Special Issue Article Open Access

Design of Multi-region SAR Segmentation by Parametric

Abstract

Synthetic Aperture Radar (SAR) is most important as a natural scenes and image segmentation purpose in Radar field. The image segmentation method name is Context-based Hierarchical Unequal Merging for SAR Image Segmentation (CHUMSIS) .We proposes an approach to represent super pixel context which uses the operation unit instead of pixels. Based on the Gestalt laws, three rules are satisfying under the condition and to manage different kinds of feature extraction from SAR image. The features are including brightness, texture, edges and spatial information of SAR images. While appearing the merging process, a hierarchical unequal merging algorithm is designed by the two stages:1) Coarse Merging Stage(CMS) and 2) Fine Merge Stage. Experiments on synthetic and real SAR images represent in this algorithm can make a balance between computation speed and segmentation accuracy. The proposed method is compared to the two state-of-the-art Markov random field models; CHUMSIS can obtain good segmentation results with short running time.

S.Narumalar, V. Madhan kumar

To read the full article Download Full Article