Band reduction for hyperspectral imagery processing

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Feature reduction denotes the group of techniques that reduce high dimensional data to a smaller set of components. In remote sensing feature reduction is a preprocessing step to many algorithms intended as a way to reduce the computational complexity and get a better data representation. Reduction can be done by either identifying bands from the original subset (selection), or by employing various transforms that produce new features (extraction). Research has noted challenges in both directions. In feature selection, identifying an "ideal" spectral band subset is a hard problem as the number of bands is increasingly large, rendering any exhaustive search unfeasible. To counter this, various approaches have been proposed that combine a search algorithm with a criterion function. However, the main drawback of feature selection remains the rather narrow bandwidths covered by the selected bands resulting in possible information loss. In feature extraction, some of the most popular techniques include Principal Component Analysis, Independent Component Analysis, Orthogonal Subspace Projection, etc. While they have been used with success in some instances, the resulting bands lack a physical relationship to the data and are mostly produced using statistical strategies. We propose a new technique for feature reduction that exploits search strategies for feature selection to extract a set of spectral bands from a given imagery. The search strategy uses dynamic programming techniques to identify 'the best set" of features.

Original languageEnglish
Title of host publicationProceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII
DOIs
StatePublished - 14 May 2010
EventComputational Imaging VIII - San Jose, CA, United States
Duration: 18 Jan 201019 Jan 2010

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume7533
ISSN (Print)0277-786X

Other

OtherComputational Imaging VIII
CountryUnited States
CitySan Jose, CA
Period18/01/1019/01/10

Fingerprint

Hyperspectral Imagery
imagery
Feature extraction
spectral bands
Processing
pattern recognition
Feature Selection
set theory
Search Strategy
dynamic programming
Feature Extraction
preprocessing
principal components analysis
Independent component analysis
remote sensing
Subset Selection
counters
Information Loss
Dynamic programming
projection

Keywords

  • Feature extraction
  • Hyperspectral imagery
  • Search methods

Cite this

Robila, S. A. (2010). Band reduction for hyperspectral imagery processing. In Proceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII [75330W] (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 7533). https://doi.org/10.1117/12.837953
Robila, Stefan A. / Band reduction for hyperspectral imagery processing. Proceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII. 2010. (Proceedings of SPIE - The International Society for Optical Engineering).
@inproceedings{5fa11b3e816245589098601bf9c0e422,
title = "Band reduction for hyperspectral imagery processing",
abstract = "Feature reduction denotes the group of techniques that reduce high dimensional data to a smaller set of components. In remote sensing feature reduction is a preprocessing step to many algorithms intended as a way to reduce the computational complexity and get a better data representation. Reduction can be done by either identifying bands from the original subset (selection), or by employing various transforms that produce new features (extraction). Research has noted challenges in both directions. In feature selection, identifying an {"}ideal{"} spectral band subset is a hard problem as the number of bands is increasingly large, rendering any exhaustive search unfeasible. To counter this, various approaches have been proposed that combine a search algorithm with a criterion function. However, the main drawback of feature selection remains the rather narrow bandwidths covered by the selected bands resulting in possible information loss. In feature extraction, some of the most popular techniques include Principal Component Analysis, Independent Component Analysis, Orthogonal Subspace Projection, etc. While they have been used with success in some instances, the resulting bands lack a physical relationship to the data and are mostly produced using statistical strategies. We propose a new technique for feature reduction that exploits search strategies for feature selection to extract a set of spectral bands from a given imagery. The search strategy uses dynamic programming techniques to identify 'the best set{"} of features.",
keywords = "Feature extraction, Hyperspectral imagery, Search methods",
author = "Robila, {Stefan A.}",
year = "2010",
month = "5",
day = "14",
doi = "10.1117/12.837953",
language = "English",
isbn = "9780819479266",
series = "Proceedings of SPIE - The International Society for Optical Engineering",
booktitle = "Proceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII",

}

Robila, SA 2010, Band reduction for hyperspectral imagery processing. in Proceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII., 75330W, Proceedings of SPIE - The International Society for Optical Engineering, vol. 7533, Computational Imaging VIII, San Jose, CA, United States, 18/01/10. https://doi.org/10.1117/12.837953

Band reduction for hyperspectral imagery processing. / Robila, Stefan A.

Proceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII. 2010. 75330W (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 7533).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Band reduction for hyperspectral imagery processing

AU - Robila, Stefan A.

PY - 2010/5/14

Y1 - 2010/5/14

N2 - Feature reduction denotes the group of techniques that reduce high dimensional data to a smaller set of components. In remote sensing feature reduction is a preprocessing step to many algorithms intended as a way to reduce the computational complexity and get a better data representation. Reduction can be done by either identifying bands from the original subset (selection), or by employing various transforms that produce new features (extraction). Research has noted challenges in both directions. In feature selection, identifying an "ideal" spectral band subset is a hard problem as the number of bands is increasingly large, rendering any exhaustive search unfeasible. To counter this, various approaches have been proposed that combine a search algorithm with a criterion function. However, the main drawback of feature selection remains the rather narrow bandwidths covered by the selected bands resulting in possible information loss. In feature extraction, some of the most popular techniques include Principal Component Analysis, Independent Component Analysis, Orthogonal Subspace Projection, etc. While they have been used with success in some instances, the resulting bands lack a physical relationship to the data and are mostly produced using statistical strategies. We propose a new technique for feature reduction that exploits search strategies for feature selection to extract a set of spectral bands from a given imagery. The search strategy uses dynamic programming techniques to identify 'the best set" of features.

AB - Feature reduction denotes the group of techniques that reduce high dimensional data to a smaller set of components. In remote sensing feature reduction is a preprocessing step to many algorithms intended as a way to reduce the computational complexity and get a better data representation. Reduction can be done by either identifying bands from the original subset (selection), or by employing various transforms that produce new features (extraction). Research has noted challenges in both directions. In feature selection, identifying an "ideal" spectral band subset is a hard problem as the number of bands is increasingly large, rendering any exhaustive search unfeasible. To counter this, various approaches have been proposed that combine a search algorithm with a criterion function. However, the main drawback of feature selection remains the rather narrow bandwidths covered by the selected bands resulting in possible information loss. In feature extraction, some of the most popular techniques include Principal Component Analysis, Independent Component Analysis, Orthogonal Subspace Projection, etc. While they have been used with success in some instances, the resulting bands lack a physical relationship to the data and are mostly produced using statistical strategies. We propose a new technique for feature reduction that exploits search strategies for feature selection to extract a set of spectral bands from a given imagery. The search strategy uses dynamic programming techniques to identify 'the best set" of features.

KW - Feature extraction

KW - Hyperspectral imagery

KW - Search methods

UR - http://www.scopus.com/inward/record.url?scp=77952034863&partnerID=8YFLogxK

U2 - 10.1117/12.837953

DO - 10.1117/12.837953

M3 - Conference contribution

AN - SCOPUS:77952034863

SN - 9780819479266

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Proceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII

ER -

Robila SA. Band reduction for hyperspectral imagery processing. In Proceedings of SPIE-IS and T Electronic Imaging - Computational Imaging VIII. 2010. 75330W. (Proceedings of SPIE - The International Society for Optical Engineering). https://doi.org/10.1117/12.837953