Title :
Mixture multiscale autoregressive modelling of SAR imagery for segmentation
Author :
Wen, Xian-Bin ; Tian, Zheng
Author_Institution :
Dept. of Comput. Sci. & Technol., Northwestern Polytech. Univ., Xi´´an, China
Abstract :
A new and efficient multiscale model for the segmentation of synthetic aperture radar imagery, namely the mixture multiscale autoregressive (MMAR) model, is presented. The estimation of parameters is easily performed via expectation maximisation algorithm. Experimental results of the MMAR model are proposed and compared with those of previous methods.
Keywords :
image segmentation; parameter estimation; radar imaging; synthetic aperture radar; SAR imagery; expectation maximisation algorithm; mixture multiscale autoregressive modelling; parameter estimation; segmentation; synthetic aperture radar;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20030805
