一个自适应的植物枝干重建方法

Adaptive Approach for Reconstructing Botanical Branch Structures

植物的真实感建模在真实场景绘制、3D游戏、农林业研究等方面都发挥着重要的作用。由于植物形态结构的复杂性和多样性,大部分植物的真实感模型的精确重建方法的计算复杂度较高。提出一种从植物枝干的三维扫描数据中重建植物枝干结构的方法,该方法可以依据植物枝干结构自适应的调整计算精度,从而在保证精度的同时大大降低计算复杂度。首先将植物枝干的三维扫描数据分割成长度相等的几个段,并分别对每段进行初始聚类;然后自适应的对包含分叉结构的聚类进行迭代细分以形成更符合植物枝干结构的聚类;最后,提取枝干骨架并形成最终的表面网格模型。实验结果表明,该方法具有较高的效率,并且能够重建出比较接近植物真实枝干结构的模型。

Realistic modeling for the plant plays an important role in real scene rendering, 3D games and research of agriculture and forestry. Due to the complexity and diversity of plant shapes, most of the precise reconstruction approaches of realistic models for plants have high computation complexity. An approach for reconstructing botanical branch structures from 3D scanned points was introduced. The proposed approach could adaptively tune the computation precision according to the complexity of the branch structures, thus it could preserve the precision while greatly reducing the computation complexity. The 3D branch points were firstly segmented into several sections and each section was then clustered. Secondly, the clusters including ramifications of the plant were adaptively subdivided. Finally, a skeleton structure was extracted and then the branch surface was computed. The experiment results show that the proposed approach is efficient as well as capable to reconstruct the branches that are more consistent with the real plant.