大型储罐的三维点云精简方法研究

针对大型储罐三维点云数据散乱、冗余点多等影响计算机显示及容积计算的问题，改进了一种储罐三维点云精简算法。该方法先利用均匀网格法，将待处理的三维点云数据分割成若干小栅格；然后根据随机抽样一致(random sample consensus，RANSAC)算法对每个栅格中的点云数据建立球模型，以保留特征点并滤除冗余数据点，达到精简点云的目的。将该方法与传统的均匀网格法和非均匀网格法进行对比，实验结果表明该方法在保证较高精简率的情况下可以更好地保留储罐点云数据特征。

3D point cloud reduction algorithm for a large storage tank

There are a large number of scattered redundant points in 3D point cloud data, which affects the computer display and volume calculations for large storage tanks. This paper proposes an improved 3D point cloud simplification algorithm for storage tanks. Firstly, the uniform grid method is used to divide the 3D point cloud data into several small grids. Then, a spherical model is established for the point cloud data in each grid which retains the feature points and filters out the redundant data points by means of a random sample consensus (RANSAC) algorithm. Finally, the simplified point cloud is obtained. Comparison of the proposed method with the traditional uniform grid method and the non-uniform grid method demonstrates that the simplified point cloud retains more characteristics of the experimental storage tank point cloud data while ensuring high simplification rate.