一种基于带噪声的三角网格模型的光顺算法

为了去除三角网格模型中的噪声,提出了一种基于均值漂移的特征保持的网格光顺算法。该方法在对模型中的三角形的法向量进行滤波的基础上鲁棒地计算了顶点的法向量,利用均值漂移方法自适应地聚类出顶点的邻域。结合顶点间几何特征的相似性,将改进后的双边滤波算子应用于顶点的位置更新,从而完成模型的光顺。实验结果证明了网格光顺算法的有效性。利用这种网格算法,可以达到光顺带噪声的三角网格模型的目的,并在光顺的同时,有效地保持了模型中原有的特征。

Smoothing algorithm based on triangle mesh with noise

In order to filter the noise in triangular mesh model, this paper presented a feature preserving mesh smoothing algorithm based on mean shift. First, it computed the normal of the vertex robustly according to the filtered normal of triangles. Then, it used the mean shift approach to acquire the neighborhood of the vertex adaptively. Finally, it improved the bilateral filter to incorporate geometric feature similarity of the vertex and the model was smoothed after updating the position of the vertex. The experiment results show that algorithm is effective. The use of this grid algorithm can smooth triangular mesh model with noise and keep some original characteristics of the model .