Spectral band selection for vegetation properties retrieval using Gaussian processes regression

Jochem Verrelst, Juan Pablo Rivera, Anatoly Gitelson, Jesus Delegido, José Moreno, Gustau Camps-Valls

Research output: Contribution to journalArticle

34 Citations (Scopus)

Abstract

With current and upcoming imaging spectrometers, automated band analysis techniques are needed to enable efficient identification of most informative bands to facilitate optimized processing of spectral data into estimates of biophysical variables. This paper introduces an automated spectral band analysis tool (BAT) based on Gaussian processes regression (GPR) for the spectral analysis of vegetation properties. The GPR-BAT procedure sequentially backwards removes the least contributing band in the regression model for a given variable until only one band is kept. GPR-BAT is implemented within the framework of the free ARTMO's MLRA (machine learning regression algorithms) toolbox, which is dedicated to the transforming of optical remote sensing images into biophysical products. GPR-BAT allows (1) to identify the most informative bands in relating spectral data to a biophysical variable, and (2) to find the least number of bands that preserve optimized accurate predictions. To illustrate its utility, two hyperspectral datasets were analyzed for most informative bands: (1) a field hyperspectral dataset (400–1100 nm at 2 nm resolution: 301 bands) with leaf chlorophyll content (LCC) and green leaf area index (gLAI) collected for maize and soybean (Nebraska, US); and (2) an airborne HyMap dataset (430–2490 nm: 125 bands) with LAI and canopy water content (CWC) collected for a variety of crops (Barrax, Spain). For each of these biophysical variables, optimized retrieval accuracies can be achieved with just 4 to 9 well-identified bands, and performance was largely improved over using all bands. A PROSAIL global sensitivity analysis was run to interpret the validity of these bands. Cross-validated RCV2 (NRMSECV) accuracies for optimized GPR models were 0.79 (12.9%) for LCC, 0.94 (7.2%) for gLAI, 0.95 (6.5%) for LAI and 0.95 (7.2%) for CWC. This study concludes that a wise band selection of hyperspectral data is strictly required for optimal vegetation properties mapping.

Original languageEnglish (US)
Pages (from-to)554-567
Number of pages14
JournalInternational Journal of Applied Earth Observation and Geoinformation
Volume52
DOIs
StatePublished - Oct 1 2016

Fingerprint

leaf area index
vegetation
Chlorophyll
Water content
chlorophyll
water content
canopy
Spectrum analysis
Sensitivity analysis
Crops
Learning systems
Spectrometers
Remote sensing
spectral analysis
soybean
sensitivity analysis
Imaging techniques
spectrometer
maize
spectral band

Keywords

  • ARTMO
  • Band selection
  • Gaussian processes regression (GPR)
  • Hyperspectral
  • Machine learning
  • Vegetation properties

ASJC Scopus subject areas

  • Global and Planetary Change
  • Earth-Surface Processes
  • Computers in Earth Sciences
  • Management, Monitoring, Policy and Law

Cite this

Spectral band selection for vegetation properties retrieval using Gaussian processes regression. / Verrelst, Jochem; Rivera, Juan Pablo; Gitelson, Anatoly; Delegido, Jesus; Moreno, José; Camps-Valls, Gustau.

In: International Journal of Applied Earth Observation and Geoinformation, Vol. 52, 01.10.2016, p. 554-567.

Research output: Contribution to journalArticle

Verrelst, Jochem ; Rivera, Juan Pablo ; Gitelson, Anatoly ; Delegido, Jesus ; Moreno, José ; Camps-Valls, Gustau. / Spectral band selection for vegetation properties retrieval using Gaussian processes regression. In: International Journal of Applied Earth Observation and Geoinformation. 2016 ; Vol. 52. pp. 554-567.
@article{41669ebb129b436f9c3ae9da978ff4ad,
title = "Spectral band selection for vegetation properties retrieval using Gaussian processes regression",
abstract = "With current and upcoming imaging spectrometers, automated band analysis techniques are needed to enable efficient identification of most informative bands to facilitate optimized processing of spectral data into estimates of biophysical variables. This paper introduces an automated spectral band analysis tool (BAT) based on Gaussian processes regression (GPR) for the spectral analysis of vegetation properties. The GPR-BAT procedure sequentially backwards removes the least contributing band in the regression model for a given variable until only one band is kept. GPR-BAT is implemented within the framework of the free ARTMO's MLRA (machine learning regression algorithms) toolbox, which is dedicated to the transforming of optical remote sensing images into biophysical products. GPR-BAT allows (1) to identify the most informative bands in relating spectral data to a biophysical variable, and (2) to find the least number of bands that preserve optimized accurate predictions. To illustrate its utility, two hyperspectral datasets were analyzed for most informative bands: (1) a field hyperspectral dataset (400–1100 nm at 2 nm resolution: 301 bands) with leaf chlorophyll content (LCC) and green leaf area index (gLAI) collected for maize and soybean (Nebraska, US); and (2) an airborne HyMap dataset (430–2490 nm: 125 bands) with LAI and canopy water content (CWC) collected for a variety of crops (Barrax, Spain). For each of these biophysical variables, optimized retrieval accuracies can be achieved with just 4 to 9 well-identified bands, and performance was largely improved over using all bands. A PROSAIL global sensitivity analysis was run to interpret the validity of these bands. Cross-validated RCV2 (NRMSECV) accuracies for optimized GPR models were 0.79 (12.9{\%}) for LCC, 0.94 (7.2{\%}) for gLAI, 0.95 (6.5{\%}) for LAI and 0.95 (7.2{\%}) for CWC. This study concludes that a wise band selection of hyperspectral data is strictly required for optimal vegetation properties mapping.",
keywords = "ARTMO, Band selection, Gaussian processes regression (GPR), Hyperspectral, Machine learning, Vegetation properties",
author = "Jochem Verrelst and Rivera, {Juan Pablo} and Anatoly Gitelson and Jesus Delegido and Jos{\'e} Moreno and Gustau Camps-Valls",
year = "2016",
month = "10",
day = "1",
doi = "10.1016/j.jag.2016.07.016",
language = "English (US)",
volume = "52",
pages = "554--567",
journal = "International Journal of Applied Earth Observation and Geoinformation",
issn = "1569-8432",
publisher = "Elsevier",

