Binary partition tree as a hyperspectral segmentation tool for tropical rainforests

Guillaume Tochon, Jean Baptiste Feret, Roberta E. Martin, Raul Tupayachi, Jocelyn Chanussot, Gregory P. Asner

Research output: Contribution to conferencePaper

7 Citations (Scopus)

Abstract

Individual tree crown delineation in tropical forests is of great interest for ecological applications. In this paper we propose a method for hyperspectral image segmentation based on binary tree partitioning. The initial partition is obtained from a watershed transformation in order to make the method computationally more efficient. Then we use a non-parametric region model based on histograms to characterize the regions and the diffusion distance to define the region merging order. The pruning strategy is based on the discontinuity of size increment observed when iteratively merging the regions. The segmentation quality is assessed visually and appears to perform well on most cases, but tree delineation could be improved by including structural information derived from LiDAR data.

Original languageEnglish (US)
Pages6368-6371
Number of pages4
DOIs
StatePublished - Dec 1 2012
Externally publishedYes
Event2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012 - Munich, Germany
Duration: Jul 22 2012Jul 27 2012

Conference

Conference2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012
CountryGermany
CityMunich
Period7/22/127/27/12

Fingerprint

Merging
rainforest
segmentation
Binary trees
Watersheds
Image segmentation
pruning
histogram
tropical forest
discontinuity
partitioning
watershed
method

Keywords

  • Binary partition tree
  • Carnegie airborne observatory
  • hyperspectral imagery
  • segmentation
  • tree crown delineation
  • tropical forest

ASJC Scopus subject areas

  • Computer Science Applications
  • Earth and Planetary Sciences(all)

Cite this

Tochon, G., Feret, J. B., Martin, R. E., Tupayachi, R., Chanussot, J., & Asner, G. P. (2012). Binary partition tree as a hyperspectral segmentation tool for tropical rainforests. 6368-6371. Paper presented at 2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012, Munich, Germany. https://doi.org/10.1109/IGARSS.2012.6352716

Binary partition tree as a hyperspectral segmentation tool for tropical rainforests. / Tochon, Guillaume; Feret, Jean Baptiste; Martin, Roberta E.; Tupayachi, Raul; Chanussot, Jocelyn; Asner, Gregory P.

2012. 6368-6371 Paper presented at 2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012, Munich, Germany.

Research output: Contribution to conferencePaper

Tochon, G, Feret, JB, Martin, RE, Tupayachi, R, Chanussot, J & Asner, GP 2012, 'Binary partition tree as a hyperspectral segmentation tool for tropical rainforests', Paper presented at 2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012, Munich, Germany, 7/22/12 - 7/27/12 pp. 6368-6371. https://doi.org/10.1109/IGARSS.2012.6352716
Tochon G, Feret JB, Martin RE, Tupayachi R, Chanussot J, Asner GP. Binary partition tree as a hyperspectral segmentation tool for tropical rainforests. 2012. Paper presented at 2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012, Munich, Germany. https://doi.org/10.1109/IGARSS.2012.6352716
Tochon, Guillaume ; Feret, Jean Baptiste ; Martin, Roberta E. ; Tupayachi, Raul ; Chanussot, Jocelyn ; Asner, Gregory P. / Binary partition tree as a hyperspectral segmentation tool for tropical rainforests. Paper presented at 2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012, Munich, Germany.4 p.
@conference{b2b486b7c9794030ae9777e44df27970,
title = "Binary partition tree as a hyperspectral segmentation tool for tropical rainforests",
abstract = "Individual tree crown delineation in tropical forests is of great interest for ecological applications. In this paper we propose a method for hyperspectral image segmentation based on binary tree partitioning. The initial partition is obtained from a watershed transformation in order to make the method computationally more efficient. Then we use a non-parametric region model based on histograms to characterize the regions and the diffusion distance to define the region merging order. The pruning strategy is based on the discontinuity of size increment observed when iteratively merging the regions. The segmentation quality is assessed visually and appears to perform well on most cases, but tree delineation could be improved by including structural information derived from LiDAR data.",
keywords = "Binary partition tree, Carnegie airborne observatory, hyperspectral imagery, segmentation, tree crown delineation, tropical forest",
author = "Guillaume Tochon and Feret, {Jean Baptiste} and Martin, {Roberta E.} and Raul Tupayachi and Jocelyn Chanussot and Asner, {Gregory P.}",
year = "2012",
month = "12",
day = "1",
doi = "10.1109/IGARSS.2012.6352716",
language = "English (US)",
pages = "6368--6371",
note = "2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012 ; Conference date: 22-07-2012 Through 27-07-2012",

}

TY - CONF

T1 - Binary partition tree as a hyperspectral segmentation tool for tropical rainforests

AU - Tochon, Guillaume

AU - Feret, Jean Baptiste

AU - Martin, Roberta E.

AU - Tupayachi, Raul

AU - Chanussot, Jocelyn

AU - Asner, Gregory P.

PY - 2012/12/1

Y1 - 2012/12/1

N2 - Individual tree crown delineation in tropical forests is of great interest for ecological applications. In this paper we propose a method for hyperspectral image segmentation based on binary tree partitioning. The initial partition is obtained from a watershed transformation in order to make the method computationally more efficient. Then we use a non-parametric region model based on histograms to characterize the regions and the diffusion distance to define the region merging order. The pruning strategy is based on the discontinuity of size increment observed when iteratively merging the regions. The segmentation quality is assessed visually and appears to perform well on most cases, but tree delineation could be improved by including structural information derived from LiDAR data.

AB - Individual tree crown delineation in tropical forests is of great interest for ecological applications. In this paper we propose a method for hyperspectral image segmentation based on binary tree partitioning. The initial partition is obtained from a watershed transformation in order to make the method computationally more efficient. Then we use a non-parametric region model based on histograms to characterize the regions and the diffusion distance to define the region merging order. The pruning strategy is based on the discontinuity of size increment observed when iteratively merging the regions. The segmentation quality is assessed visually and appears to perform well on most cases, but tree delineation could be improved by including structural information derived from LiDAR data.

KW - Binary partition tree

KW - Carnegie airborne observatory

KW - hyperspectral imagery

KW - segmentation

KW - tree crown delineation

KW - tropical forest

UR - http://www.scopus.com/inward/record.url?scp=84873153579&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873153579&partnerID=8YFLogxK

U2 - 10.1109/IGARSS.2012.6352716

DO - 10.1109/IGARSS.2012.6352716

M3 - Paper

AN - SCOPUS:84873153579

SP - 6368

EP - 6371

ER -