q. Du, J. M. Bioucas-Dias, and A. Plaza, "Hyperspectral band selection using a collaborative sparse model," in 2012 IEEE International Geoscience and Remote Sensing Symposium, 2012, pp. 3054-3057.
[36] Z. Guo, H. Yang, X. Bai, Z. Zhang, and J. Zhou, "Semi-supervised hyperspectral band selection via sparse linear regression and hypergraph models," in 2013 IEEE International Geoscience and Remote Sensing Symposium - IGARSS, 2013, pp. 1474-1477.
[37] J. A. Tropp and A. C. Gilbert, "Signal Recovery From Random Measurements Via Orthogonal Matching Pursuit," IEEE Transactions on Information Theory, vol. 53, pp. 4655-4666, 2007.
[38] J. Mairal, F. Bach, J. Ponce, G. Sapiro, and A. Zisserman, "Discriminative learned dictionaries for local image analysis," in Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on, 2008, pp. 1-8.
[39] E. Elhamifar, G. Sapiro, and R. Vidal, "See all by looking at a few: Sparse modeling for finding representative objects," in Computer Vision and Pattern Recognition (CVPR), 2012 IEEE Conference on, 2012, pp. 1600-1607.
[40] G. Swirszcz, N. Abe, and A. C. Lozano, "Grouped Orthogonal Matching Pursuit for Variable Selection and Prediction," in Advances in Neural Information Processing Systems (NIPS)22, 2009, pp. 1150-1158.
[41] Y. W. Chao, Y. R. Yeh, Y. W. Chen, Y. J. Lee, and Y. C. F. Wang, "Locality-constrained group sparse representation for robust face recognition," in 2011 18th IEEE International Conference on Image Processing, 2011, pp. 761-764.
[42] H. L. Wu and S. Wang, "Compressive Wideband Spectrum Sensing Using Group Sparsity.," Journal of Signal Processing vol. 30, pp. 355-362, 2014.
[43] B. Mojaradia, H. Emamib, M. Varshosazc, and S. J. D, "A novel band selection method for hyperspectral data analysis," presented at the The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Beijing, 2008.
[44] H. Su, H. Yang, Q. Du, and Y. Sheng, "Semisupervised Band Clustering for Dimensionality Reduction of Hyperspectral Imagery," IEEE Geoscience and Remote Sensing Letters, vol. 8, pp. 1135-1139, 2011.
[45] P. Gurram and H. Kwon, "Contextual SVM for hyperspectral classification using Hilbert Space Embedding," in 2012 IEEE International Geoscience and Remote Sensing Symposium, 2012, pp. 5470-5473.
[46] K. H. Liu, Y. Y. Lin, and C. S. Chen, "Linear Spectral Mixture Analysis via Multiple-Kernel Learning for Hyperspectral Image Classification," IEEE Transactions on Geoscience and Remote Sensing, vol. 53, pp. 2254-2269, 2015.