Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery

Steven P. Brumby, James Theiler, Jeffrey J. Bloch, Neal R. Harvey, Simon Perkins, John J. Szymanski, A. Cody Young

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

15 Citations (Scopus)

Abstract

The Cerro Grande/Los Alamos forest fire devastated over 43,000 acres (17,500 ha) of forested land, and destroyed over 200 structures in the town of Los Alamos and the adjoining Los Alamos National Laboratory. The need to measure the continuing impact of the fire on the local environment has led to the application of a number of remote sensing technologies. During and after the fire, remote-sensing data was acquired from a variety of aircraft- and satellite-based sensors, including Landsat 7 Enhanced Thematic Mapper (ETM+). We now report on the application of a machine learning technique to the automated classification of land cover using multi-spectral and multi-temporal imagery. We apply a hybrid genetic programming/supervised classification technique to evolve automatic feature extraction algorithms. We use a software package we have developed at Los Alamos National Laboratory, called GENIE, to carry out this evolution. We use multispectral imagery from the Landsat 7 ETM+ instrument from before and after the wildfire. Using an existing land cover classification based on a 1992 Landsat 5 TM scene for our training data, we evolve algorithms that distinguish a range of land cover categories, and an algorithm to mask out clouds and cloud shadows. We report preliminary results of combining individual classification results using a K-means clustering approach. The details of our evolved classification are compared to the manually produced land-cover classification.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
EditorsM.R. Descour, S.S. Shen
Pages120-129
Number of pages10
Volume4480
DOIs
StatePublished - 2002
Externally publishedYes
EventImaging Spectrometry VII - San Diego, CA, United States
Duration: Aug 1 2001Aug 3 2001

Other

OtherImaging Spectrometry VII
CountryUnited States
CitySan Diego, CA
Period8/1/018/3/01

Fingerprint

imagery
Landsat 7
Fires
remote sensing
Remote sensing
Landsat 5
thematic mappers (LANDSAT)
forest fires
machine learning
Genetic programming
programming
Software packages
pattern recognition
aircraft
Learning systems
Feature extraction
Masks
education
masks
Aircraft

Keywords

  • Feature Extraction
  • Genetic programming
  • K-means clustering
  • Land cover
  • Multi-spectral imagery
  • Supervised classification
  • Wildfire

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Brumby, S. P., Theiler, J., Bloch, J. J., Harvey, N. R., Perkins, S., Szymanski, J. J., & Young, A. C. (2002). Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery. In M. R. Descour, & S. S. Shen (Eds.), Proceedings of SPIE - The International Society for Optical Engineering (Vol. 4480, pp. 120-129) https://doi.org/10.1117/12.453331

Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery. / Brumby, Steven P.; Theiler, James; Bloch, Jeffrey J.; Harvey, Neal R.; Perkins, Simon; Szymanski, John J.; Young, A. Cody.

Proceedings of SPIE - The International Society for Optical Engineering. ed. / M.R. Descour; S.S. Shen. Vol. 4480 2002. p. 120-129.

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

Brumby, SP, Theiler, J, Bloch, JJ, Harvey, NR, Perkins, S, Szymanski, JJ & Young, AC 2002, Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery. in MR Descour & SS Shen (eds), Proceedings of SPIE - The International Society for Optical Engineering. vol. 4480, pp. 120-129, Imaging Spectrometry VII, San Diego, CA, United States, 8/1/01. https://doi.org/10.1117/12.453331
Brumby SP, Theiler J, Bloch JJ, Harvey NR, Perkins S, Szymanski JJ et al. Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery. In Descour MR, Shen SS, editors, Proceedings of SPIE - The International Society for Optical Engineering. Vol. 4480. 2002. p. 120-129 https://doi.org/10.1117/12.453331
Brumby, Steven P. ; Theiler, James ; Bloch, Jeffrey J. ; Harvey, Neal R. ; Perkins, Simon ; Szymanski, John J. ; Young, A. Cody. / Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery. Proceedings of SPIE - The International Society for Optical Engineering. editor / M.R. Descour ; S.S. Shen. Vol. 4480 2002. pp. 120-129
@inproceedings{7c8f727e5d944be1bb1cf072ef084c9e,
title = "Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery",
abstract = "The Cerro Grande/Los Alamos forest fire devastated over 43,000 acres (17,500 ha) of forested land, and destroyed over 200 structures in the town of Los Alamos and the adjoining Los Alamos National Laboratory. The need to measure the continuing impact of the fire on the local environment has led to the application of a number of remote sensing technologies. During and after the fire, remote-sensing data was acquired from a variety of aircraft- and satellite-based sensors, including Landsat 7 Enhanced Thematic Mapper (ETM+). We now report on the application of a machine learning technique to the automated classification of land cover using multi-spectral and multi-temporal imagery. We apply a hybrid genetic programming/supervised classification technique to evolve automatic feature extraction algorithms. We use a software package we have developed at Los Alamos National Laboratory, called GENIE, to carry out this evolution. We use multispectral imagery from the Landsat 7 ETM+ instrument from before and after the wildfire. Using an existing land cover classification based on a 1992 Landsat 5 TM scene for our training data, we evolve algorithms that distinguish a range of land cover categories, and an algorithm to mask out clouds and cloud shadows. We report preliminary results of combining individual classification results using a K-means clustering approach. The details of our evolved classification are compared to the manually produced land-cover classification.",
keywords = "Feature Extraction, Genetic programming, K-means clustering, Land cover, Multi-spectral imagery, Supervised classification, Wildfire",
author = "Brumby, {Steven P.} and James Theiler and Bloch, {Jeffrey J.} and Harvey, {Neal R.} and Simon Perkins and Szymanski, {John J.} and Young, {A. Cody}",
year = "2002",
doi = "10.1117/12.453331",
language = "English (US)",
volume = "4480",
pages = "120--129",
editor = "M.R. Descour and S.S. Shen",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

