Efficient Representation of Geometric Tree Models with Level-of-Detail Using Compressed 3D Chain Code

Area of Science:

• Computer Graphics
• Geometric Modeling
• Forestry Science

Background:

• Geometric tree models are crucial for simulations but require significant storage.
• Existing methods often struggle to efficiently represent branching structures and radii.

Purpose of the Study:

• To develop a space-efficient method for representing geometric tree models.
• To enable level-of-detail construction directly from encoded data.

Main Methods:

• A novel differential 3D chain code is proposed to encode orientation changes of branch segments.
• The method encodes complete branching geometry, including branch radii.
• Interpolative encoding is used for tree descriptors and radii sequences.

Main Results:

• Storage requirements for geometric forest models reduced to less than 15% of raw size.
• Structural fidelity of tree models is preserved.
• The method is the first to encode complete branching geometry and radii.

Conclusions:

• The proposed 3D chain code offers a highly compressible and accurate representation for geometric tree models.
• This approach facilitates efficient storage and direct level-of-detail construction for complex forest ecosystems.