摘要
激光雷达数据有其自身的特点,在三维空间中呈不规则分布的点云。这种数据形式在为信息提取带来便利的同时,也对数据表达方式提出更高的要求。不恰当的数据组织方法不仅会损失部分信息,还可能为信息提取带来某种困难。常用的数据表达方式包括规则格网、不规则三角网、剖面、体元等。本文对这些方法进行了比较分析,指出它们各自的优势和不足。此外,通过对激光雷达数据、表达方式、算法三者间相互关系的分析,探讨了数据表达方式在信息提取中所起的作用。
With remarkable development, the small-footprint multi-return LIDAR system has proven to be satisfactory for a variety of application, such as high-resolution elevation model generation, three-dimensional city environment mapping, vegetation structure estimation, etc. While the irregular distributed 3D points （points cloud） lend themselves to analyses and studies, they also necessitate an appropriate representation, especially for some particular purposes or in some specific applications. On the contrary, an improper one might lead to not only information loss but also erroneous interpretation. In this paper, some representing methods are examined, including raster grid, Triangulated Irregular Network （TIN）, profile and volumetric pixel （voxel）, while some other related ones are also mentioned （e. g. Pseudo- Grid and Octree Structure）. And, an exploration of the role, that representations play in feature extractions, is also presented, by an analysis of the relationships among LIDAR data, representations and algorithms.
出处
《遥感信息》
CSCD
2005年第6期60-64,T0001,共6页
Remote Sensing Information
