Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions
Estimating forest canopy height from large-footprint satellite LiDAR waveforms is challenging given the complex interaction between LiDAR waveforms, terrain, and vegetation, especially in dense tropical and equatorial forests. In this study, canopy height in French Guiana was estimated using multiple linear regression models and the Random Forest technique (RF). This analysis was either based on LiDAR waveform metrics extracted from the GLAS (Geoscience Laser Altimeter System) spaceborne LiDAR data and terrain information derived from the SRTM (Shuttle Radar Topography Mission) DEM (Digital Elevation Model) or on Principal Component Analysis (PCA) of GLAS waveforms. Results show that the best statistical model for estimating forest height based on waveform metrics and digital elevation data is a linear regression of waveform extent, trailing edge extent, and terrain index (RMSE of 3.7 m). For the PCA based models, better canopy height estimation results were observed using a regression model that incorporated both the first 13 principal components (PCs) and the waveform extent (RMSE = 3.8 m). Random Forest regressions revealed that the best configuration for canopy height estimation used all the following metrics: waveform extent, leading edge, trailing edge, and terrain index (RMSE = 3.4 m). Waveform extent was the variable that best explained canopy height, with an importance factor almost three times higher than those for the other three metrics (leading edge, trailing edge, and terrain index). Furthermore, the Random Forest regression incorporating the first 13 PCs and the waveform extent had a slightly-improved canopy height estimation in comparison to the linear model, with an RMSE of 3.6 m. In conclusion, multiple linear regressions and RF regressions provided canopy height estimations with similar precision using either LiDAR metrics or PCs. However, a regression model (linear regression or RF) based on the PCA of waveform samples with waveform extent information is an interesting alternative for canopy height estimation as it does not require several metrics that are difficult to derive from GLAS waveforms in dense forests, such as those in French Guiana.
Main Authors: | , , , , , , , |
---|---|
Format: | article biblioteca |
Language: | eng |
Subjects: | K01 - Foresterie - Considérations générales, U30 - Méthodes de recherche, U10 - Informatique, mathématiques et statistiques, P01 - Conservation de la nature et ressources foncières, forêt tropicale, http://aims.fao.org/aos/agrovoc/c_24904, http://aims.fao.org/aos/agrovoc/c_3093, http://aims.fao.org/aos/agrovoc/c_3081, |
Online Access: | http://agritrop.cirad.fr/574501/ http://agritrop.cirad.fr/574501/1/document_574501.pdf |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
dig-cirad-fr-574501 |
---|---|
record_format |
koha |
spelling |
dig-cirad-fr-5745012022-04-15T09:18:14Z http://agritrop.cirad.fr/574501/ http://agritrop.cirad.fr/574501/ Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions. Fayad Ibrahim, Baghdadi Nicolas, Bailly Jean Stéphane, Barbier Nicolas, Gond Valéry, El Hajj Mahmoud, Fabre Frédéric, Bourgine Bernard. 2014. Remote Sensing, 6 (12) : 11883-11914.https://doi.org/10.3390/rs61211883 <https://doi.org/10.3390/rs61211883> Researchers Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions Fayad, Ibrahim Baghdadi, Nicolas Bailly, Jean Stéphane Barbier, Nicolas Gond, Valéry El Hajj, Mahmoud Fabre, Frédéric Bourgine, Bernard eng 2014 Remote Sensing K01 - Foresterie - Considérations générales U30 - Méthodes de recherche U10 - Informatique, mathématiques et statistiques P01 - Conservation de la nature et ressources foncières forêt tropicale http://aims.fao.org/aos/agrovoc/c_24904 Guyane française France http://aims.fao.org/aos/agrovoc/c_3093 http://aims.fao.org/aos/agrovoc/c_3081 Estimating forest canopy height from large-footprint satellite LiDAR waveforms is challenging given the complex interaction between LiDAR waveforms, terrain, and vegetation, especially in dense tropical and equatorial forests. In this study, canopy height in French Guiana was estimated using multiple linear regression