基于凸壳与有向包围盒的骨架提取方法

Skeleton Extraction Method Based on Convex Hull and OBB

为获取三维模型的几何及拓扑信息,提出一种基于凸壳与有向包围盒(OBB)的线性骨架提取方法.首先将三维网格模型进行分割生成多个子网格模型;然后对各子网格中的点集求取凸壳作为该子网格点集的近似,由凸壳顶点的形心构成原始骨架点;再用OBB进行重叠计算求出相交点集,以生成关节骨架点;最后对原始骨架点与关节骨架点进行连接,经冗余检测后形成完整骨架.实验结果表明,该方法快速、有效,提取出的骨架能保证连通性与中心性且能很好地提取关节骨架点,为蒙皮关节动画、模型形状分析等提供有效信息.

In order to obtain the geometry and topology of a 3D model, a skeleton extraction method based on the convex hull and the oriented bounding box （OBB） is presented in this paper. It firstly divides the mesh into several parts; computes the convex hull of each part＇s point set as its approximation, and uses the convex hull＇s centroid as the original skeleton point. Then it looks for the overlapping point set between two adjacent parts by using OBB, determines the centroid of the overlapping point set as the joint point. It finally connects original skeleton points with joint points and removes redundant points to get the entire skeleton. Experimental results show that this skeleton extraction method is simple, fast and effective. It guarantees the connectivity and centrality of the skeleton, which can provide useful information for skinning joint animation, 3D model shape analysis, etc.