AdTree

AdTree reconstructs detailed three-dimensional tree models from laser-scanned point cloud data to enable accurate geometric characterization of tree architecture.

Key Features:

• Automatic Reconstruction: Automates 3D tree modeling from laser-scanned point cloud data using computational extraction methods.
• High Accuracy: Produces models with a reported fitting error of less than 10 cm.
• Robust Methodology: Applies algorithmic extraction and refinement to recover tree structure from point clouds.

Scientific Applications:

• Tree Attribute Estimation: Enables precise calculation of tree attributes for ecological and forestry research.
• Urban Landscape Visualization: Supports creation of realistic urban green-space models for planning and design studies.

Methodology:

Initial skeleton extraction from the input point cloud using a minimum spanning tree approach based on Dijkstra's shortest-path algorithm; iterative pruning of the initial skeleton to remove redundant components; and an optimization-based method that fits a sequence of cylinders to approximate branch geometry.