A Tool Assessing Optimal Multi-Scale Image Segmentation
- PDF / 13,895,737 Bytes
- 11 Pages / 595.276 x 790.866 pts Page_size
- 56 Downloads / 162 Views
RESEARCH ARTICLE
A Tool Assessing Optimal Multi-Scale Image Segmentation A. Mohan Vamsee1 • P. Kamala1 • Tapas R. Martha2 • K. Vinod Kumar2 G. Jai sankar1 • E. Amminedu1
•
Received: 11 April 2016 / Accepted: 30 April 2017 Ó Indian Society of Remote Sensing 2017
Abstract Image segmentation to create representative objects by region growing image segmentation techniques such as multi resolution segmentation (MRS) is mostly done through interactive selection of scale parameters and is still a subject of great research interest in object-based image analysis. In this study, we developed an optimum scale parameter selector (OSPS) tool for objective determination of multiple optimal scales in an image by MRS using eCognition software. The ready to use OSPS tool consists of three modules and determines optimum scales in an image by combining intrasegment variance and intersegment spatial autocorrelation. The tool was tested using WorldView-2 and Resourcesat-2 LISS-IV Mx images having different spectral and spatial resolutions in two areas to find optimal objects for ground features such as water bodies, trees, buildings, road, agricultural fields and landslides. Quality of the objects created for these features using scale parameters obtained from the OSPS tool was evaluated quantitatively using segmentation goodness metrics. Results show that OSPS tool is able determine optimum scale parameters for creation of representative objects from high resolution satellite images by MRS method. Keywords MRS GEOBIA ESP Segment optimisation Spatial autocorrelation
& Tapas R. Martha [email protected]; [email protected] 1
Department of Geo-Engineering, Andhra University College of Engineering (A), Visakhapatnam 530 003, India
2
Geosciences Group, National Remote Sensing Centre (NRSC), Indian Space Research Organisation (ISRO), Hyderabad 500 037, India
Introduction The primary purpose of classifying a satellite image is to extract features of interest and prepare thematic maps. With the increase in spatial resolution of satellites, features of interest have also been redefined. For example, individual trees and buildings are gradually becoming features of interest unlike forests and urban areas in the past. Since features of interest are gradually becoming small, tone and texture alone may not produce desirable image classification accuracy. Object-based methods, which use shape, size and context in addition to tone and texture, have been shown to be useful for achieving higher classification accuracy in comparison to pixel-based methods (Akcay and Aksoy 2008; Blaschke 2010; Gholoobi and Kumar 2015; Kartikeyan et al. 1994; Laliberte et al. 2004; Pradhan et al. 2014). One of the pre-requisites for object-based image analysis (OBIA) is the creation of objects, which is defined as a group of homogeneous pixels in an image (Benz et al. 2004; Blaschke et al. 2006). Image segmentation by region growing techniques are mainly used to group pixels based on their homogeneity and create objects that can be used in OBIA (P
Data Loading...
