Gibbs random field models for image content characterization
M. Schroder, K. Seidel, Mihai P. Datcu · 2002
Satellite images contain an enormous amount of spatial information. To capture that information the authors propose, in the framework of a stochastic modelling of the image, the use of Gibbs Markov random fields. They expand on a particular model suitable for the use with typical remote sensing images. They demonstrate the capabilities of that model with two examples. In particular, they perform directed queries for specific spatial information.