一种基于曲线拟合与采样的点云数据压缩方法

在反求工程中,从样件表面采集得到的通常是数量非常庞大的点云数据,严重影响了曲面重建算法的效率;另外,基于四边域曲面的重建算法通常要求型值点数据呈拓扑矩形排列,而采样得到的散乱点云通常不满足这一拓扑要求。文中提出了一种新的数据压缩技术路线,用于把海量散乱点云数据合理地压缩到四边域曲面重建所要求的数据量和拓扑形式。该方法的核心是B样条曲线的拟合和采样。为使采样点更好地反映原始模型的外形特征,给出了一种基于曲率的自由曲线自适应采样算法。应用实例表明本文提出的方法达到了预期的效果。

A Point Cloud Data Compression Approach Based on Curve Fitting and Sampling

In reverse engineering, the data acquired from the surface of a sample product always results in a huge quantity of point cloud data, which greatly influences the efficiency of free-form surface reconstruction. Moreover, the scattered point cloud thus acquired does not meet the requirements of a topologically rectangular permutation. A new data compression approach whose key technique is B-spline curve fitting and sampling is presented. The approach compresses huge quantities of scattered point data and produces a reasonably smaller-scale and topologically rectangular point cloud. In order for the sampled points to fully represent the external features of their prototype, an adaptive free-form curve sampling algorithm based on curvature was presented. Application instances show that the approach achieves expected results.