Extracting Individual Tree Metrics from 3D Point Clouds: UAV-LiDAR vs Digital Aerial Photogrammetry.
2024
Taylor, Eleanor | Nichol, Dr Caroline
A forest’s structure has been used to determine how the system functions as a whole. Since the technological advancement of deploying high resolution cameras and LiDAR sensors on uncrewed aerial vehicles (UAVs), these active and passive optical remote sensing techniques have been in competition regarding estimating structural metrics through three-dimensional (3D) point cloud datasets. This study focused on deriving tree height and DBH measurements using a structure from motion (SfM) approach from 246 RGB images, and from discrete-return LiDAR. To fill in gaps in literature where coverage of estimating structural metrics within a UK commercial forest, homogenous in its species, the study area focused on a Scots pine (pinus Sylvestris) stand. Regression analyses were used to determine the accuracy of the approaches in deriving tree height and DBH metrics. These resulted in weak correlation (R2 = 0.33) with high error (RMSE = 1.69) between those derived from point cloud datasets and ground validation measurements. This study highlights how integral extensive field measurements are when modelling structural attributes within forests, for uses within forest monitoring and management schemes. Despite the low correlations, production of the 3D forest reconstructions were successful and individual trees were accurately extracted from the entire forest scene. Thus, calls for future work include the development of more systematic comparisons between the effectiveness of UAV-derived LiDAR and photogrammetric datasets in extracting tree structural attributes from 3D reconstructions of UK commercial forest stands.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par University of Edinburgh
Découvrez la collection de ce fournisseur de données dans AGRIS