白噪声干扰下三维点云数据重建方法仿真
|
摘要
为解决当前方法进行白噪声干扰下三维点云数据重建过程中,获取的三维点云数据噪声造成的边缘数据缺失,存在三维重建精度较低的问题,提出一种基于曲率图的白噪声干扰下三维点云数据重建方法。所提方法先采用多帧数据融合理论获取白噪声干扰下的三维点云数据并对三维点云数据进行初步平滑处理,运用深度数据双边滤波理论滤除三维点云数据离群点噪声和内部高频噪声,依据去噪处理后的高质量三维点云数据各点的邻域特性度量各点及其近邻邻域点的曲率值,构造相应点的曲率图,通过在三维点云数据多比例空间下曲率图的特征保持分析,提取三维点云数据特征的特征点集并进行三维点云数据特征点对匹配,由匹配的三维点云数据特征点对进行坐标变换求解,在此基础上采用迭代最近点理论进行三维点云数据配准,以此为依据完成白噪声干扰下三维点云数据重建。实验结果表明,所提方法能够有效去除白噪声干扰影响,保留边缘数据的完整性,且三维点云数据重建精度较高。
This article proposes a reconstruction method of point cloud data with three dimensions(3D) under white noise interference based on curvature plot.Firstly,the author used data fusion theory with multi-frame to acquire the data and carried out initial smoothing processing for the data,and then applied theory of bilateral filter of depth data to filter outlier noise and internal high-frequency noise of the data.Moreover,the research measured curvature value of each point and its neighborhood point according to neighborhood character of the point cloud data with high quality after de-noising processing and structured curvature figure of relevant point.Through analyzing curvature figure character in multi-scale space of cloud data,the research extracted feature point set of the data feature and matched dot pair of the feature,and carried out solution of transformation of coordinate by the matched dot pair.On that basis,the research used theory of iterative closest point to carry out rectification for the point cloud data.Finally,according to that,the reconstruction was completed.Experimental results show that the method can eliminate influence of white noise interference effectively,retain integrity of marginal data and has high reconstruction precision.
This article proposes a reconstruction method of point cloud data with three dimensions(3D) under white noise interference based on curvature plot.Firstly,the author used data fusion theory with multi-frame to acquire the data and carried out initial smoothing processing for the data,and then applied theory of bilateral filter of depth data to filter outlier noise and internal high-frequency noise of the data.Moreover,the research measured curvature value of each point and its neighborhood point according to neighborhood character of the point cloud data with high quality after de-noising processing and structured curvature figure of relevant point.Through analyzing curvature figure character in multi-scale space of cloud data,the research extracted feature point set of the data feature and matched dot pair of the feature,and carried out solution of transformation of coordinate by the matched dot pair.On that basis,the research used theory of iterative closest point to carry out rectification for the point cloud data.Finally,according to that,the reconstruction was completed.Experimental results show that the method can eliminate influence of white noise interference effectively,retain integrity of marginal data and has high reconstruction precision.
引文
[1]蔡剑.基于数学形态学的图像椒盐噪声去除方法[J].电子设计工程2016,13(24):182-184.
[2]王向玉,等.基于地面激光雷达点云数据的单木三维重建[J].遥感技术与应用,2015,30(3):455-460.
[3]何东健,等.Kineet获取植物三维点云数据的去噪方法[J].农业机械学报,2016,47(1):331-336.
[4]韩凯,等.基于深度扫描仪的高辨识度三维人体模型重建方法[J].图学学报,2015,36(4):503-510.
[5]林祥国,等.一种机载LiDAR点云电力线三维重建方法[J].测绘科学,2016,41(1):109-114.
[6]贾松敏,等.基于变分模型的单目视觉三维重建方法[J].光学学报,2014,34(4):162-168.
[7]翟志强,等.基于Rank变换的农田场景三维重建方法[J].农业工程学报,2015,31(20):157-164.
[8]吴福培,等.基于单目视觉系统的三维重建方法[J].测试技术学报,2016,30(5):369-374.
[9]陈大平,王栋,刘衷瑞.一种基于双目视觉的三维重建方法[J].现代测绘,2016,39(1):14-16.
[10]王铭军,朱文耀.古代建筑三维虚拟重建方法研究与仿真[J].计算机仿真,2014,31(3):437-440.
[2]王向玉,等.基于地面激光雷达点云数据的单木三维重建[J].遥感技术与应用,2015,30(3):455-460.
[3]何东健,等.Kineet获取植物三维点云数据的去噪方法[J].农业机械学报,2016,47(1):331-336.
[4]韩凯,等.基于深度扫描仪的高辨识度三维人体模型重建方法[J].图学学报,2015,36(4):503-510.
[5]林祥国,等.一种机载LiDAR点云电力线三维重建方法[J].测绘科学,2016,41(1):109-114.
[6]贾松敏,等.基于变分模型的单目视觉三维重建方法[J].光学学报,2014,34(4):162-168.
[7]翟志强,等.基于Rank变换的农田场景三维重建方法[J].农业工程学报,2015,31(20):157-164.
[8]吴福培,等.基于单目视觉系统的三维重建方法[J].测试技术学报,2016,30(5):369-374.
[9]陈大平,王栋,刘衷瑞.一种基于双目视觉的三维重建方法[J].现代测绘,2016,39(1):14-16.
[10]王铭军,朱文耀.古代建筑三维虚拟重建方法研究与仿真[J].计算机仿真,2014,31(3):437-440.