Special Issue Article Open Access

SAR Image Segmentation Based On Hierarchical Merging Method

Abstract

Image segmentation is an important tool in satellite image processing and serves as an efficient front end to sophisticated algorithm and thereby simplify subsequent processing. It used to extract the meaningful objects lying in the image. The aim of the paper is to obtain the segmentation of the Synthetic Aperture Radar (SAR) image with minimum run time of the algorithm. The algorithm used for the segmentation is named as hierarchical unequal merging algorithm. In this paper instead of pixel, the superpixels are used as operation units. The preprocessing stage consist of formation of superpixel. The analysis of superpixel is performed by using three Gestalt law. In this edge detection, feature extraction are computed from the superpixel content. Based on this the merging of superpixel take place in two phase namely 1) Coarse merging stage 2) Fine merging stage. It will use less running time for the superpixels which are not present in the boundaries of different pattern and more running time in the superpixels which are in doubtful regions. The proposed algorithm is effectively reduces the process of segmentation and computational complexity.

To read the full article Download Full Article