Identifying classes in SAR sea ice imagery using correlated texture

Leen-Kiat Soh, Costas Tsatsoulis

Research output: Contribution to conferencePaper

Abstract

This paper presents a new technique in identifying classes in Synthetic Aperture Radar (SAR) sea ice imagery using correlated texture. First, we employ dynamic local thresholding to generate a histogram of thresholds. Then, we use a multi-resolution peak-detection method, a strategy used in digital image quantization field, to extract significant intensity thresholds from the histogram and provide an initial segmentation. Next, we compute correlated texture of the result and create a matrix of spatial, probabilistic relationships among the classes. Given the texture, we cluster the classes into different groups. The clustering concept is based on an innovative `solidification' model that strives to obtain similar auto-correlated textural values for all groups. This process produces a second segmentation with the correct number of classes. We have tested our technique in more than 200 SAR sea ice imagery successfully. The entire process is fully automated and fast.

Original languageEnglish (US)
Pages1177-1179
Number of pages3
StatePublished - Jan 1 1997
EventProceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4) - Singapore, Singapore
Duration: Aug 3 1997Aug 8 1997

Other

OtherProceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4)
CitySingapore, Singapore
Period8/3/978/8/97

Fingerprint

Sea ice
Synthetic aperture radar
sea ice
synthetic aperture radar
imagery
Textures
texture
histogram
segmentation
solidification
digital image
detection method
Solidification
matrix

ASJC Scopus subject areas

  • Software
  • Geology

Cite this

Soh, L-K., & Tsatsoulis, C. (1997). Identifying classes in SAR sea ice imagery using correlated texture. 1177-1179. Paper presented at Proceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4), Singapore, Singapore, .

Identifying classes in SAR sea ice imagery using correlated texture. / Soh, Leen-Kiat; Tsatsoulis, Costas.

1997. 1177-1179 Paper presented at Proceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4), Singapore, Singapore, .

Research output: Contribution to conferencePaper

Soh, L-K & Tsatsoulis, C 1997, 'Identifying classes in SAR sea ice imagery using correlated texture' Paper presented at Proceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4), Singapore, Singapore, 8/3/97 - 8/8/97, pp. 1177-1179.
Soh L-K, Tsatsoulis C. Identifying classes in SAR sea ice imagery using correlated texture. 1997. Paper presented at Proceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4), Singapore, Singapore, .
Soh, Leen-Kiat ; Tsatsoulis, Costas. / Identifying classes in SAR sea ice imagery using correlated texture. Paper presented at Proceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4), Singapore, Singapore, .3 p.
@conference{26eee18c79144f40ba4145b1660d7a8e,
title = "Identifying classes in SAR sea ice imagery using correlated texture",
abstract = "This paper presents a new technique in identifying classes in Synthetic Aperture Radar (SAR) sea ice imagery using correlated texture. First, we employ dynamic local thresholding to generate a histogram of thresholds. Then, we use a multi-resolution peak-detection method, a strategy used in digital image quantization field, to extract significant intensity thresholds from the histogram and provide an initial segmentation. Next, we compute correlated texture of the result and create a matrix of spatial, probabilistic relationships among the classes. Given the texture, we cluster the classes into different groups. The clustering concept is based on an innovative `solidification' model that strives to obtain similar auto-correlated textural values for all groups. This process produces a second segmentation with the correct number of classes. We have tested our technique in more than 200 SAR sea ice imagery successfully. The entire process is fully automated and fast.",
author = "Leen-Kiat Soh and Costas Tsatsoulis",
year = "1997",
month = "1",
day = "1",
language = "English (US)",
pages = "1177--1179",
note = "Proceedings of the 1997 IEEE International Geoscience and Remote Sensing Symposium, IGARSS'97. Part 3 (of 4) ; Conference date: 03-08-1997 Through 08-08-1997",

}

TY - CONF

T1 - Identifying classes in SAR sea ice imagery using correlated texture

AU - Soh, Leen-Kiat

AU - Tsatsoulis, Costas

PY - 1997/1/1

Y1 - 1997/1/1

N2 - This paper presents a new technique in identifying classes in Synthetic Aperture Radar (SAR) sea ice imagery using correlated texture. First, we employ dynamic local thresholding to generate a histogram of thresholds. Then, we use a multi-resolution peak-detection method, a strategy used in digital image quantization field, to extract significant intensity thresholds from the histogram and provide an initial segmentation. Next, we compute correlated texture of the result and create a matrix of spatial, probabilistic relationships among the classes. Given the texture, we cluster the classes into different groups. The clustering concept is based on an innovative `solidification' model that strives to obtain similar auto-correlated textural values for all groups. This process produces a second segmentation with the correct number of classes. We have tested our technique in more than 200 SAR sea ice imagery successfully. The entire process is fully automated and fast.

AB - This paper presents a new technique in identifying classes in Synthetic Aperture Radar (SAR) sea ice imagery using correlated texture. First, we employ dynamic local thresholding to generate a histogram of thresholds. Then, we use a multi-resolution peak-detection method, a strategy used in digital image quantization field, to extract significant intensity thresholds from the histogram and provide an initial segmentation. Next, we compute correlated texture of the result and create a matrix of spatial, probabilistic relationships among the classes. Given the texture, we cluster the classes into different groups. The clustering concept is based on an innovative `solidification' model that strives to obtain similar auto-correlated textural values for all groups. This process produces a second segmentation with the correct number of classes. We have tested our technique in more than 200 SAR sea ice imagery successfully. The entire process is fully automated and fast.

UR - http://www.scopus.com/inward/record.url?scp=0030717895&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030717895&partnerID=8YFLogxK

M3 - Paper

SP - 1177

EP - 1179

ER -