Burned area estimations derived from Landsat ETM+ and OLI data: Comparing Genetic Programming with Maximum Likelihood and Classification and Regression Trees

Ana I.R. Cabral, Sara Silva, Pedro C. Silva, Leonardo Vanneschi, Maria J. Vasconcelos

Research output: Contribution to journalReview article

1 Citations (Scopus)

Abstract

Every year, large areas of savannas and woodlands burn due to natural conditions and land management practices. Given the relevant level of greenhouse gas emissions produced by biomass burning in tropical regions, it is becoming even more important to clearly define historic fire regimes so that prospective emission reduction management strategies can be well informed, and their results Measured, Reported, and Verified (MRV). Thus, developing tools for accurately, and periodically mapping burned areas, based on cost advantageous, expedite, and repeatable rigorous approaches, is important. The main objective of this study is to investigate the potential of novel Genetic Programming (GP) methodologies for classifying burned areas in satellite imagery over savannas and tropical woodlands and to assess if they can improve over the popular and currently applied methods of Maximum Likelihood classification and Classification and Regression Tree analysis. The tests are performed using three Landsat images from Brazil (South America), Guinea-Bissau (West Africa) and the Democratic Republic of Congo (Central Africa). Burned areas were digitized on-screen to produce mapped information serving as surrogate ground-truth. Validation results show that all methods provide an overestimation of burned area, but GP achieves higher accuracy values in two of the three cases. GP is the most versatile machine learning method available today, but still largely underused in remote sensing. This study shows that standard GP can produce better results than two classical methods, and illustrates its versatility and potential in becoming a mainstream method for more difficult tasks involving the large amounts of newly available data.

LanguageEnglish
Pages94-105
Number of pages12
JournalISPRS Journal of Photogrammetry and Remote Sensing
Volume142
DOIs
Publication statusPublished - 1 Aug 2018

Fingerprint

Genetic programming
programming
Maximum likelihood
Landsat
regression analysis
savanna
woodland
Democratic Republic of Congo
land management
Guinea
Satellite imagery
satellite imagery
biomass burning
Gas emissions
Greenhouse gases
machine learning
ground truth
greenhouses
Learning systems
Remote sensing

Keywords

  • Burned area mapping
  • Classification and Regression Trees
  • Genetic Programming
  • Landsat ETM+/OLI
  • Maximum Likelihood
  • Savana woodlands

Cite this

@article{6cd946f398914f52aca6f1a50bc875ee,
title = "Burned area estimations derived from Landsat ETM+ and OLI data: Comparing Genetic Programming with Maximum Likelihood and Classification and Regression Trees",
abstract = "Every year, large areas of savannas and woodlands burn due to natural conditions and land management practices. Given the relevant level of greenhouse gas emissions produced by biomass burning in tropical regions, it is becoming even more important to clearly define historic fire regimes so that prospective emission reduction management strategies can be well informed, and their results Measured, Reported, and Verified (MRV). Thus, developing tools for accurately, and periodically mapping burned areas, based on cost advantageous, expedite, and repeatable rigorous approaches, is important. The main objective of this study is to investigate the potential of novel Genetic Programming (GP) methodologies for classifying burned areas in satellite imagery over savannas and tropical woodlands and to assess if they can improve over the popular and currently applied methods of Maximum Likelihood classification and Classification and Regression Tree analysis. The tests are performed using three Landsat images from Brazil (South America), Guinea-Bissau (West Africa) and the Democratic Republic of Congo (Central Africa). Burned areas were digitized on-screen to produce mapped information serving as surrogate ground-truth. Validation results show that all methods provide an overestimation of burned area, but GP achieves higher accuracy values in two of the three cases. GP is the most versatile machine learning method available today, but still largely underused in remote sensing. This study shows that standard GP can produce better results than two classical methods, and illustrates its versatility and potential in becoming a mainstream method for more difficult tasks involving the large amounts of newly available data.",
keywords = "Burned area mapping, Classification and Regression Trees, Genetic Programming, Landsat ETM+/OLI, Maximum Likelihood, Savana woodlands",
author = "Cabral, {Ana I.R.} and Sara Silva and Silva, {Pedro C.} and Leonardo Vanneschi and Vasconcelos, {Maria J.}",
note = "Cabral, A. I. R., Silva, S., Silva, P. C., Vanneschi, L., & Vasconcelos, M. J. (2018). Burned area estimations derived from Landsat ETM+ and OLI data: Comparing Genetic Programming with Maximum Likelihood and Classification and Regression Trees. ISPRS Journal of Photogrammetry and Remote Sensing, 142, 94-105. DOI: 10.1016/j.isprsjprs.2018.05.007",
year = "2018",
month = "8",
day = "1",
doi = "10.1016/j.isprsjprs.2018.05.007",
language = "English",
volume = "142",
pages = "94--105",
journal = "ISPRS Journal of Photogrammetry and Remote Sensing",
issn = "0924-2716",
publisher = "Elsevier Science B.V., Amsterdam.",

}

Burned area estimations derived from Landsat ETM+ and OLI data : Comparing Genetic Programming with Maximum Likelihood and Classification and Regression Trees. / Cabral, Ana I.R.; Silva, Sara; Silva, Pedro C.; Vanneschi, Leonardo; Vasconcelos, Maria J.

