Mapping land-cover modifications over large areas: A comparison of machine learning algorithms

John Rogan, Janet Franklin, Doug Stow, Jennifer Miller, Curtis Woodcock, Dar Roberts

Research output: Contribution to journalArticle

127 Citations (Scopus)

Abstract

Large area land-cover monitoring scenarios, involving large volumes of data, are becoming more prevalent in remote sensing applications. Thus, there is a pressing need for increased automation in the change mapping process. The objective of this research is to compare the performance of three machine learning algorithms (MLAs); two classification tree software routines (S-plus and C4.5) and an artificial neural network (ARTMAP), in the context of mapping land-cover modifications in northern and southern California study sites between 1990/91 and 1996. Comparisons were based on several criteria: overall accuracy, sensitivity to data set size and variation, and noise. ARTMAP produced the most accurate maps overall (∼ 84%), for two study areas - in southern and northern California, and was most resistant to training data deficiencies. The change map generated using ARTMAP has similar accuracies to a human-interpreted map produced by the U.S. Forest Service in the southern study area. ARTMAP appears to be robust and accurate for automated, large area change monitoring as it performed equally well across the diverse study areas with minimal human intervention in the classification process.

Original languageEnglish (US)
Pages (from-to)2272-2283
Number of pages12
JournalRemote Sensing of Environment
Volume112
Issue number5
DOIs
StatePublished - May 15 2008
Externally publishedYes

Fingerprint

artificial intelligence
land cover
Learning algorithms
Learning systems
USDA Forest Service
monitoring
automation
neural networks
remote sensing
Monitoring
artificial neural network
Remote sensing
Automation
Neural networks
software
comparison
machine learning

Keywords

  • Land-cover change
  • Large area monitoring
  • Machine learning

ASJC Scopus subject areas

  • Computers in Earth Sciences
  • Earth-Surface Processes
  • Environmental Science(all)
  • Management, Monitoring, Policy and Law

Cite this

Mapping land-cover modifications over large areas : A comparison of machine learning algorithms. / Rogan, John; Franklin, Janet; Stow, Doug; Miller, Jennifer; Woodcock, Curtis; Roberts, Dar.

In: Remote Sensing of Environment, Vol. 112, No. 5, 15.05.2008, p. 2272-2283.

Research output: Contribution to journalArticle

Rogan, J, Franklin, J, Stow, D, Miller, J, Woodcock, C & Roberts, D 2008, 'Mapping land-cover modifications over large areas: A comparison of machine learning algorithms', Remote Sensing of Environment, vol. 112, no. 5, pp. 2272-2283. https://doi.org/10.1016/j.rse.2007.10.004
Rogan, John ; Franklin, Janet ; Stow, Doug ; Miller, Jennifer ; Woodcock, Curtis ; Roberts, Dar. / Mapping land-cover modifications over large areas : A comparison of machine learning algorithms. In: Remote Sensing of Environment. 2008 ; Vol. 112, No. 5. pp. 2272-2283.
@article{933c7b80e644457893460c63a1f6c1e6,
title = "Mapping land-cover modifications over large areas: A comparison of machine learning algorithms",
abstract = "Large area land-cover monitoring scenarios, involving large volumes of data, are becoming more prevalent in remote sensing applications. Thus, there is a pressing need for increased automation in the change mapping process. The objective of this research is to compare the performance of three machine learning algorithms (MLAs); two classification tree software routines (S-plus and C4.5) and an artificial neural network (ARTMAP), in the context of mapping land-cover modifications in northern and southern California study sites between 1990/91 and 1996. Comparisons were based on several criteria: overall accuracy, sensitivity to data set size and variation, and noise. ARTMAP produced the most accurate maps overall (∼ 84{\%}), for two study areas - in southern and northern California, and was most resistant to training data deficiencies. The change map generated using ARTMAP has similar accuracies to a human-interpreted map produced by the U.S. Forest Service in the southern study area. ARTMAP appears to be robust and accurate for automated, large area change monitoring as it performed equally well across the diverse study areas with minimal human intervention in the classification process.",
keywords = "Land-cover change, Large area monitoring, Machine learning",
author = "John Rogan and Janet Franklin and Doug Stow and Jennifer Miller and Curtis Woodcock and Dar Roberts",
year = "2008",
month = "5",
day = "15",
doi = "10.1016/j.rse.2007.10.004",
language = "English (US)",
volume = "112",
pages = "2272--2283",
journal = "Remote Sensing of Environment",
issn = "0034-4257",
publisher = "Elsevier Inc.",
number = "5",

}

TY - JOUR

T1 - Mapping land-cover modifications over large areas

T2 - A comparison of machine learning algorithms

AU - Rogan, John

AU - Franklin, Janet

AU - Stow, Doug

AU - Miller, Jennifer

AU - Woodcock, Curtis

AU - Roberts, Dar

PY - 2008/5/15

Y1 - 2008/5/15

N2 - Large area land-cover monitoring scenarios, involving large volumes of data, are becoming more prevalent in remote sensing applications. Thus, there is a pressing need for increased automation in the change mapping process. The objective of this research is to compare the performance of three machine learning algorithms (MLAs); two classification tree software routines (S-plus and C4.5) and an artificial neural network (ARTMAP), in the context of mapping land-cover modifications in northern and southern California study sites between 1990/91 and 1996. Comparisons were based on several criteria: overall accuracy, sensitivity to data set size and variation, and noise. ARTMAP produced the most accurate maps overall (∼ 84%), for two study areas - in southern and northern California, and was most resistant to training data deficiencies. The change map generated using ARTMAP has similar accuracies to a human-interpreted map produced by the U.S. Forest Service in the southern study area. ARTMAP appears to be robust and accurate for automated, large area change monitoring as it performed equally well across the diverse study areas with minimal human intervention in the classification process.

AB - Large area land-cover monitoring scenarios, involving large volumes of data, are becoming more prevalent in remote sensing applications. Thus, there is a pressing need for increased automation in the change mapping process. The objective of this research is to compare the performance of three machine learning algorithms (MLAs); two classification tree software routines (S-plus and C4.5) and an artificial neural network (ARTMAP), in the context of mapping land-cover modifications in northern and southern California study sites between 1990/91 and 1996. Comparisons were based on several criteria: overall accuracy, sensitivity to data set size and variation, and noise. ARTMAP produced the most accurate maps overall (∼ 84%), for two study areas - in southern and northern California, and was most resistant to training data deficiencies. The change map generated using ARTMAP has similar accuracies to a human-interpreted map produced by the U.S. Forest Service in the southern study area. ARTMAP appears to be robust and accurate for automated, large area change monitoring as it performed equally well across the diverse study areas with minimal human intervention in the classification process.

KW - Land-cover change

KW - Large area monitoring

KW - Machine learning

UR - http://www.scopus.com/inward/record.url?scp=41249103454&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=41249103454&partnerID=8YFLogxK

U2 - 10.1016/j.rse.2007.10.004

DO - 10.1016/j.rse.2007.10.004

M3 - Article

AN - SCOPUS:41249103454

VL - 112

SP - 2272

EP - 2283

JO - Remote Sensing of Environment

JF - Remote Sensing of Environment

SN - 0034-4257

IS - 5

ER -