Modélisation de la canopée forestière par photogrammétrie depuis des images acquises par drone

Abstract

Recent development of small unmanned aerial systems opens the door for their use in forest mapping, as both the spatial and temporal resolution of drone imagery better suit local-scale investigation than traditional remote sensing tools. An original photogrammetric workflow, based on the open source toolbox MICMAC, was set up to model the forest canopy surface from low-altitude aerial images. In combination with a LiDAR digital terrain model, the elevation of vegetation was determined after a fine co-registration of the photogrammetric canopy surface model. The investigation of different images matching strategies is performed within MICMAC and their performance in modelling the outer canopy is compared. Although photogrammetric reconstruction do not account for small peaks and gaps in the canopy surface, our results have shown the potential of drones to accurately estimate canopy height in broadleaf stands, confirming thus the feasibility of modeling height growth from UAV images time series.