文摘
Texture segmentation is one of the most difficult and important tasks in the remote sensing image processing. Through the years, several methods, based on statistical features, structural features and other image features, have been proposed for texture segmentation. This paper proposes a new feature based on Contourlet transform and Fractal analysis for texture segmentation. The performance of the proposed feature is compared with that of fractal, statistical, fuzzy and power spectrum features in presence of noise and the obtained results show superiority of the new proposed feature.