Classification de données LiDAR bi-spectral topo-bathymétriques par une approche multi-échelle : Application en milieu fluvial

Arthur Le Guennec 1 Dimitri Lague 1 Sébastien Lefèvre 2 Thomas Corpetti 3
2 OBELIX - Environment observation with complex imagery
UBS - Université de Bretagne Sud, IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
3 LETG - Rennes - Littoral, Environnement, Télédétection, Géomatique
LETG - Littoral, Environnement, Télédétection, Géomatique UMR 6554
Abstract : The monitoring of a natural/semi-natural space often requires the automatic identification of the various objects present such as vegetation, soil, water, buildings, etc. In the fluvial context, the detection of bathymetric classes (surface and mainly water bottom) is essential for many applications , such as the management of rivers or streams. In order to obtain them, the topo-bathymetric LiDAR is an interesting tool because it makes it possible to construct two 3D point clouds at a very high spatial resolution of the scenes scanned from two specific wavelengths: 1064 nm and 532 nm. The topographic aspect, by the wavelength 1064 nm, allows to recover the spatial structure of the different objects. The bathymetric aspect comes from the wavelength 532 nm entering the water and sometimes allowing access to the bottom of the river. Thus, classifying the different topo-bathymetric objects from these two points clouds is an essential issue. In this article, we extend the multi-scale approach to characterize spatial structures, which has already demonstrated its performance for point cloud analysis, with new descriptors taking into consideration the bi-spectral aspect. The classification is carried out using a technique of Random Forest which offers the possibility to analyze the most contributive descriptors in the classification and thus to better understand the impact of the bispectral data.
Document type :
Conference papers
Complete list of metadatas

Cited literature [21 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02070083
Contributor : Sébastien Lefèvre <>
Submitted on : Saturday, March 16, 2019 - 5:03:54 PM
Last modification on : Tuesday, July 16, 2019 - 4:44:02 PM
Long-term archiving on : Monday, June 17, 2019 - 3:37:13 PM

File

CFPT2018_paper_leguennec.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02070083, version 1

Citation

Arthur Le Guennec, Dimitri Lague, Sébastien Lefèvre, Thomas Corpetti. Classification de données LiDAR bi-spectral topo-bathymétriques par une approche multi-échelle : Application en milieu fluvial. Conférence Française de Photogrammétrie et de Télédétection, 2018, Marne-la-Vallée, France. ⟨hal-02070083⟩

Share

Metrics

Record views

200

Files downloads

49