Support vector machines for recognition of semi-arid vegetation types using MISR multi-angle imagery

Lihong Su, Mark J. Chopping, Albert Rango, John V. Martonchik, Debra P.C. Peters

Research output: Contribution to journalArticle

55 Citations (Scopus)

Abstract

Accurately mapping community types is one of the main challenges for monitoring arid and semi-arid grasslands with remote sensing. The multi-angle approach has been proven useful for mapping vegetation types in desert grassland. The Multi-angle Imaging Spectro-Radiometer (MISR) provides 4 spectral bands and 9 angular reflectance. In this study, 44 classification experiments have been implemented to find the optimal combination of MISR multi-angular data to mine the information carried by MISR data as effectively as possible. These experiments show the following findings: 1) The combination of MISR's 4 spectral bands at nadir and red and near infrared bands in the C, B, and A cameras observing off-nadir can obtain the best vegetation type differentiation at the community level in New Mexico desert grasslands. 2) The k parameter at red band of Modified-Rahman-Pinty-Verstraete (MRPV) model and the structural scattering index (SSI) can bring useful additional information to land cover classification. The information carried by these two parameters, however, is less than that carried by surface anisotropy patterns described by the MRPV model and a linear semi-empirical kernel-driven bidirectional reflectance distribution function model, the RossThin-LiSparseMODIS (RTnLS) model. These experiments prove that: 1) multi-angular reflectance raise overall classification accuracy from 45.8% for nadir-only reflectance to 60.9%. 2) With surface anisotropy patterns derived from MRPV and RTnLS, an overall accuracy of 68.1% can be obtained when maximum likelihood algorithms are used. 3) Support Vector Machine (SVM) algorithms can raise the classification accuracy to 76.7%. This research shows that multi-angular reflectance, surface anisotropy patterns and SVM algorithms can improve desert vegetation type differentiation importantly.

Original languageEnglish
Pages (from-to)299-311
Number of pages13
JournalRemote Sensing of Environment
Volume107
Issue number1-2
DOIs
StatePublished - 15 Mar 2007

Fingerprint

MISR
radiometers
Radiometers
vegetation types
vegetation type
reflectance
Support vector machines
nadir
imagery
image analysis
Imaging techniques
anisotropy
desert
grassland
taxonomy
deserts
Anisotropy
grasslands
bidirectional reflectance
surface reflectance

Keywords

  • Classification
  • MISR
  • Multi-angle observations
  • Semi-arid vegetation
  • Support vector machine

Cite this

Su, Lihong ; Chopping, Mark J. ; Rango, Albert ; Martonchik, John V. ; Peters, Debra P.C. / Support vector machines for recognition of semi-arid vegetation types using MISR multi-angle imagery. In: Remote Sensing of Environment. 2007 ; Vol. 107, No. 1-2. pp. 299-311.
@article{528cb908ebb64fa8a9b029fc67a29940,
title = "Support vector machines for recognition of semi-arid vegetation types using MISR multi-angle imagery",
abstract = "Accurately mapping community types is one of the main challenges for monitoring arid and semi-arid grasslands with remote sensing. The multi-angle approach has been proven useful for mapping vegetation types in desert grassland. The Multi-angle Imaging Spectro-Radiometer (MISR) provides 4 spectral bands and 9 angular reflectance. In this study, 44 classification experiments have been implemented to find the optimal combination of MISR multi-angular data to mine the information carried by MISR data as effectively as possible. These experiments show the following findings: 1) The combination of MISR's 4 spectral bands at nadir and red and near infrared bands in the C, B, and A cameras observing off-nadir can obtain the best vegetation type differentiation at the community level in New Mexico desert grasslands. 2) The k parameter at red band of Modified-Rahman-Pinty-Verstraete (MRPV) model and the structural scattering index (SSI) can bring useful additional information to land cover classification. The information carried by these two parameters, however, is less than that carried by surface anisotropy patterns described by the MRPV model and a linear semi-empirical kernel-driven bidirectional reflectance distribution function model, the RossThin-LiSparseMODIS (RTnLS) model. These experiments prove that: 1) multi-angular reflectance raise overall classification accuracy from 45.8{\%} for nadir-only reflectance to 60.9{\%}. 2) With surface anisotropy patterns derived from MRPV and RTnLS, an overall accuracy of 68.1{\%} can be obtained when maximum likelihood algorithms are used. 3) Support Vector Machine (SVM) algorithms can raise the classification accuracy to 76.7{\%}. This research shows that multi-angular reflectance, surface anisotropy patterns and SVM algorithms can improve desert vegetation type differentiation importantly.",
keywords = "Classification, MISR, Multi-angle observations, Semi-arid vegetation, Support vector machine",
author = "Lihong Su and Chopping, {Mark J.} and Albert Rango and Martonchik, {John V.} and Peters, {Debra P.C.}",
year = "2007",
month = "3",
day = "15",
doi = "10.1016/j.rse.2006.05.023",
language = "English",
volume = "107",
pages = "299--311",
journal = "Remote Sensing of Environment",
issn = "0034-4257",
publisher = "Elsevier Inc.",
number = "1-2",

}

Support vector machines for recognition of semi-arid vegetation types using MISR multi-angle imagery. / Su, Lihong; Chopping, Mark J.; Rango, Albert; Martonchik, John V.; Peters, Debra P.C.

