A REVIEW ON VARIOUS METHODS OF IMAGE SEGMENTATION BASED ON REMOTE SENSING APPLICATIONS

Page 1

IJIRST –International Journal for Innovative Research in Science & Technology| Volume 1 | Issue 6 | November 2014 ISSN (online): 2349-6010

A Review on Various Methods of Image Segmentation Based on Remote Sensing Applications G. Priyadharsini PG Scholar Department of Information Technology SNS College of Technology, Coimbatore

C.Senthil Kumar Assistant Professor Department of Information Technology SNS College of Technology, Coimbatore

M.Udhayamoorthy Assistant Professor Department of Information Technology SNS College of Technology, Coimbatore

Abstract Information extraction in high-spatial resolution imagery has been the idea of many researchers all around the world. Several methods are used to inherit information from remote sensing data. The arrival of high-spatial resolution imagery necessitates new refined image processing algorithms for varied remote sensing applications such as segmentation of various regions in image. Image segmentation is defined as the process of isolating an image into non-overlapping and homogenous regions which is a necessary step toward higher level image processing namely automatic image interpretation, image analysis etc. Its performance decides the concluding result of a computer visual task. This paper surveys various methods which developed recently used for Remote sensing image segmentation. Each method is differentiated with other surveyed method and comparative measures of methods are presented which provides the merits and demerits of various image segmentation methods. Keywords: Image Segmentation, Merging, Multiscale Segmentation, Region-Based Image Segmentation, Remote Sensing Image, Statistical Region Merging. _______________________________________________________________________________________________________

I. INTRODUCTION Systems are sufficient for Remote Sensing Image (RSI) because of following With the evolution of remote sensing satellite image technology, the larger growth in the spatial resolution of remote sensing image is explored. Exploring such a efficient and faster information extraction methods and high-resolution remote sensing image processing has turn into an significant research theme in remote sensing applications. Segmentation of image and inheriting exact regions of interest is the primary stage to automatic extraction of ground objects in an image by computer vision system, and act as the foundation of articulating and measuring ground objects. Consequently, image segmentation has grown to be one of the chief researches that inheriting ground targets in the images of high resolution remote sensing application based system. Still, due to the enormous data and composite details of applications of high-resolution remote sensing image, the method of segmentation is altered from the usual natural images. Remote sensing image segmentation is a method to segregate an image into homogenous regions and to identify interested regions of objects, which is an important step toward advanced stage image processing. Since remote sensing images are multispectral, multi sensor and multi resolution, they enclose shape, spectrum, texture and various characteristics information. The redundancy and complexity are increased significantly so that common image segmentation techniques cannot attain acceptable results. Recently, the techniques based on the Theory of Machine Learning have got serious attentions in the areas such as local filtering, region growing, and so on. These techniques or methods no longer observes exact solution for image segmentation process, instead these process seeks for improved approximation of the exact results. These systems do not rely on particular part information to establish most favourable solutions, so they are more appropriate for remote sensing image segmentation.Yet, not all the segmentation particulars.  The redundancy and complexity of the Remote Sensing Image increases considerably.  The RSI offers more information namely context shape, spectral and texture.  The RSI is precisely bothered by noises, illuminance and so on. The following literature surveys various methods for Remote sensing image segmentation. And merits and demerits of each method are represented in the comparative table which is described in the following section.

All rights reserved by www.ijirst.org

184


Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.
A REVIEW ON VARIOUS METHODS OF IMAGE SEGMENTATION BASED ON REMOTE SENSING APPLICATIONS by IJIRST - Issuu