models and the Random Forest technique (RF). This analysis was either based on LiDAR waveform metrics extracted from the GLAS (Geoscience Laser Altimeter System) spaceborne LiDAR data and terrain information derived from the SRTM (Shuttle Radar Topography Mission) DEM (Digital Elevation Model) or on Principal Component Analysis (PCA) of GLAS waveforms. Results show that the best statistical model for estimating forest height based on waveform metrics and digital elevation data is a linear regression of waveform extent, trailing edge extent, and terrain index (RMSE of 3.7 m). For the PCA based models, better canopy height estimation results were observed using a regression model that incorporated both the first 13 principal components (PCs) and the waveform extent (RMSE = 3.8 m). Random Forest regressions revealed that the best configuration for canopy height estimation used all the following metrics: waveform extent, leading edge, trailing edge, and terrain index (RMSE = 3.4 m). Waveform extent was the variable that best explained canopy height, with an importance factor almost three times higher than those for the other three metrics (leading edge, trailing edge, and terrain index). Furthermore, the Random Forest regression incorporating the first 13 PCs and the waveform extent had a slightly-improved canopy height estimation in comparison to the linear model, with an RMSE of 3.6 m. In conclusion, multiple linear regressions and RF regressions provided canopy height estimations with similar precision using either LiDAR metrics or PCs. However, a regression model (linear regression or RF) based on the PCA of waveform samples with waveform extent information is an interesting alternative for canopy height estimation as it does not require several metrics that are difficult to derive from GLAS waveforms in dense forests, such as those in French Guiana. article info:eu-repo/semantics/article Journal Article info:eu-repo/semantics/publishedVersion http://agritrop.cirad.fr/574501/1/document_574501.pdf application/pdf Cirad license info:eu-repo/semantics/openAccess https://agritrop.cirad.fr/mention_legale.html https://doi.org/10.3390/rs61211883 10.3390/rs61211883 info:eu-repo/semantics/altIdentifier/doi/10.3390/rs61211883 info:eu-repo/semantics/altIdentifier/purl/https://doi.org/10.3390/rs61211883 |
institution |
CIRAD FR |
collection |
DSpace |
country |
Francia |
countrycode |
FR |
component |
Bibliográfico |
access |
En linea |
databasecode |
dig-cirad-fr |
tag |
biblioteca |
region |
Europa del Oeste |
libraryname |
Biblioteca del CIRAD Francia |
language |
eng |
topic |
K01 - Foresterie - Considérations générales U30 - Méthodes de recherche U10 - Informatique, mathématiques et statistiques P01 - Conservation de la nature et ressources foncières forêt tropicale http://aims.fao.org/aos/agrovoc/c_24904 http://aims.fao.org/aos/agrovoc/c_3093 http://aims.fao.org/aos/agrovoc/c_3081 K01 - Foresterie - Considérations générales U30 - Méthodes de recherche U10 - Informatique, mathématiques et statistiques P01 - Conservation de la nature et ressources foncières forêt tropicale http://aims.fao.org/aos/agrovoc/c_24904 http://aims.fao.org/aos/agrovoc/c_3093 http://aims.fao.org/aos/agrovoc/c_3081 |
spellingShingle |
K01 - Foresterie - Considérations générales U30 - Méthodes de recherche U10 - Informatique, mathématiques et statistiques P01 - Conservation de la nature et ressources foncières forêt tropicale http://aims.fao.org/aos/agrovoc/c_24904 http://aims.fao.org/aos/agrovoc/c_3093 http://aims.fao.org/aos/agrovoc/c_3081 K01 - Foresterie - Considérations générales U30 - Méthodes de recherche U10 - Informatique, mathématiques et statistiques P01 - Conservation de la nature et ressources foncières forêt tropicale http://aims.fao.org/aos/agrovoc/c_24904 http://aims.fao.org/aos/agrovoc/c_3093 http://aims.fao.org/aos/agrovoc/c_3081 Fayad, Ibrahim Baghdadi, Nicolas Bailly, Jean Stéphane Barbier, Nicolas Gond, Valéry El Hajj, Mahmoud Fabre, Frédéric Bourgine, Bernard Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions |
description |
