A contextual classification system for remote sensing using a multivariate Gaussian MRF model

Author

Yamazaki, Tsutomu ; Gingras, Denis

Author_Institution

Commun. Res. Lab., Minist. of Posts & Telecommun., Kobe, Japan

Volume

2

fYear

1996

fDate

12-15 May 1996

Firstpage

648

Abstract

We propose a spatial contextual classification system for remote sensing images. In the system the observed multispectral images are modeled with a multivariate Gaussian Markov Random Field (GMRF) model and the hidden classified image is modeled with another type of MRF model. The classification is carried out from the viewpoint of Maximum a Posteriori (MAP) estimation. One of the well-known problems of MAP estimation is its high computational complexity. One way to avoid this problem is a pixelwise classification that is successfully implemented on a computer with a clique-type block matrix notation of a multivariate GMRF local conditional density function (LCDF). The proposed system is applied to real remote sensing data

Keywords

Markov processes; computational complexity; image classification; iterative methods; maximum likelihood estimation; remote sensing; Markov random field; clique-type block matrix notation; computational complexity; hidden classified image; local conditional density function; maximum a posteriori estimation; multispectral images; multivariate Gaussian MRF model; pixelwise classification; remote sensing images; spatial contextual classification system; Context modeling; Data analysis; Density functional theory; Laboratories; Markov random fields; Multispectral imaging; Optical filters; Optical sensors; Remote sensing; Smoothing methods;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 1996. ISCAS '96., Connecting the World., 1996 IEEE International Symposium on

Conference_Location

Atlanta, GA

Print_ISBN

0-7803-3073-0

Type

conf

DOI

10.1109/ISCAS.1996.541808

Filename

541808