Monitoring restored riparian vegetation: how can recent developments in remote sensing sciences help?

Riparian vegetation restoration projects require appropriate tools to monitor actions efficiency. On a large scale remote sensing approaches can provide continuous and detailed data to describe riparian vegetation. In this paper, we illustrated recent developments and perspectives for riparian vegetation monitoring purposes through three examples of image sources: Light Detection And Ranging (LiDAR), radar and Unmanned Aerial Vehicule (UAV) images. We notably focused on the potential of such images to provide 3D information for narrow strips of riparian vegetation with high temporal resolution to allow fine monitoring following restoration program. LiDAR data allows canopy structure identification with a high accuracy level and automatic classifications for heterogeneous riparian corridors. Radar images allow a good identification of riparian vegetation but also of the structure and phenology of vegetation through time with an analysis of the Shannon entropy of the signal. The UAV system used here is a very flexible approach that can easily provide RGB mosaic but also a local digital surface model with very high spatial resolution. Lastly, we discuss the advantages and limitations of each approach from an applied perspective, in terms of flexibility, resolution and technicality.Original Abstract: Le suivi des projets de restauration de la ripisylve necessite des outils specifiques. Dans cet article, nous illustrons et discutons comment les developpements recents dans le domaine de la teledetection permettent une description detaillee, continue et a large echelle des ripisylves restaurees a partir de trois exemples d'images : laser (LiDAR), radar et drone. Nous analysons notamment la capacite et le potentiel de ces images a fournir une information volumetrique de ripisylves etroites avec une forte resolution temporelle afin de permettre un suivi fin des actions de restauration. Les donnees LiDAR permettent une description de la structure de la canopee avec une tres bonne precision ainsi qu'une classification automatique des ripisylves heterogenes. Les images radar permettent une bonne identification non seulement de la vegetation riveraine mais aussi de sa structure et de sa phenologie par analyse de l'entropie du signal. La technologie drone deployee ici est tres flexible et facile a mettre en uvre; elle donne acces a des mosaiques de photographie a tres haute resolution spatiale et a faible resolution spectrale. Mais elle permet aussi de generer facilemen