nbul 2004 
fon (PA) / 
2002. A 
uation of 
mposium, 
Springer, 
utomated 
gery.. In: 
Sensing, 
es using 
ammmet, ry 
. ISPRS 
4. Graph 
urnal of 
—165. 
pairwise 
entation. 
Antipolis 
2004. 
pdate of 
try and 
eodaten. 
rmation 
252003. 
DA In: 
nimetric 
ation of 
(eds.), 
1, Hong 
otes in 
uctures. 
fachine 
d road 
trv and 
sensing 
mmetry 
en aus 
nission. 
on and 
ives of 
t B3/2, 
AUTOMATED ROAD EXTRACTION AND UPDATING USING THE ATOMI 
SYSTEM - PERFORMANCE COMPARISON BETWEEN AERIAL FILM, ADSA0, 
IKONOS AND QUICKBIRD ORTHOIMAGERY 
E. Baltsavias?, L. O'Sullivan", C. Zhang“ 
Institute of Geodesy and Photogrammetry, Swiss Federal Institute of Technology (ETH) Zürich, ETH-Hóngegerberg, CH-8093 
Zürich, Switzerland - (chunsun, manos)@ geod. baug.ethz.ch 
® Swiss Federal Office of Topography, Seftigenstr. 264, CH-3084 Wabern, Switzerland - Liam.OSullivan(swisstopo.ch 
Commission III, Working Group III/4 
KEY WORDS: 3D road reconstruction, automation, updating, mapping, performance evaluation, orthoimage, IKONOS, Quickbird, 
digital airborne sensor, high-resolution satellites 
ABSTRACT: 
In the recent years, the automated extraction of roads from digital images has drawn considerable attention due to the need for the 
efficient acquisition and updating of road data for geodatabases. The development of new digital aerial sensors and high-resolution 
satellite sensors signifies a revolutionary change in image acquisition and the possibility of fully digital processing from image 
acquisition to the generation of value-added products for various applications. At ETH Zurich in cooperation with and funded by the 
Swiss Federal Office of Topography (swisstopo), we have developed an operational system for the automated extraction of 3D road 
networks from imagery that integrates the processing of colour image data and existing digital spatial databases. The system focuses 
on rural areas, can use stereo or orthoimages and can determine 3D road axes, and possibly other attributes like width if the roads 
have a minimum width of ca. 3 pixels. Colour is of advantage but not a must, while a DTM or DSM is required. If no road database 
exists, it can be generated from scratch, using manual measurement of characteristic road seed points. The system has been 
extensively tested, mainly by swiss/opo, on areas with diverse terrain relief and landcover types using different resolution stereo and 
orthoimages with good results. