1. 南京师范大学虚拟地理环境教育部重点实验室, 江苏 南京 210046 2. 密西西比州立大学电子与计算机工程系, Starkville, MS 39759, USA
Orthogonal Projection Divergence-Based Hyperspectral Band Selection
SU Hong-jun1,2，SHENG Ye-hua1*，Yang He2，Du Qian2
1. Key Lab of Virtual Geographic Environment (Ministry of Education), Nanjing Normal University, Nanjing 210046, China 2. Department of Electrical and Computer Engineering, Mississippi State University, Starkville, MS 39759, USA
Abstract：Due to the high data dimensionality of a hyperspectral image, dimensionality reduction algorithm has attracted much attention in hyperspectral image analysis. Band selection algorithm, which selects appropriate bands from the original set of spectral bands, can preserve original information from the data and is useful for image classification and recognition. In the present paper, a novel band selection algorithm based on orthogonal projection divergence (OPD) is proposed, it aims to discriminate the interesting objects from background and noise information, maximize the spectral similarity between different spectral vectors by projecting the original data to feature space. Two HYDICE Washington DC Mall images and an HYMAP Purdue campus image data were experimented, and support vector machine (SVM) classifier was used for classification. The selected band number varies from 5 to 40 in order to study the impacts of different band selection algorithms on different features. For the computation complex, the sequential floating forward search (SFFS) was used to get the appropriate bands. The experiments have proved that our proposed OPD algorithm can outperform other traditional band selection methods such as SAM, ED, SID, and LCMV-BCC for hyperspectral image analysis. It is proven that OPD band selection is effective and robust in hyperspectral remote sensing dimensionality reduction.
[1] Chang C I. Hyperspectral Imaging: Signal Processing Algorithm Design and Analysis, NJ: John Wiley & Sons, 2009. 1. [2] Hughes G F. IEEE Trans. Information Theory, 1968, IT-14: 55. [3] Chang C I, Du Q. IEEE Trans. Geoscience Remote Sensing, 2004, 42(3): 608. [4] Nascimento J, Dias J. IEEE Trans. Geoscience Remote Sensing, 2005, 43(4): 898. [5] Nadakuditi R R, Edelman A. IEEE Trans. Signal Processing, 2008, 56(7): 2625. [6] Bioucas-Dias J M, Nascimento J M P. IEEE Trans. Geoscience Remote Sensing, 2008, 46(8): 2435. [7] Debba P, Cho M, Mathieu R. Pro. Whispers, 2009, 5289068: 1. [8] Keshava N. IEEE Trans. Geoscience Remote Sensing, 2004, 42(7): 1552. [9] Chang C I. IEEE Trans. on Information Theory, 2000, 46(5): 1927. [10] Chang C I, Wang S. IEEE Trans. Geoscience Remote Sensing, 2006, 44(6): 1575. [11] Su H J, Sheng Y H, Du P J. Proceedings of the ISPRS 2008. Vol. XXXVII. Part B7: 279. [12] Bajcsy P, Groves P. Photogrammetric Engineering & Remote Sensing, 2004, 70: 793. [13] SU Hong-jun, DU Pei-jun, SHENG Ye-hua(苏红军, 杜培军, 盛业华). Application Research of Computers(计算机应用研究), 2008, 25(4): 1093. [14] Harsanyi J C, Chang C I. IEEE Trans. Geoscience Remote Sensing, 1994, 32: 779. [15] Chang C I. Hyperspectral Imaging: Techniques for Spectral Detection and Classification. New York: Kluwer, 2003. 15. [16] Pudil P, Novovicová J, Kittler J. Pattern Recognition Letters, 1994, 15: 1119. [17] Du Q, Yang H. IEEE Geoscience and Remote Sensing Letters, 2008, 5(4): 564. [18] Burges C J C. Data Mining and Knowledge Discovery, 1998, 2: 121.