}

TY - GEN

T1 - Evolving land cover classification algorithms for multi-spectral and multi-temporal imagery

AU - Brumby, Steven P.

AU - Theiler, James

AU - Bloch, Jeffrey J.

AU - Harvey, Neal R.

AU - Perkins, Simon

AU - Szymanski, John J.

AU - Young, A. Cody

PY - 2002

Y1 - 2002

N2 - The Cerro Grande/Los Alamos forest fire devastated over 43,000 acres (17,500 ha) of forested land, and destroyed over 200 structures in the town of Los Alamos and the adjoining Los Alamos National Laboratory. The need to measure the continuing impact of the fire on the local environment has led to the application of a number of remote sensing technologies. During and after the fire, remote-sensing data was acquired from a variety of aircraft- and satellite-based sensors, including Landsat 7 Enhanced Thematic Mapper (ETM+). We now report on the application of a machine learning technique to the automated classification of land cover using multi-spectral and multi-temporal imagery. We apply a hybrid genetic programming/supervised classification technique to evolve automatic feature extraction algorithms. We use a software package we have developed at Los Alamos National Laboratory, called GENIE, to carry out this evolution. We use multispectral imagery from the Landsat 7 ETM+ instrument from before and after the wildfire. Using an existing land cover classification based on a 1992 Landsat 5 TM scene for our training data, we evolve algorithms that distinguish a range of land cover categories, and an algorithm to mask out clouds and cloud shadows. We report preliminary results of combining individual classification results using a K-means clustering approach. The details of our evolved classification are compared to the manually produced land-cover classification.

AB - The Cerro Grande/Los Alamos forest fire devastated over 43,000 acres (17,500 ha) of forested land, and destroyed over 200 structures in the town of Los Alamos and the adjoining Los Alamos National Laboratory. The need to measure the continuing impact of the fire on the local environment has led to the application of a number of remote sensing technologies. During and after the fire, remote-sensing data was acquired from a variety of aircraft- and satellite-based sensors, including Landsat 7 Enhanced Thematic Mapper (ETM+). We now report on the application of a machine learning technique to the automated classification of land cover using multi-spectral and multi-temporal imagery. We apply a hybrid genetic programming/supervised classification technique to evolve automatic feature extraction algorithms. We use a software package we have developed at Los Alamos National Laboratory, called GENIE, to carry out this evolution. We use multispectral imagery from the Landsat 7 ETM+ instrument from before and after the wildfire. Using an existing land cover classification based on a 1992 Landsat 5 TM scene for our training data, we evolve algorithms that distinguish a range of land cover categories, and an algorithm to mask out clouds and cloud shadows. We report preliminary results of combining individual classification results using a K-means clustering approach. The details of our evolved classification are compared to the manually produced land-cover classification.

KW - Feature Extraction

KW - Genetic programming

KW - K-means clustering

KW - Land cover

KW - Multi-spectral imagery

KW - Supervised classification

KW - Wildfire

UR - http://www.scopus.com/inward/record.url?scp=0036394985&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036394985&partnerID=8YFLogxK

U2 - 10.1117/12.453331

DO - 10.1117/12.453331

M3 - Conference contribution

VL - 4480

SP - 120

EP - 129

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

A2 - Descour, M.R.

A2 - Shen, S.S.

ER -