This article is cited in 1 paper

Modeling and classification of multichannel remotely sensed images via copulas

V. A. Krylov

M. V. Lomonosov Moscow State University, Faculty of Computational Mathematics and Cybernetics

Abstract: A novel approach to modeling of multichannel remotely sensed images is proposed. This approach suggests to use the classical statistical probability distribution estimation methods for single channels and then the construction of the joint probability distribution of amultichannel image via copulas. An integration of the developed copula-based approach with aMarkov random field model is proposed for supervised Bayesian image classification. Experiments with real remotely sensed images captured by a synthetic aperture radar demonstrate high accuracy classification results proving the efficiency of the developed approach as compared to state-of-the-art methods.

Keywords: multichannel image; copula; Markov random field; Bayesian classification.