To improve the reconstructive precision and efficiency of massive scattered point cloud,this paper put forward some improvements which aimed at the flaws existed in practical use.The flaws were the phenomenon of data blank and partial details of rebuilding surface which were not easy to grasp.The way to get new sampling points was by analyzing the abnormal points of sampling points in detail,reducing noises according to the results and adopting bicubic spline function to fit curved surfaces and patch point cloud hole.This paper proposed the least square method to calculate accurately and adjust normal vector of point cloud data.These improved the problem of overall shifting.Experiments solve the problem of quality in the reconstruction of traditional algorithm and make the external details more distinct.The result of experiments shows the great adaptability and high reconstructive efficiency and precision of this method.%为了提高大规模散乱点云的重建精度和效率,针对泊松算法在实际工程应用过程中产生的数据空白现象以及不能很好地捕捉重建表面局部细节的缺陷提出了改进.通过对采样点中的异常点进行详细分析,根据分析结果进行相应的降噪后处理,利用双三次样条插值方程拟合曲面,能够很好地修复孔洞,解决点云模型全局偏移的问题,形成新的采样点,采用最小二乘法精确计算并调整了点云数据法向量;实验解决了传统算法重建的面片质量问题,使重建出的表面细节更加显著.实验结果表明,该方法具有良好的适用性,具有较高的重建效率和精度.
展开▼
