Academic journal article Cartography and Geographic Information Science

Interpretation and Generalization of 3D Landscapes from LiDAR Data

Academic journal article Cartography and Geographic Information Science

Interpretation and Generalization of 3D Landscapes from LiDAR Data

Article excerpt

Introduction

Traditionally, 2D cartography has employed a workflow of different processes (such as aggregation, simplification, enhancement, and displacement) to create the desired cartographic product. However, the digital era we are in offers far greater flexibility in terms of both the available data and potential applications. For example, the availability of dense 3D data offers new possibilities concerning the communication of topographic information, particularly in the context of dynamic and adaptive maps created for instantaneous use. The communication process thus needs to be viewed as a set of individually configurable processes that adapt to the underlying data. Achieving this requires different extraction and interpretation operations (depending on the data), and appropriate generalization operations that conform to the general objective of the communication process.

The research reported here instantiates modules of this communication process applicable for roads extracted from raw 3D LiDAR data, and their cartographic generalization and exaggeration within a 3D terrain model. The communication pipeline for this scenario is demonstrated in Figure 1. The first step is the extraction of spatial information from the point cloud. This information can be in the form of terrain only or a combination of terrain and other topographic features, such as roads. Additional information from external sources, e.g., features from a GIS database, can also be input into this process. The next step is the generalization of the information; one can generalize the terrain, the other topographic features, or both. The choice of generalization processes depends in part on the topography.

[FIGURE 1 OMITTED]

For example, if roads cross relatively flat areas, one can generalize them using such traditional cartographic signatures as overlay of thick lines. For rolling terrain or mountainous areas, traditional cartographic signatures will create a disturbing effect, as they will partly cover the original road and partly the surrounding topography. Therefore, in these areas the generalization of both terrain and road elements becomes significant. The final output is the visualization of the information as a communication tool.

The following sections provide an overview of related work, discussions of the interpretation and generalization of 3D topography, a description of the experimental results from our work, and conclusions.

Related Work

Interpretation of Topographic Features from LiDAR Data

The conversion of raw laser data into a generalized terrain model with explicit road networks requires the detection and the extraction of both entities from the raw point cloud. Extraction of the terrain from the raw laser data has been a subject of intensive research in recent years, leading to a variety of techniques. These can be classified into the following categories: morphological algorithms (e.g., Vosselman 2000), densification based models (Axelsson 1999), analysis of surface discontinuities (Brovelli et al. 2004), models based on robust surface fitting (Briese and Pfeifer 2001), and models based on surface segmentation (Sithole and Vosselman 2003; Nardinocchi et al. 2003). Recently, models that combine individual techniques have been proposed as a means of compensating for the shortcomings of the individual approaches (see Abo-Akel et al. 2004, Tovari and Pfeifer 2005). A comprehensive review of the principal filtering approaches and an analysis of their performance is presented in Sithole and Vosselman (2004). We utilize the filtering model proposed in Abo-Akel et al. (2004) because of its good performance over regions with complex topographic features (see Abo-Akel et al. 2007).

The detection of road networks has been the subject of research involving imagery from air- and spaceborne platforms (see e.g., Wiedemann and Hinz 1999; Hinz and Baumgartner 2003; and most notably the collection of manuscripts in the special issue on road extraction in Photogrammetric Engineering & Remote Sensing 70(12)). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.