In: ISPRS Journal of Photogrammetry and Remote Sensing, Vol. 142, 01.08.2018, p. 94-105.

Research output: Contribution to journalReview article

TY - JOUR

T1 - Burned area estimations derived from Landsat ETM+ and OLI data

T2 - ISPRS Journal of Photogrammetry and Remote Sensing

AU - Cabral, Ana I.R.

AU - Silva, Sara

AU - Silva, Pedro C.

AU - Vanneschi, Leonardo

AU - Vasconcelos, Maria J.

N1 - Cabral, A. I. R., Silva, S., Silva, P. C., Vanneschi, L., & Vasconcelos, M. J. (2018). Burned area estimations derived from Landsat ETM+ and OLI data: Comparing Genetic Programming with Maximum Likelihood and Classification and Regression Trees. ISPRS Journal of Photogrammetry and Remote Sensing, 142, 94-105. DOI: 10.1016/j.isprsjprs.2018.05.007

PY - 2018/8/1

Y1 - 2018/8/1

N2 - Every year, large areas of savannas and woodlands burn due to natural conditions and land management practices. Given the relevant level of greenhouse gas emissions produced by biomass burning in tropical regions, it is becoming even more important to clearly define historic fire regimes so that prospective emission reduction management strategies can be well informed, and their results Measured, Reported, and Verified (MRV). Thus, developing tools for accurately, and periodically mapping burned areas, based on cost advantageous, expedite, and repeatable rigorous approaches, is important. The main objective of this study is to investigate the potential of novel Genetic Programming (GP) methodologies for classifying burned areas in satellite imagery over savannas and tropical woodlands and to assess if they can improve over the popular and currently applied methods of Maximum Likelihood classification and Classification and Regression Tree analysis. The tests are performed using three Landsat images from Brazil (South America), Guinea-Bissau (West Africa) and the Democratic Republic of Congo (Central Africa). Burned areas were digitized on-screen to produce mapped information serving as surrogate ground-truth. Validation results show that all methods provide an overestimation of burned area, but GP achieves higher accuracy values in two of the three cases. GP is the most versatile machine learning method available today, but still largely underused in remote sensing. This study shows that standard GP can produce better results than two classical methods, and illustrates its versatility and potential in becoming a mainstream method for more difficult tasks involving the large amounts of newly available data.

AB - Every year, large areas of savannas and woodlands burn due to natural conditions and land management practices. Given the relevant level of greenhouse gas emissions produced by biomass burning in tropical regions, it is becoming even more important to clearly define historic fire regimes so that prospective emission reduction management strategies can be well informed, and their results Measured, Reported, and Verified (MRV). Thus, developing tools for accurately, and periodically mapping burned areas, based on cost advantageous, expedite, and repeatable rigorous approaches, is important. The main objective of this study is to investigate the potential of novel Genetic Programming (GP) methodologies for classifying burned areas in satellite imagery over savannas and tropical woodlands and to assess if they can improve over the popular and currently applied methods of Maximum Likelihood classification and Classification and Regression Tree analysis. The tests are performed using three Landsat images from Brazil (South America), Guinea-Bissau (West Africa) and the Democratic Republic of Congo (Central Africa). Burned areas were digitized on-screen to produce mapped information serving as surrogate ground-truth. Validation results show that all methods provide an overestimation of burned area, but GP achieves higher accuracy values in two of the three cases. GP is the most versatile machine learning method available today, but still largely underused in remote sensing. This study shows that standard GP can produce better results than two classical methods, and illustrates its versatility and potential in becoming a mainstream method for more difficult tasks involving the large amounts of newly available data.

KW - Burned area mapping

KW - Classification and Regression Trees

KW - Genetic Programming

KW - Landsat ETM+/OLI

KW - Maximum Likelihood

KW - Savana woodlands

UR - http://www.scopus.com/inward/record.url?scp=85047983235&partnerID=8YFLogxK

U2 - 10.1016/j.isprsjprs.2018.05.007

DO - 10.1016/j.isprsjprs.2018.05.007

M3 - Review article

VL - 142

SP - 94

EP - 105

JO - ISPRS Journal of Photogrammetry and Remote Sensing

JF - ISPRS Journal of Photogrammetry and Remote Sensing

SN - 0924-2716

ER -