【作者】 张熠斌；

【导师】 隋立春；

【作者基本信息】 长安大学 ， 摄影测量与遥感， 2010， 硕士

【摘要】 机载LiDAR作为一种新兴的主动式遥感技术,在地表三维空间信息的实时获取方面取得了突破性的进展,引起了测绘、林业等相关行业的浓厚兴趣。近几年,随着相关技术的发展以及社会需求的不断扩大,机载LiDAR技术的发展日新月异,为获取高时空分辨率地球空间信息提供了一种全新的技术手段,使数据的获取和处理朝智能化和自动化的方向发展,代表了对地观测领域一个新的发展方向。相对硬件技术的飞速发展,LiDAR点云数据后处理算法(如自动滤波和分类)目前仍处于研究和发展阶段,有待进一步深入研究。本文系统地探讨了机载LiDAR系统及其数据数据后处理理论与技术,主要创新点如下：(1)在数据管理方面。采用建立规则格网索引的方法,实现了大规模点云数据的快速访问；引进内存映射文件技术,实现了大规模点云数据的快速读写；基于"Mesh"面几何体模拟密集的点云数据,实现了大规模点云数据的快速绘制；使用分块绘制和稀疏显示的方法,实现了大规模点云数据的实时刷新。(2)在滤波算法方面。改进了现有数学形态学和渐进三角网滤波算法。实验证明,改进后的算法,算法稳定,执行效率高,具有良好的可靠性与实用性。(3)在点云数据后处理技术方面。基于MicroStation设计与开发了LiDAR点云数据后处理软件TopLiDAR,引进了经改进后的数学形态学滤波算法和渐进三角网滤波算法,并设计了点、线、面和连续动态断面编辑等手工分类工具,提高了LiDAR点云数据滤波和分类的效率。研究LiDAR点云数据滤波分类算法,开发相应的数据后处理软件,是提高其数据精度的有效途径之一,也是目前其数据后处理领域研究的热点之一。本文的研究对LiDAR点云数据的后处理具有一定的实用和参考价值。更多还原

【Abstract】 As a new active remote sensing technology, airborne LiDAR has produced the gross breakthrough in the aspect of three-dimensional earth spatial information acquisition, and has caused great interests in the surveying and mapping, forestry and other related industries. In recent years, with the development of related technologies and the expansion of the social needs, airborne LiDAR which represents a new development direction in the field of earth observation, offers a kind of completely new technical means for the high temporal-spatial resolution earth spatial information acquisition, make data acquisition, processing automation and intellectualization to be possible. Compared with the rapid development of hardware technology, the post-processing algorithms (such as automatic filtering and classification) of LiDAR point cloud data are still in the stage of being researched, and need to be further studied.This paper systematically discusses the airborne LiDAR system and the theories and techniques of data post-processing. The major innovations are listed below:(1) Data management. Large-scale point cloud data can be fast accessed by establishing regular grid index,and can also be read and written rapidly on the basis of memory-mapped file technology, drew quickly based on the "mesh" geometry which used to simulate dense point cloud data,and can be real-time refreshed with the method of block mapping and sparse displaying.(2) Filtering algorithm. Mathematical morphological and progressive TIN filtering algorithms have been improved in this paper. Experiments proves that the modified algorithms are stable and efficient with good reliability and practicability.(3) Point cloud data post-processing technology. The TopLiDAR that used to process LiDAR point cloud data,which is designed and developed based on MicroStation. It has mathematical morphological and progressive TIN filtering algorithms which have been improved, and has manual classification tools included point,line, surface, continuous dynamic profile editing, etc., which improve the efficiency of LiDAR point cloud data filtering and classification.Studying filtering and classification algorithms and developing the appropriate post-processing software of the LiDAR point cloud data are one of the effective ways to improve its data accuracy, also one of the researched focuses in the field of its data post-processing. The research in this paper may be helpful for it.更多还原