基于高斯曲率的航空模锻件点云降噪方法研究

Research on Point Cloud Noise Reduction Method for Aerospace Die Forgings Based on Gaussian Curvature

为了让大型航空模锻结构件全型面海量三维点云数据在高效降噪过程中避免过度光顺而丢失边缘特征点,论文提出了基于高斯曲率的分区域双边点云降噪方法。首先将散乱点云栅格化进行降采样,剔除了第一类噪声点;然后基于高斯曲率对点云分区域,同时添加欧式距离权重来保持点云空间特征,利用改进的双边滤波方法去除第二类噪声。最后在开源PCL1.10.1平台上,验证了论文点云降噪方法的有效性,试验表明论文点云滤波方法在点云降噪的同时,在避免曲面过度光顺和保持边缘特征方面有一定的效果。

In order to avoid excessive smoothing and loss of edge feature points during the high-efficiency noise reduction pro-cess for the massive 3D point cloud data of the full-surface large-scale die forging structure,this paper proposes a regional bilateral point cloud noise reduction method based on Gaussian curvature.First,the scattered point cloud is rasterized for down-sampling to eliminate the first type of noise points.Then the point cloud is divided into regions based on Gaussian curvature,while the Euclide-an distance weight is added to maintain the spatial characteristics of the point cloud,and this improved bilateral filtering way is for removing the second type of noise.Finally,on the open source PCL1.10.1 platform,the effectiveness of the point cloud noise reduc-tion method in this paper is verified.The trial shows that this point cloud filtering way can reduce the noise of the point cloud while avoiding excessive smoothing of the surface and maintaining edge features.