Target classification from HR sonar images

Olga Lopera; Yves Dupont

doi:10.1109/OCEANS-Bergen.2013.6608183

Target classification from HR sonar images

Olga Lopera, Yves Dupont

Remote Sensing

Belgian Navy

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

This paper presents two integrated techniques for target classification from high-resolution (HR) sonar images. Both recognition procedures start with a despeckling algorithm based on the anisotropic diffusion filter. As a second step, a fuzzymorpho-based segmentation procedure is applied to the filtered images, which partitions the image into highlights and shadow areas. A number of geometrical features are extracted from these areas, and are then used to classify targets using two techniques: (i) a Markov Chain Monte Carlo (MCMC) approach and (ii) a Decision Tree Classifier (DTC). A comparison of both recognition techniques is drawn, and classification performance is estimated by ROC curves. Very promising results are obtained.

Original language	English
Title of host publication	OCEANS 2013 MTS/IEEE Bergen
Subtitle of host publication	The Challenges of the Northern Dimension
DOIs	https://doi.org/10.1109/OCEANS-Bergen.2013.6608183
Publication status	Published - 2013
Event	OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension - Bergen, Norway Duration: 10 Jun 2013 → 13 Jun 2013

Publication series

Name	OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension

Conference

Conference	OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension
Country/Territory	Norway
City	Bergen
Period	10/06/13 → 13/06/13

Keywords

Automated target recognition
image processing
synthetic aperture sonar

Access to Document

10.1109/OCEANS-Bergen.2013.6608183

Cite this

@inproceedings{76ae9261dc7f40eba2057340558f7d81,

title = "Target classification from HR sonar images",

abstract = "This paper presents two integrated techniques for target classification from high-resolution (HR) sonar images. Both recognition procedures start with a despeckling algorithm based on the anisotropic diffusion filter. As a second step, a fuzzymorpho-based segmentation procedure is applied to the filtered images, which partitions the image into highlights and shadow areas. A number of geometrical features are extracted from these areas, and are then used to classify targets using two techniques: (i) a Markov Chain Monte Carlo (MCMC) approach and (ii) a Decision Tree Classifier (DTC). A comparison of both recognition techniques is drawn, and classification performance is estimated by ROC curves. Very promising results are obtained.",

keywords = "Automated target recognition, image processing, synthetic aperture sonar",

author = "Olga Lopera and Yves Dupont",

year = "2013",

doi = "10.1109/OCEANS-Bergen.2013.6608183",

language = "English",

isbn = "9781479900015",

series = "OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension",

booktitle = "OCEANS 2013 MTS/IEEE Bergen",

note = "OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension ; Conference date: 10-06-2013 Through 13-06-2013",

}

Lopera, O & Dupont, Y 2013, Target classification from HR sonar images. in OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension., 6608183, OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension, OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension, Bergen, Norway, 10/06/13. https://doi.org/10.1109/OCEANS-Bergen.2013.6608183

TY - GEN

T1 - Target classification from HR sonar images

AU - Lopera, Olga

AU - Dupont, Yves

PY - 2013

Y1 - 2013

N2 - This paper presents two integrated techniques for target classification from high-resolution (HR) sonar images. Both recognition procedures start with a despeckling algorithm based on the anisotropic diffusion filter. As a second step, a fuzzymorpho-based segmentation procedure is applied to the filtered images, which partitions the image into highlights and shadow areas. A number of geometrical features are extracted from these areas, and are then used to classify targets using two techniques: (i) a Markov Chain Monte Carlo (MCMC) approach and (ii) a Decision Tree Classifier (DTC). A comparison of both recognition techniques is drawn, and classification performance is estimated by ROC curves. Very promising results are obtained.

AB - This paper presents two integrated techniques for target classification from high-resolution (HR) sonar images. Both recognition procedures start with a despeckling algorithm based on the anisotropic diffusion filter. As a second step, a fuzzymorpho-based segmentation procedure is applied to the filtered images, which partitions the image into highlights and shadow areas. A number of geometrical features are extracted from these areas, and are then used to classify targets using two techniques: (i) a Markov Chain Monte Carlo (MCMC) approach and (ii) a Decision Tree Classifier (DTC). A comparison of both recognition techniques is drawn, and classification performance is estimated by ROC curves. Very promising results are obtained.

KW - Automated target recognition

KW - image processing

KW - synthetic aperture sonar

UR - https://www.scopus.com/pages/publications/84886426630

U2 - 10.1109/OCEANS-Bergen.2013.6608183

DO - 10.1109/OCEANS-Bergen.2013.6608183

M3 - Conference contribution

AN - SCOPUS:84886426630

SN - 9781479900015

T3 - OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension

BT - OCEANS 2013 MTS/IEEE Bergen

T2 - OCEANS 2013 MTS/IEEE Bergen: The Challenges of the Northern Dimension

Y2 - 10 June 2013 through 13 June 2013

ER -

Target classification from HR sonar images

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this