针对光顺带噪声的三角网格模型并使光顺的结果逼近模型的原始设计意图,提出了一种基于局部曲面逼近的网格光顺算法.首先获取顶点的二阶邻域内的顶点信息,利用邻域顶点的多元L1中值获得邻域确定的局部曲面的2次逼近的点集;然后将顶点沿着其法矢方向向该点集上投影,获得顶点在点集上的投影点;最后将顶点移动到投影点的坐标处,得到光顺后的三角网格.该方法在光顺的同时有效地逼近了网格的原始特征.实验结果表明了算法的有效性和鲁棒性.
To fair triangular meshes with noise and to approximate the original meshes of the 3D model designed,a smoothing algorithm based local surface approximation is presented.Specific local information of the vertex was acquired and multivariate L1 medians of the neighbor vertexes were used to determine the point set of local second order approximation of surface.The vertex was then projected to the point set along the direction of its normal.The smoothing was finally completed by moving vertex to the position of the projective point.Experiments demonstrates that the algorithm is efficient and robust.