Processing of Laser Scanner Data and Extraction of Structure Lines Using Methods of the Image Processing

摘要

Airborne laser scanner data represents a new and independent technology for the highly automated generation of digital elevation models (DEM) and surface models. Airborne laser scanner raw data are unstructured and irregular points which need intensive post processing. The first step of a post processing is the separation of valuable points which lie on the ground surface (ground points) from outliners which are located above it (non-ground points). Another task is the extraction of structure lines from the separated ground points. Structure lines play an important role for the construction of precise DEM. The methods, based on mathematical morphology and robust parameter estimation, are explored for the detection of ground points and for the separation of ground points from non-ground points. The presented and developed method for the extraction of structure lines is based on the digital image processing methods and on the differential geometry of curves. This operator was initially developed for image processing and is now used in a new context. The developed and implemented methodology is open for further extension and compatibility. It is hence possible to use it in other fields of application. The presented work shows that image processing methods yield good results for the extraction of structure lines without the need of specific knowledge of objects.%激光扫描数据提供了一种新的手段用于获取高精度的数字地形表面模型. 原始的航空激光扫描数据表达的是一些非规则分布的"点云", 这些非规则分布的点需要进行有效的事后处理. 这种事后处理有2个目的:一是将那些分布在地表面上的点(即地面点)与分布在非地表面上的点(譬如树木、房屋或汽车上的点, 即非地面点)进行有效的分离;二是从分离后的地面点中提取结构线, 用于建立高精度的数字地面模型. 作者发展了一系列的基于数字形态学理论和稳健参数估计理论的方法用于分离和探测地面点. 这里所介绍和开发的提取结构线的算法建立在数字图像处理和表面曲率理论的基础上. 这些算法同样可以扩展地用于其他领域. 所介绍的基于数字图像处理理论处理原始的航空激光扫瞄数据和提取结构线的方法取得了很好的结果. 这一结论可以在本文中通过一系列的插图得到有力的证明.