Share Email Print
cover

Proceedings Paper

Multivariate change detection based on canonical transformation
Author(s): Jie Yang; Deren Li; Pan Zhu; Wen Yang
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

This article introduces the multivariate change detection which is based on the established canonical correlation analysis. It also proposes using post processing of the change detected by the multivariate change detection variables using maximum autocorrelation factor analysis. Differing from traditional schemes, the strategy takes two multivariate satellite images covering the same geographic area acquired at different points in time as a random whole sample and transforms two sets of random variables into one set of new random multivariate by using the so-called canonical transformation introduced in the paper. In doing so, the correlation between spectral bands in the same image and in the two different images is removed out as much as possible that the actual changes in all channels simultaneously can be accurately detected. The strategy is invariant to linear scaling. Therefore, it is insensitive to differences in gaining settings in a measuring device, or to linear radiometric and atmosphere correction schemes. The experimental results show the fact that the presented method is exactly creditable and effective on multivariate change detection of remote sensing satellite data.

Paper Details

Date Published: 25 September 2001
PDF: 6 pages
Proc. SPIE 4548, Multispectral and Hyperspectral Image Acquisition and Processing, (25 September 2001); doi: 10.1117/12.441380
Show Author Affiliations
Jie Yang, Wuhan Univ. (China)
Deren Li, Wuhan Univ. (China)
Pan Zhu, Wuhan Univ. (Thailand)
Wen Yang, Wuhan Univ. (China)


Published in SPIE Proceedings Vol. 4548:
Multispectral and Hyperspectral Image Acquisition and Processing

© SPIE. Terms of Use
Back to Top