Share Email Print
cover

Proceedings Paper

Classification of hyperspectral images by enhancing absorption bands in spectral dimension
Author(s): Delian Liu; Liang Han; Jingguo Zong; Shaoze Zhang
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The spectral signatures in most hyperspectral classification approaches are generally treated as random vectors, which is inappropriate in denoting their typical physical characteristics, such as central wavelengths, widths, and depths of absorption bands. In this paper, we present a new classification approach by enhancing the absorption bands of spectral signatures to boost their physical information. Firstly, an analysis is made of the characteristics of absorption bands of spectral signatures. Next, an absorption bands enhancing approach is proposed based on the discussion of the approach of fusing spectral signatures and their derivative. Finally, the proposed approach is applied on two real hyperspectral subimages. The experimental results show that our proposed approach can significantly enhance the differences of spectral signatures of a hyperspectral images. And thus can improve the classification performance of hyperspectral images.

Paper Details

Date Published: 15 October 2012
PDF: 7 pages
Proc. SPIE 8419, 6th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optoelectronic Materials and Devices for Sensing, Imaging, and Solar Energy, 84192R (15 October 2012); doi: 10.1117/12.978168
Show Author Affiliations
Delian Liu, Xidian Univ. (China)
Liang Han, Xidian Univ. (China)
Jingguo Zong, Xidian Univ. (China)
Shaoze Zhang, Xidian Univ. (China)


Published in SPIE Proceedings Vol. 8419:
6th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optoelectronic Materials and Devices for Sensing, Imaging, and Solar Energy
Yadong Jiang; Junsheng Yu; Zhifeng Wang, Editor(s)

© SPIE. Terms of Use
Back to Top