保持地形特征的网格模型简化算法

Mesh Simplification Algorithm for Terrain Model Based on Feature Degree

针对目前地形简化算法在低分辨率的状态下忽略地形本身的重要起伏特征,如山峰、河谷等典型区域,导致视觉上的失真的问题,在Garland算法的基础上,结合地形学中山谷线和山脊线定义提出顶点特征度的概念,并将其作为权值嵌入到误差测度中,使简化后的地形模型能够保留更多重要特征信息.与Garland算法的对比实验表明,文中算法在保持原始算法高效性能的基础上,提高了地形模型精度,误差更小,并且保留了更多的重要特征信息.

Most of the existing simplification algorithms for terrain model might lost some important shape features, such as valley and channel, especially at the low levels of the simplified model, and this will lead to the visual distortion. Based on Garland＇s greedy insert algorithm, a feature degree of a vertex is introduced. It can change the order of candidates of error metric through calculating the feature degree of vertexes. So the more important vertex of the original model can be preserved. The experimental result shows that the proposed algorithm can maintain more terrain features at the same decimation rate as greedy insert algorithm, while reducing visual distortion effectively.