Estimating forest canopy height from large-footprint satellite LiDAR waveforms is challenging given the complex interaction between LiDAR waveforms, terrain, and vegetation, especially in dense tropical and equatorial forests. In this study, canopy height in French Guiana was estimated using multiple linear regression models and the Random Forest technique (RF). This analysis was either based on LiDAR waveform metrics extracted from the GLAS (Geoscience Laser Altimeter System) spaceborne LiDAR data and terrain information derived from the SRTM (Shuttle Radar Topography Mission) DEM (Digital Elevation Model) or on Principal Component Analysis (PCA) of GLAS waveforms. Results show that the best statistical model for estimating forest height based on waveform metrics and digital elevation data is a linear regression of waveform extent, trailing edge extent, and terrain index (RMSE of 3.7 m). For the PCA based models, better canopy height estimation results were observed using a regression model that incorporated both the first 13 principal components (PCs) and the waveform extent (RMSE = 3.8 m). Random Forest regressions revealed that the best configuration for canopy height estimation used all the following metrics: waveform extent, leading edge, trailing edge, and terrain index (RMSE = 3.4 m). Waveform extent was the variable that best explained canopy height, with an importance factor almost three times higher than those for the other three metrics (leading edge, trailing edge, and terrain index). Furthermore, the Random Forest regression incorporating the first 13 PCs and the waveform extent had a slightly-improved canopy height estimation in comparison to the linear model, with an RMSE of 3.6 m. In conclusion, multiple linear regressions and RF regressions provided canopy height estimations with similar precision using either LiDAR metrics or PCs. However, a regression model (linear regression or RF) based on the PCA of waveform samples with waveform extent information is an interesting alternative for canopy height estimation as it does not require several metrics that are difficult to derive from GLAS waveforms in dense forests, such as those in French Guiana. |
format |
article |
topic_facet |
K01 - Foresterie - Considérations générales U30 - Méthodes de recherche U10 - Informatique, mathématiques et statistiques P01 - Conservation de la nature et ressources foncières forêt tropicale http://aims.fao.org/aos/agrovoc/c_24904 http://aims.fao.org/aos/agrovoc/c_3093 http://aims.fao.org/aos/agrovoc/c_3081 |
author |
Fayad, Ibrahim Baghdadi, Nicolas Bailly, Jean Stéphane Barbier, Nicolas Gond, Valéry El Hajj, Mahmoud Fabre, Frédéric Bourgine, Bernard |
author_facet |
Fayad, Ibrahim Baghdadi, Nicolas Bailly, Jean Stéphane Barbier, Nicolas Gond, Valéry El Hajj, Mahmoud Fabre, Frédéric Bourgine, Bernard |
author_sort |
Fayad, Ibrahim |
title |
Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions |
title_short |
Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions |
title_full |
Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions |
title_fullStr |
Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions |
title_full_unstemmed |
Canopy height estimation in French Guiana with LiDAR ICESat/GLAS data using principal component analysis and random forest regressions |
title_sort |
canopy height estimation in french guiana with lidar icesat/glas data using principal component analysis and random forest regressions |
url |
http://agritrop.cirad.fr/574501/ http://agritrop.cirad.fr/574501/1/document_574501.pdf |
work_keys_str_mv |
AT fayadibrahim canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions AT baghdadinicolas canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions AT baillyjeanstephane canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions AT barbiernicolas canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions AT gondvalery canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions AT elhajjmahmoud canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions AT fabrefrederic canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions AT bourginebernard canopyheightestimationinfrenchguianawithlidaricesatglasdatausingprincipalcomponentanalysisandrandomforestregressions |
_version_ |
1758024311611850752 |