Novel feature extraction method for hyperspectral remote sensing image

Chunhong Liu; Huijie Zhao

doi:10.1117/12.750371

15 November 2007 Novel feature extraction method for hyperspectral remote sensing image

Chunhong Liu, Huijie Zhao

Proceedings Volume 6787, MIPPR 2007: Multispectral Image Processing; 67871S (2007) https://doi.org/10.1117/12.750371
Event: International Symposium on Multispectral Image Processing and Pattern Recognition, 2007, Wuhan, China

Abstract

In order to reduce high dimensions of hyperspectral remote sensing image and concentrate optimal information to reduced bands, this paper proposed a new method of feature extraction. The new method has two steps. The first step is to reduce the high dimensions by selecting high informative and low correlative bands according to the indexes calculated by a smart band selection method. The criterions that SBS method complied are: (1) The selected bands have the most information; (2) The selected bands have the smallest correlation with other bands. The second step is to decompose the selected bands by a novel second generation wavelet, predicting and updating subimages on rectangle and quincunx grids by Neville filters, finally using variance weighting as fusion weight. A 126-band HYMAP hyperspectral data was experimented in order to test the effect of the new method. The results showed classification accuracy is increased by using the novel feature extraction method.

Citation Download Citation

Chunhong Liu and Huijie Zhao "Novel feature extraction method for hyperspectral remote sensing image", Proc. SPIE 6787, MIPPR 2007: Multispectral Image Processing, 67871S (15 November 2007); https://doi.org/10.1117/12.750371

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available