Please use this identifier to cite or link to this item:
Title: Analyzing the Effect of Enhanced Preprocessing Techniques on Classification of Sonar Images
Researcher: K.S. Jeen Marseline
Guide(s): Dr. C.Meena
Keywords: Preprocessing
University: Avinashilingam Deemed University For Women
Completed Date: 08.08.2015
Abstract: Sonar imaging technology is a field that is used in the study of seafloor through the analysis of high resolution images provided by sonar devices 2D imaging sonar also referred to as acoustic cameras which can be operated from both moving and stationary positions are used to capture the oceanic images The study of sonar images are performed by sonar imaging systems and consist of techniques for efficient analysis and understanding of the water surface of the Earth Frequently sonar images are analyzed to understand the sediments of seafloor region For this purpose an Automatic Seafloor Sediment Classification System ASSCS newlineusing sonar images is proposed In particular the research work analyzes enhancement techniques like noise removal and image fusion, to improve the process of classification The newlineASSCS used for this purpose performs classification in four steps. They are preprocessing segmentation feature extraction and classification The preprocessing step focuses on algorithms that can be used to enhance the sonar images using techniques like contrast adjustment illumination and lighting variation newlinecorrection and noise removal The contrast and illumination variations are corrected using the conventional histogram equalization method The sonar images are degraded by the presence of noise and to remove it two methods that enhance the operations of wavelets and curvelets are proposed The study also proposes block based image fusion methods to enhance the sonar newlineimage The main goal of the study is to analyze the effect of using these preprocessing techniques on classification A sonar image is composed of three main regions namely echo region shadow region and sediment region An enhanced Markov Random Field segmentation algorithm combined with an edge preserving Expectation Maximization algorithm is used for this purpose After newlinesegmentation the 11 types of wavelet packet based texture features are extracted The dimensionality of the extracted feature vector is reduced using Principle Component Analysis
Pagination: 223 p.
Appears in Departments:Department of Computer Science

Files in This Item:
File Description SizeFormat 
ksjeenmarseline_chapter1.pdfAttached File1.22 MBAdobe PDFView/Open
ksjeenmarseline_chapter2.pdf586.05 kBAdobe PDFView/Open
ksjeenmarseline_chapter3.pdf1.2 MBAdobe PDFView/Open
ksjeenmarseline_chapter4.pdf83.54 kBAdobe PDFView/Open
ksjeenmarseline_chapter5.pdf186.39 kBAdobe PDFView/Open
ksjeenmarseline_intro.pdf174.76 kBAdobe PDFView/Open

Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.