测绘通报 ›› 2018, Vol. 0 ›› Issue (10): 27-31.doi: 10.13474/j.cnki.11-2246.2018.0309

• 学术研究 • 上一篇    下一篇

一种机载LiDAR点云缺失数据的填补方法

王丽英1, 夏玉红2, 徐艳2, 赵元丁1   

  1. 1. 辽宁工程技术大学, 辽宁 阜新 123000;
    2. 北京金景科技有限公司, 北京 100094
  • 收稿日期:2018-01-23 出版日期:2018-10-25 发布日期:2018-10-31
  • 作者简介:王丽英(1982-),女,博士,副教授,主要研究方向为激光雷达基础理论及应用研究。E-mail:wanglyinglntu@163.com
  • 基金资助:
    辽宁省自然科学基金面上项目(20170540419);国家自然科学基金(41471351)

A Method for Filling Absence Data of Airborne LiDAR Point Cloud

WANG Liying1, XIA Yuhong2, XU Yan2, ZHAO Yuanding1   

  1. 1. LiaoNing Technical University, Fuxin 123000, China;
    2. Beijing Brisight Technology Company, Beijing 100094, China
  • Received:2018-01-23 Online:2018-10-25 Published:2018-10-31

摘要: 针对机载激光雷达点云数据内部存在的缺失数据,提出了一种缺失数据辨识及填补方法。首先,利用二值数据空洞图检测数据缺失的存在;其次,利用连通区域标记算法将数据空洞图分割为多个数据空洞;然后,对大面积的数据空洞的外边界轮廓数据点的凸包依据方差和平均曲率条件进行等距扩大,获得数据空洞的有效影响区域;最后,研究不同的缺失数据类型的最优空间插值方案,并基于影响区域内的点云,利用最优空间插值方案对不同类型的缺失数据进行填补。试验基于ISPRS提供的包含了不同类型的数据空洞的点云数据测试了影响区域的有效性及不同类型的缺失数据的最优空间插值方案,从而验证了提出方法的有效性。

关键词: 缺失数据, 数据空洞, 填补, 影响区域, 空间插值

Abstract: For the absence data existed in the airborne LiDAR point cloud, a method for identification and filling the absence data is proposed. The proposed method first detects absence data depending on binary data void image. Next the data void image is segmentd into multiple data voids by means of connected-component labeling algorithm. Then influence region of each data void is obtained by isometrically expanding the outer boundary contour based on the variance and mean curvature criterion. Finally, the optimal spatial interpolation scheme is determined related to different data void types, and then it is used to fill the data voids based on the point clouds in influence region of each data void. ISPRS LiDAR datasets, which are representative of absence data of diverse types, are used to analyze the effectiveness of "influence region" and to detemine the assess optimal spatial interpolation scheme for different data void types. And thus the feasibility and effectiveness of the proposed method is verified.

Key words: absence data, data void, filling, influence region, spatial interpolation

中图分类号: