边界特征保持的网格模型分级二次误差简化算法

在参考张量投票理论的基础上,结合二次误差简化算法,提出一种边界特征保持的几何网格模型分级二次误差简化算法.首先根据张量投票理论对三角网格顶点进行面点、边点、角点类型分类；然后对各边对按照二次误差简化算法进行边折叠代价计算；再将分类顶点按照设定的等级权重加入边对折叠代价中,从而保证渐进式简化过程中能够对顶点进行分级简化.实验结果表明,该算法不但能实现渐进简化,而且能按需保留模型的整体特征和细节特征.

New Quadric Metric for Simplifying Meshes to Retain the Feature Edge

Using tensor voting theory to classify the vertices in triangle meshes and connecting with the quadric error metric surface simplification algorithm,we presented a new quadric error metric algorithm for simplifying meshes to retain the feature edge of the original model.First,all vertices are classified into a plane,a sharp edge,a corner,according to the tensor voting theory.Then every edge is classified into several types,each type with a level weight.And the level weight is added into the cost of contracting pair in the quadric error metric surface simplification algorithm.Finally,iteratively removing the edge pair,a feature edge retained progressive meshes are generated.The empirical results show that the proposed algorithm is able to retain the overall characteristic features and small details of the original model.