Revealing the skeleton from imperfect point clouds

Beschreibung

Quantifying our surrounding environment in terms of sizes and orders has always been of interest, because it enables us to visualize, describe and interpret our environment. In the last decade terrestrial laser scanners became available as a tool to measure objects in our surrounding environment. Terrestrial laser scanning samples surfaces with millions of 3D points to be stored as a point cloud. These point clouds contain information of sizes and orders of the sampled objects. Recently, trees became an object of high interest, because they contain a rich and complex structure, which is to be exploited in terms of branch length and diameters and branching hierarchy. Such information can be used for forestry, hydrological, ecological and visualization applications. This thesis introduces the SkelTre algorithm as a skeleton extraction method, which enables us to extract the branching hierarchy and the size parameters of branches from point clouds. SkelTre is the abbreviation of Skeletonization of Trees. Because the data contains only measurement samples on the surface and not inside the object, a skeletonization method was developed which operates only on the given surface samples.
Quantifying our surrounding environment in terms of sizes and orders has always been of interest, because it enables us to visualize, describe and interpret our environment. In the last decade terrestrial laser scanners became available as a tool to measure objects in our surrounding environment. Terrestrial laser scanning samples surfaces with millions of 3D points to be stored as a point cloud. These point clouds contain information of sizes and orders of the sampled objects. Recently, trees became an object of high interest, because they contain a rich and complex structure, which is to be exploited in terms of branch length and diameters and branching hierarchy. Such information can be used for forestry, hydrological, ecological and visualization applications. This thesis introduces the SkelTre algorithm as a skeleton extraction method, which enables us to extract the branching hierarchy and the size parameters of branches from point clouds. SkelTre is the abbreviation of Skeletonization of Trees. Because the data contains only measurement samples on the surface and not inside the object, a skeletonization method was developed which operates only on the given surface samples.

Autor*in

Alexander Bucksch

Stimmen zu »Revealing the skeleton from imperfect point clouds«

Details

Link teilen