}

TY - JOUR

T1 - Spectral band selection for vegetation properties retrieval using Gaussian processes regression

AU - Verrelst, Jochem

AU - Rivera, Juan Pablo

AU - Gitelson, Anatoly

AU - Delegido, Jesus

AU - Moreno, José

AU - Camps-Valls, Gustau

PY - 2016/10/1

Y1 - 2016/10/1

N2 - With current and upcoming imaging spectrometers, automated band analysis techniques are needed to enable efficient identification of most informative bands to facilitate optimized processing of spectral data into estimates of biophysical variables. This paper introduces an automated spectral band analysis tool (BAT) based on Gaussian processes regression (GPR) for the spectral analysis of vegetation properties. The GPR-BAT procedure sequentially backwards removes the least contributing band in the regression model for a given variable until only one band is kept. GPR-BAT is implemented within the framework of the free ARTMO's MLRA (machine learning regression algorithms) toolbox, which is dedicated to the transforming of optical remote sensing images into biophysical products. GPR-BAT allows (1) to identify the most informative bands in relating spectral data to a biophysical variable, and (2) to find the least number of bands that preserve optimized accurate predictions. To illustrate its utility, two hyperspectral datasets were analyzed for most informative bands: (1) a field hyperspectral dataset (400–1100 nm at 2 nm resolution: 301 bands) with leaf chlorophyll content (LCC) and green leaf area index (gLAI) collected for maize and soybean (Nebraska, US); and (2) an airborne HyMap dataset (430–2490 nm: 125 bands) with LAI and canopy water content (CWC) collected for a variety of crops (Barrax, Spain). For each of these biophysical variables, optimized retrieval accuracies can be achieved with just 4 to 9 well-identified bands, and performance was largely improved over using all bands. A PROSAIL global sensitivity analysis was run to interpret the validity of these bands. Cross-validated RCV2 (NRMSECV) accuracies for optimized GPR models were 0.79 (12.9%) for LCC, 0.94 (7.2%) for gLAI, 0.95 (6.5%) for LAI and 0.95 (7.2%) for CWC. This study concludes that a wise band selection of hyperspectral data is strictly required for optimal vegetation properties mapping.

AB - With current and upcoming imaging spectrometers, automated band analysis techniques are needed to enable efficient identification of most informative bands to facilitate optimized processing of spectral data into estimates of biophysical variables. This paper introduces an automated spectral band analysis tool (BAT) based on Gaussian processes regression (GPR) for the spectral analysis of vegetation properties. The GPR-BAT procedure sequentially backwards removes the least contributing band in the regression model for a given variable until only one band is kept. GPR-BAT is implemented within the framework of the free ARTMO's MLRA (machine learning regression algorithms) toolbox, which is dedicated to the transforming of optical remote sensing images into biophysical products. GPR-BAT allows (1) to identify the most informative bands in relating spectral data to a biophysical variable, and (2) to find the least number of bands that preserve optimized accurate predictions. To illustrate its utility, two hyperspectral datasets were analyzed for most informative bands: (1) a field hyperspectral dataset (400–1100 nm at 2 nm resolution: 301 bands) with leaf chlorophyll content (LCC) and green leaf area index (gLAI) collected for maize and soybean (Nebraska, US); and (2) an airborne HyMap dataset (430–2490 nm: 125 bands) with LAI and canopy water content (CWC) collected for a variety of crops (Barrax, Spain). For each of these biophysical variables, optimized retrieval accuracies can be achieved with just 4 to 9 well-identified bands, and performance was largely improved over using all bands. A PROSAIL global sensitivity analysis was run to interpret the validity of these bands. Cross-validated RCV2 (NRMSECV) accuracies for optimized GPR models were 0.79 (12.9%) for LCC, 0.94 (7.2%) for gLAI, 0.95 (6.5%) for LAI and 0.95 (7.2%) for CWC. This study concludes that a wise band selection of hyperspectral data is strictly required for optimal vegetation properties mapping.

KW - ARTMO

KW - Band selection

KW - Gaussian processes regression (GPR)

KW - Hyperspectral

KW - Machine learning

KW - Vegetation properties

UR - http://www.scopus.com/inward/record.url?scp=84997822665&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84997822665&partnerID=8YFLogxK

U2 - 10.1016/j.jag.2016.07.016

DO - 10.1016/j.jag.2016.07.016

M3 - Article

AN - SCOPUS:84997822665

VL - 52

SP - 554

EP - 567

JO - International Journal of Applied Earth Observation and Geoinformation

JF - International Journal of Applied Earth Observation and Geoinformation

SN - 1569-8432

ER -