Abstract: A new adaptive subdivision scheme based on vertex’s importance is presented. The vertex’s average plane is replaced by the plane that is formed by three endpoints of three longer edges connected to the vertex in the triangle mesh, and then the distance between the vertex and its average plane is regarded as the criterion for the vertex’s importance so that it can be decided whether the vertex needs to be subdivided. Since the original triangular faces are of high density and shape similarity, the error derived from the vertex-to-plane distance as subdivision measure can be limited to a volume element. Compared with the algorithm of amending the vertex’s normal vector repeatedly, this method decreases the amount of computation greatly. Experimentally, the approach proposed can both accelerate the processing speed effectively and reduce the amount of faces greatly.

Key words: triangle mesh, average plane, vertex’s importance, loop scheme, adaptive subdivision scheme

摘要： 提出了一种新的自适应细分算法，在顶点的1-邻域内，用与顶点相连较长三条边的端点构成的平面去替代其平均平面，将顶点到其平均平面的距离作为判断顶点重要度的标准，对三角网格进行自适应细分。由于原始三角面片的高密度和形状相似性，以点面距离为细分尺度所产生的误差，可被限制在一个体元之内，与反复修正顶点法矢算法相比，该算法大大减少了计算量。实验结果表明，所提方法在三角网格细分过程中，简化了数据模型，提高了处理速度。

关键词: 三角网格, 平均平面, 顶点重要度, Loop细分, 自适应细分