Download now Free registration required
In the paper, the authors present a novel approach to modeling plants from images by detecting apex features. First, an effective algorithm is proposed to extract apex features in volumetric data recovered from the images. It provides position and pose information for assigning 3D generic leaves. Then, the 3D leaf shapes are determined by an optimization based on the volume. Finally, branches are modeled by using a particle flow approach. The proposed method is simply with limited manual intervention and has the obvious benefit of knowing a leaf by its visible apex part.
- Format: PDF
- Size: 468.2 KB