Segmentation Of SAR Images Using Quadtree And Potts Model
Olimpia Arellano Neri, M. Moctezuma, F. Parmiggiani · Biblioteca Digital da Memória Científica do INPE (National Institute for Space Research) · 1998
Abstract: This paper presents a contextual classifier based on quadtree structures and Markov random fields theory. The initial classification is realized by a clustering algorithm, then for each level of the tree, boundary regions are found. Pixels of boundary regions are classified by using a combination of nearest class mean criterion, Mahalanobis distance criterion and finally a Markov model. Our scheme is simple to implement and performs well, giving satisfactory results for SAR images.