In: Remote Sensing of Environment, Vol. 107, No. 1-2, 15.03.2007, p. 299-311.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Support vector machines for recognition of semi-arid vegetation types using MISR multi-angle imagery

AU - Su, Lihong

AU - Chopping, Mark J.

AU - Rango, Albert

AU - Martonchik, John V.

AU - Peters, Debra P.C.

PY - 2007/3/15

Y1 - 2007/3/15

N2 - Accurately mapping community types is one of the main challenges for monitoring arid and semi-arid grasslands with remote sensing. The multi-angle approach has been proven useful for mapping vegetation types in desert grassland. The Multi-angle Imaging Spectro-Radiometer (MISR) provides 4 spectral bands and 9 angular reflectance. In this study, 44 classification experiments have been implemented to find the optimal combination of MISR multi-angular data to mine the information carried by MISR data as effectively as possible. These experiments show the following findings: 1) The combination of MISR's 4 spectral bands at nadir and red and near infrared bands in the C, B, and A cameras observing off-nadir can obtain the best vegetation type differentiation at the community level in New Mexico desert grasslands. 2) The k parameter at red band of Modified-Rahman-Pinty-Verstraete (MRPV) model and the structural scattering index (SSI) can bring useful additional information to land cover classification. The information carried by these two parameters, however, is less than that carried by surface anisotropy patterns described by the MRPV model and a linear semi-empirical kernel-driven bidirectional reflectance distribution function model, the RossThin-LiSparseMODIS (RTnLS) model. These experiments prove that: 1) multi-angular reflectance raise overall classification accuracy from 45.8% for nadir-only reflectance to 60.9%. 2) With surface anisotropy patterns derived from MRPV and RTnLS, an overall accuracy of 68.1% can be obtained when maximum likelihood algorithms are used. 3) Support Vector Machine (SVM) algorithms can raise the classification accuracy to 76.7%. This research shows that multi-angular reflectance, surface anisotropy patterns and SVM algorithms can improve desert vegetation type differentiation importantly.

AB - Accurately mapping community types is one of the main challenges for monitoring arid and semi-arid grasslands with remote sensing. The multi-angle approach has been proven useful for mapping vegetation types in desert grassland. The Multi-angle Imaging Spectro-Radiometer (MISR) provides 4 spectral bands and 9 angular reflectance. In this study, 44 classification experiments have been implemented to find the optimal combination of MISR multi-angular data to mine the information carried by MISR data as effectively as possible. These experiments show the following findings: 1) The combination of MISR's 4 spectral bands at nadir and red and near infrared bands in the C, B, and A cameras observing off-nadir can obtain the best vegetation type differentiation at the community level in New Mexico desert grasslands. 2) The k parameter at red band of Modified-Rahman-Pinty-Verstraete (MRPV) model and the structural scattering index (SSI) can bring useful additional information to land cover classification. The information carried by these two parameters, however, is less than that carried by surface anisotropy patterns described by the MRPV model and a linear semi-empirical kernel-driven bidirectional reflectance distribution function model, the RossThin-LiSparseMODIS (RTnLS) model. These experiments prove that: 1) multi-angular reflectance raise overall classification accuracy from 45.8% for nadir-only reflectance to 60.9%. 2) With surface anisotropy patterns derived from MRPV and RTnLS, an overall accuracy of 68.1% can be obtained when maximum likelihood algorithms are used. 3) Support Vector Machine (SVM) algorithms can raise the classification accuracy to 76.7%. This research shows that multi-angular reflectance, surface anisotropy patterns and SVM algorithms can improve desert vegetation type differentiation importantly.

KW - Classification

KW - MISR

KW - Multi-angle observations

KW - Semi-arid vegetation

KW - Support vector machine

UR - http://www.scopus.com/inward/record.url?scp=33846901577&partnerID=8YFLogxK

U2 - 10.1016/j.rse.2006.05.023

DO - 10.1016/j.rse.2006.05.023

M3 - Article

AN - SCOPUS:33846901577

VL - 107

SP - 299

EP - 311

JO - Remote Sensing of Environment

JF - Remote Sensing of Environment

SN - 0034-4257

IS - 1-2

ER -