›› 2013, Vol. 42 ›› Issue (3): 0-0.

• 学术论文 •    下一篇

一种从车载激光扫描数据中提取复杂建筑物立面的方法

杨必胜1,董震2   

  1. 1. 武汉大学测绘遥感信息工程国家重点实验室
    2. 武汉大学
  • 收稿日期:2012-03-19 修回日期:2012-07-04 出版日期:2013-06-20 发布日期:2014-01-23
  • 通讯作者: 董震 E-mail:dongzhenwh@126.com
  • 基金资助:
    国家973计划课题;国家自然科学基金面上项目

Extracting complex Building Facades from Mobile Laser Scanning Data

  • Received:2012-03-19 Revised:2012-07-04 Online:2013-06-20 Published:2014-01-23

摘要: 本文提出了一种从车载激光扫描数据中提取复杂建筑物立面的新方法。该方法首先利用“维数特征”方法确定每个扫描点的最佳邻域,进而计算得到每个扫描点精确的局部几何特征(法向量、主方向、维数特征);然后基于“维数特征”对扫描点进行粗分类,并设置相应的生长准则对不同类别的扫描数据分别进行分割;最后综合建筑物立面的语义知识对建筑物立面区域进行精确提取。实验结果和比较分析表明,本文的方法不但能提取建筑物平面和非平面立面,而且消除了点密度差异(变化)对建筑物立面提取结果的影响,提高了建筑物立面提取的正确率和完整性。

关键词: 点云分割, 建筑物立面提取, 自适应邻域, 局部几何特征 , 车载 LiDAR

Abstract: This paper proposes an efficient method of extracting complex building facades from mobile LiDAR data in large scale urban environment. The proposed method firstly eliminates the noise in the data. An adaptive neighborhood algorithm based on dimensionality is then adopted to calculate the local geometric features of each point, such as local normal vector and dimensionality feature. Then, a region growing segmentation method with adaptive growing rules is applied to segment the point cloud. Finally, Knowledge based feature recognition method is developed for detecting facade planar patches, which are combined to represent the whole facades of complex buildings. Experiments show that the proposed method has a promising solution for building facade extracting from mobile LiDAR point clouds.

Key words: point cloud segmentation, building facade extraction, adaptive neighborhood, local geometrical feature, mobile LiDAR

中图分类号: