Skip to Main content Skip to Navigation
Theses

Spécification de paramètres techniques et stratégie d'échantillonnage pour la conception de nouveaux capteurs lidars dédiés à la cartographie de forêts

Abstract : Foresters need tools to map the tree species, tree heights, stand structure and biomass. Although the airborne lidar (Light detection and ranging) technology does not give access to all these variables, it can provide quick, accurate and spatially explicit measurements of tree heights and biomass over large surfaces. However, lidar systems currently used have not been specially designed to performed vegetation studies. The adjustment of the technical characteristics of such systems is expected to improve the accuracy of retrieved forest parameters. Consequently, the objective of this thesis is to determine configurations of lidar sensors dedicated to the study of forest vegetation, and to propose methods designed to extract forest parameters depending on the different configurations. The ability of different resolutions (footprint size and spatial sampling), wavelengths and sampling modes of the backscattered signal to measure forest parameters (canopy height and density, crown size and also volume and biomass) was evaluated. The studies were conducted from tree to stand level, on experimental or simulated data. In a first part, we developed methods to process classic airborne lidar data (scanner system, tens of centimeter footprint, 5 measurements/m², near-infrared laser) for the estimation of the biomass of individual trees. In this study, we demonstrated the contribution of new data called "Full-waveform" (recording the entire signal) compared to traditional multi-echoes data (extraction of the most significant echoes). In a second part, we performed an experiment using a Comissariat of Energy Atomique (CEA)'s lidar prototype onboard an ultra-light aircraft (profiler system, 2.4 m footprint, 2.4 m spacing between two measurements along the flight line, ultraviolet laser). Such a configuration did not allow to measure individual trees, but we were able to study variations in forest structure at the plot level (30 m diameter). Having demonstrated the ability of an ultraviolet lidar to perform vegetation studies, this sensor opens the way to the development of bi-functional lidar for both atmosphere and vegetation remote sensing. In a third part, large footprint lidar signals (tens of meters on the ground) were simulated from the aggregation of classical airborne lidar signals. We proposed a method for modeling the signal dynamics of satellite lidars in different forest types, in order to calibrate the energy to emit for a future space-borne mission. The main problem with large-footprint size is the strong influence of topography on accurate measurements of tree heights in steep areas. We consequently developed a method to correct this effect, thus increasing the accuracy of tree height retrieval. This approach also opened new perspectives in topography assessment from large-footprint data in forest environments.
Complete list of metadata

Cited literature [138 references]  Display  Hide  Download

https://pastel.archives-ouvertes.fr/tel-01127548
Contributor : ABES STAR :  Contact
Submitted on : Saturday, March 7, 2015 - 3:24:58 AM
Last modification on : Friday, August 5, 2022 - 2:38:11 PM
Long-term archiving on: : Monday, June 8, 2015 - 3:51:38 PM

File

2011AGPT0086.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01127548, version 1

Citation

Tristan Allouis. Spécification de paramètres techniques et stratégie d'échantillonnage pour la conception de nouveaux capteurs lidars dédiés à la cartographie de forêts. Sylviculture, foresterie. AgroParisTech, 2011. Français. ⟨NNT : 2011AGPT0086⟩. ⟨tel-01127548⟩

Share

Metrics

Record views

319

Files downloads

109