Voxel space analysis of terrestrial laser scans in forests for wind field modelling

Abstract

Meteorological simulation tools to model gas exchange phenomena within forests require well defined information of forest structure (e.g., 3D forest models) as a basis for the computation of the turbulent flow shaped by the drag of the vegetation. The paper describes techniques to obtain 3D data describing forest stands from dense terrestrial laser scanner point clouds. In a first step, stems are automatically detected from the laser scanner data, forming a basis for the determination of tree density, distance patterns and average stem distance. In a second step, the 3D point cloud is translated into a voxel structure representing the forest. A method to segment voxel clusters with the goal of a tree wise interpretation, is presented. From this voxel structure, drag coefficients can be derived via the local density and distribution of stems, branches and leaves. Therefore different voxel attributes are calculated.