散乱点云的补偿滤波

Scatter Point Cloud Compensation Filtering

针对散乱点云的滤波问题,提出了一种体积保持的滤波方法.由于采用双边滤波器对散乱点云模型多次迭代滤波时,在高曲率区域会产生较大的形状变形,因此提出一种新的补偿量滤波算子,对滤波移动量进行平滑处理,获得形状补偿量,并依据形状补偿量将降噪点位置向原始位置方向调整,从而减少了滤波产生的形状变形.实验表明,该方法具有减小体积收缩的能力,对于噪声点云模型滤波,该方法简单、有效,与传统的双边滤波相比,能够减少滤波中的变形,获得具有更多细节特征的点模型.

For fairing complex noisy point cloud model,a method preserving geometric features and volumes is proposed.When a complex scatter point model is smoothed iteratively by bilateral filter,more distortions appear in high curvature areas,and a new compensation filter is considered to smooth displacements of bilateral filtering to achieve compensation distances.The final results is obtained by adjusting smoothed points back to its previous position with the compensation distance.Compared with bilateral filtering,the new method is able to preserve more fine features and decrease shape shrinkage obviously in a series of numeric experiments.