1253 
DIGITAL ELEVATION MODEL DATABASE W42 - A SCALABLE SYSTEM FOR 
SPATIAL DATA 
M. Habermeyer, U. Marschalk, A. Roth 
German Aerospace Center (DLR), German Remote Sensing Data Center (DFD) 
82234 Wessling, Germany, (martin.habermeyer,ursula.marschalk,achim.roth)@dlr.de 
Commission I, SS-1 
KEY WORDS: DEM/DTM, Mapping, Mosaic, Software, System, Spatial Databases 
ABSTRACT: 
Digital Elevation Models (DEMs) provide elementary information for various application fields. From a global point of view, there 
are many different sources to furnish height information of the Earth's surface. These sources can be heterogeneous in coverage, as 
well as in resolution, information content and quality. A database aiming at the provision of world-wide DEMs has to consider these 
differences in the design of the system with respect to the structure and the algorithms.The DEM database W42 is a scalable spatial 
database system capable of holding, extracting, mosaicing, and fusing spatial data. Design aspects for this task can be specified as 
holding spatial data in a unique data structure and providing unique access functions to the data. These are subject of this paper.The 
database's structural design is presented in a Four-Layer-Concept. Each layer provides a clearly specified interface to 
subsequent/preceding layers, as well as each layer is scalable in nature to meet future requirements. The scalability of the system 
enables the integration of the database into operational processors: at the moment it is serving as part of the geocoding-processor of 
the TerraSAR-X mission and as part of the generic processing chain for airborne hyperspectral data. It further will be utilized in 
satellite missions as an element of the Payload Ground Segment in the future. The extensions made for TanDEM-X are explained in 
this context. 
1. INTRODUCTION 
A design of a scalable system for holding spatial data in general 
and Digital Elevation Models (DEMs) in specific, has to 
account for some basic conditions resulting from the different 
sources DEMs are produced from. The sources can be 
heterogeneous in coverage, as well as in resolution, information 
content and quality. A database aiming at the provision of 
world-wide DEMs has to consider these differences in the 
design with respect to the structure and the algorithms. 
The W42-DEM-Database is a stand-alone database system 
capable of holding spatial data in a unique system architecture. 
It provides functions allowing import and export of spatial data 
in raster and vector-like structures. The purpose of this 
document is the clarification of the database design and the 
description of the provided functionality. 
This work extends the papers and reports previously published 
about the database see e.g. (Roth et al. 2002), (DLR, 2006). The 
database was part of the geocoding system for radar missions: 
SRTM/X-SAR (Roth et al., 1999) and TerraS AR-X (Huber et 
al., 2006). It is also used in an operational processing chain for 
optical airborne data (Habermeyer et al., 2003). The focus in 
this paper is on design aspects with regard to the TanDEM-X 
mission, for which the database has been adopted to be a part of 
the Mosaicking and Calibration Processor (Wessel et al., 2008) 
and to support accompanying studies. 
The document is structured as follows: 
Section 2 explains the System Design in a Four-Layer-Concept, 
comprising Physical Layer, File System Layer, Database 
Access Layer and File Access Layer. 
Three different kinds of data (Data Tiles, Image Chips, and 
Tracks) can be stored in the database system at the moment. 
These are specified in section 3. 
The data types are abstracted to data formats (section 4). 
For each data type (Data Tile, Image Chip, track) the 
organisation of the file sytem is explained (section 5). 
The interaction of the database access routines with the file 
system is subject of discussion in section 6: the needed routines 
for import to, export from, update of, inventory of and 
maintenance of the database are specified and their impact on 
database components are discussed. 
The utilities for the manipulation of the database content are 
based on programs providing functionalities to handle raster- 
and vector data. These are explained in section 7, without a 
consideration of an organisational structure, only working on a 
file level without spatial semantics. 
The functions of the File Access Layer cover file format 
conversion, file manipulation, information about the data 
content, visualization of files, fusion of information provided by 
different layers and different formats, mosaicking them together, 
and resampling routines to change their resolution. 
2. SYSTEM DESIGN 
The database's structure is organized in four layers (fig. 1). The 
system is based on the "Physical Layer" represented by media 
mounted on workstations or PCs. The second layer, which is 
named "File System Layer" spans the directory tree over the