PROJECT TITLE :

A Multi-Region Segmentation Method for SAR Images Based on the Multi-Texture Model With Level Sets - 2018

ABSTRACT:

Artificial aperture radar (SAR) image segmentation is a difficult downside due to the presence of sturdy multiplicative noise. To achieve multi-region segmentation for SAR images, this Project presents a parametric segmentation method based mostly on the multi-texture model with level sets. Segmentation is achieved by solving level set functions obtained from minimizing the proposed energy functional. To totally utilize image info, edge feature and region data are both included in the energy useful. For the need of level set evolution, the ratio of exponentially weighted averages operator is modified to get edge feature. Region data is obtained by the improved edgeworth series enlargement, that will adaptively model a SAR image distribution with respect to various sorts of regions. The performance of the proposed technique is verified by 3 high resolution SAR images. The experimental results demonstrate that SAR pictures will be segmented into multiple regions accurately while not any speckle pre-processing steps by the proposed method.