The localization and reconstruction of individual trees as well as the extraction of their geometrical parameters is an important field of research in both forestry and remote sensing. While the current state-of-the-art mostly focuses on the exploitation of optical imagery and airborne {LiDAR} data, modern {SAR} sensors have not yet met the interest of the research community in that regard. This paper presents a prototypical processing chain for the reconstruction of individual deciduous trees: First, single-pass multi-baseline {InSAR} data acquired from multiple aspect angles are used for the generation of a layover- and shadow-free 3D point cloud by tomographic {SAR} processing. The resulting point cloud is then segmented by unsupervised mean shift clustering, before ellipsoid models are fitted to the points of each cluster. From these 3D ellipsoids the relevant geometrical tree parameters are extracted. Evaluation with respect to a manually derived reference dataset prove that almost 74% of all trees are successfully segmented and reconstructed, thus providing a promising perspective for further research toward individual tree recognition from {SAR} data.     «

The localization and reconstruction of individual trees as well as the extraction of their geometrical parameters is an important field of research in both forestry and remote sensing. While the current state-of-the-art mostly focuses on the exploitation of optical imagery and airborne {LiDAR} data, modern {SAR} sensors have not yet met the interest of the research community in that regard. This paper presents a prototypical processing chain for the reconstruction of individual deciduous trees:...     »