Analyse des données lidar aéroportées à Retour d'Onde Complète pour la cartographie des milieux urbains

Abstract : The recent development of airborne full-waveform lidar systems, that are able to provide more than a simple 3D representation of the topographic surfaces, has arisen many questions with respect to the analysis of urban areas, and in relation with traditional multiple pulse sensors. Indeed, the recorded lidar waveforms provide additional information about the geometry and the radiometry of the targets. The final aim of this thesis is to process the raw data in order to achieve a 3D automatic land-cover classification, as a basis for various subsequent applications. The proposed workflow is composed of two main stages. The first step consists in processing the waveform in order to generate an improved 3D point cloud, and retrieve morphological information about these points. Two distinct methods have been proposed. The first one aims to improve the standard approach, which assumes that all echoes can be modelled by the Gaussian curve. The second method hypothesizes a mixture of several models -- each echo is then characterized separately -- and allows to propose a simple physical and flexible solution. Afterwards, the generated 3D points are classified using both spatial features and the morphological ones retrieved from the previous processing step. A supervised approach using Support Vector Machines is adopted in order to label ground, building, and vegetation areas. Furthermore, the classifier is coupled with a feature selection step. Eventually, results with high accuracy are achieved, and enhance the relevance of full-waveform data for automatically mapping urban areas.
Complete list of metadatas

https://pastel.archives-ouvertes.fr/pastel-00566992
Contributor : Clément Mallet <>
Submitted on : Thursday, February 17, 2011 - 4:38:35 PM
Last modification on : Wednesday, February 20, 2019 - 2:41:26 PM

Identifiers

  • HAL Id : pastel-00566992, version 1

Citation

Clément Mallet. Analyse des données lidar aéroportées à Retour d'Onde Complète pour la cartographie des milieux urbains. Traitement du signal et de l'image [eess.SP]. Télécom ParisTech, 2010. Français. ⟨pastel-00566992⟩

Share

Metrics

Record views

985

Files downloads

1399