Realistic Procedural Plant Modeling from Multiple View Images.

Area of Science:

• Computer Graphics
• Computational Photography
• Plant Sciences

Background:

• Generating realistic 3D plant models is challenging.
• Previous methods often require extensive user interaction for segmentation and branch recovery.
• Modeling complex or occluded plant structures, like bushes, remains difficult.

Purpose of the Study:

• To develop an automated procedural modeling technique for realistic plant reconstruction from multi-view photographs.
• To overcome limitations of user-dependent methods in plant modeling.
• To enable the synthesis of parametric plant models from real-world image data.

Main Methods:

• Utilizing an efficient stereophotogrammetry approach to automatically estimate depth maps and generate 3D dense point clouds from multi-view images.
• Employing the 3D point cloud as a soft constraint to fit a parametric plant representation, simulating growth.
• Developing a novel procedural modeling technique for plant synthesis.

Main Results:

• Successfully generated realistic 3D plant models with complex branching and self-occlusions.
• Demonstrated the framework's capability to reconstruct ground-covering plants (bushes, shrubs), an area often overlooked.
• Synthesized parametric plant models directly from photographic data and 3D point clouds.

Conclusions:

• The proposed automated procedural modeling technique offers a robust and efficient solution for realistic plant reconstruction.
• The method significantly reduces reliance on manual user input, enhancing accessibility.
• This approach advances the state-of-the-art in 3D plant modeling, particularly for challenging plant types.