摘要
为了提高三角网格模型的质量,满足模型后续处理的要求,提出了一种基于邻域相似性的网格光顺算法。类比图像像素的灰度值,首先构造顶点的双边滤波微分算子,作为其几何灰度值;然后计算顶点邻域之间的相似性,作为顶点几何灰度值的权值,并对顶点的邻域顶点的几何灰度值进行加权平均,得到该顶点的最终几何灰度值;最后将顶点沿着其法矢量方向移动几何灰度值大小的距离,得到光顺后的三角网格模型。实验证明,该算法在光顺模型的同时有效地保持了网格的几何特征。
In order to enhance the quality of triangular mesh model and meet the requiremented of follow-up treatment, this paper presented a mesh smoothing algorithm based on neighborhood similarity. Firstly, contrasting to the gray values of image pixels, bilateral filtering differential operator was constructed as geometric gray value of the vertex. Then neighborhood similarity was calculated between vertexes and the result was used as weight of geometric gray value for each vertex. The final geometric gray value of a vertex was the average weight of its neighbor vertexes' geometric gray values. Finally, the vertex moved geometric gray value size of distance along its normal direction, and then we would get the smoothed model. Experiments demonstrate that this algorithm can acquire smoothing models and maintains their geometrical features effectively at the same time.
出处
《计算机科学》
CSCD
北大核心
2010年第3期289-291,共3页
Computer Science
基金
国家863高技术研究发展计划(2007AA04Z137)资助
关键词
三角网格
几何灰度值
双边滤波
几何特征
Triangular mesh, Geometric gray value, Bilateral filter, Geometric feature