基于支持向量机(SVM)特征加权/选择的光谱匹配算法

doi:10.3964/j.issn.1000-0593(2009)03-0735-05

Abstract
References
Related Citation (15)

Download: PDF (1101 KB)
Export: BibTeX | EndNote (RIS)

Abstract In the present article, feature selection/weighting based on SVM was employed to improve the algorithm of choosing reference spectrum through a multi-objective optimization approach proposed in reference [18]. Based on the sensitive analysis, half of features having low weights in SVM classification model were eliminated iteratively. Two criteria, matching accuracy and classification confidence, were used to select the best-performing feature subset. Three scenarios were designed: (1) only feature subset selected by SVM was used; (2) both feature subset and global weights were used, in which global weights were the coefficients of selected features in the SVM classification model; (3) both feature subset and local weights, which changed with the distance of a sample point to the SVM separation plan, were used. Experiment executed on the popular Indiana AVIRIS data set indicate that under all the three scenarios, spectral matching accuracies were increased by 13%-17% compared to the situation without feature selection. The result obtained under scenario 3 is the most accurate and the most stable, which can be primarily attributed to the ability of local weights to accurately describe local distribution of spectra from the same class in feature space. Moreover, scenario 3 can be regarded as the extension of scenario 2 because when spectra far away from the separation plane are selected as reference spectrums for matching, the features’ weights will not be considered. The results obtained under scenario 1 and 2 are very similar, indicating that considering global weights is not necessary. The research presented in this paper advanced the spectrum analysis using SVM to a higher level.

Key words：Hyperspectral;Spectrum Matching;Support Vector Machine;Feature selection

Received: 2007-11-16 Accepted: 2008-02-20

ZTFLH:

O657.3

Corresponding Authors: LI Jing E-mail: lijing@ires.cn

Cite this article:

WANG Yuan-yuan,CHEN Yun-hao,LI Jing. Improving Hyperspectral Matching Method through Feature-Selection/Weighting Based on SVM[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2009, 29(03): 735-739.

URL:

https://www.gpxygpfx.com/EN/10.3964/j.issn.1000-0593(2009)03-0735-05 OR https://www.gpxygpfx.com/EN/Y2009/V29/I03/735

[1] XU Wei-dong, YIN Qiu, KUANG Ding-bo (许卫东，尹球，匡定波). Journal of Infrared and Millimeter Waves(红外与毫米波学报), 2005, 24(4): 296.
[2] Rowan L C, Mars J C. Remote Sensing of Environment, 2003, 84: 350.
[3] Shrestha D P, Margate D E, van der Meer F, et al. International Journal of Applied Earth Observation and Geoinformation, 2005, 7: 85.
[4] XU Da-qi, NI Guo-qiang, SHEN Yuan-ting, et al(徐大琦，倪国强，沈渊婷，等). Spectroscopy and Spectral Analysis(光谱学与光谱分析)，2007, 27(3): 521.
[5] MA Yi, ZHANG Jie, CUI Ting-wei(马毅，张杰，崔廷伟). Spectroscopy and Spectral Analysis(光谱学与光谱分析)，2006, 26(12): 2302.
[6] SU Li-hong, LI Xiao-wen, WANG Jin-di, et al(苏理宏，李小文，王锦地，等). Advance in Earth Sciences(地球科学进展)，2003, 18(2)：185.
[7] TONG Qing-xi, ZHANG Bing, ZHENG Lan-fen(童庆禧，张兵，郑兰芬). Hyperspectral Remote Sensing: Principle, Technology and Application(高光谱遥感：原理、技术与应用). Beijing: Higher Education Press(北京：高等教育出版社)，2006. 206.
[8] Chen X F, Warner T A, Campagna D J. Remote Sensing of Environment, 2007, 110(3): 344.
[9] BAI Ji-wei, ZHAO Yong-chao, ZHANG Bing, et al(白继伟，赵永超，张兵，等). Computation Engineering and Applications(计算机工程与应用)，2003，39(13): 88.
[10] Clark M L, Roberts D A, Clark D B. Remote Sensing of Environment, 2005, 96: 375.
[11] Gallo I, Binaghi E, Boschetti M，et al. PASCAL, Pattern Analysis, Statistical Modeling and Computational Learning, 2004, 11, 3 http://eprints.pascal-network.org/archive/00000269/.
[12] van der Meer F. International Journal of Applied Earth Observation and Geoinformation, 2004, 5(1): 55.
[13] Du Y Z, Chang C I, Ren H, et al. Optical Engineering, 2004, 43(8): 1777.
[14] Chang C I. IEEE Transaction on Information Theory, 2000, 46(5): 1927.
[15] Filippi A M, Jensen J R. Remote Sensing of Environment, 2006, 100: 512.
[16] Glenn N F, Mundt J T, Weber K T. Remote Sensing of Environment, 2005, 95: 399.
[17] Dennison P E, Roberts D A. Remote Sensing of Environment, 2003, 87：123.
[18] Wang Y Y, Chen Y H, Li J. Progress in Intelligence Computation and Applications. Wuhan: Chinese University Gaosciences Press, 2007. 111.
[19] Wang S, Chang C I. IEEE Transaction on Geoscience and Remote Sensing, 2006, 44(6): 1575.
[20] Wang J, Chang C I. IEEE Transaction on Geoscience and Remote Sensing, 2006, 44(6): 1586.
[21] Foody G M, Mathur A. Remote Sensing of Environment, 2006，103: 179.
[22] Mahesh Pal, Mather P M. Future Generation Computer Systems, 2004，20：1215.
[23] AN Xin, SU Shi-guang, WANG Tao(安欣，苏时光，王韬). Spectroscopy and Spectral Analysis(光谱学与光谱分析)，2007, 27(8): 1619.
[24] Guyon I, Weston J, Barnhill S, et al. Machine Learning, 2002, 46(13): 389.
[25] Domeniconi C, Peng J, Gunopulos D, IEEE Transactions on Pattern Analysis and Machine Intelligence(TPAMI), 2002, 24(9): 1281.
[26] http://dynamo.ecn.purdue.edu/~biehl/MultiSpec.
[27] YANG Zhe-hai, LI Zhi-xin, HAN Jian-feng, et al(杨哲海，李之歆，韩建峰，等). Journal of Institute of Surveying and Mapping(测绘学院学报), 2004, 21(4): 253.