Application of mixed models of random fields for the segmentation of satellite images

Abstract

The problem of images segmentation is considered in the article. A brief overview of the existing segmentation methods is provided. We suggested to use estimates obtained in the course of nonlinear recurrent filtering for segmentation of inhomogeneous images. The proposed segmentation algorithm was investigated when working with generated images and real ones. It is shown that effective estimation of model parameters can provide the best quality of segmentation in comparison with the ISODATA algorithm. In addition, it is shown how it is possible to modify the segmentation model used to find the boundaries between objects.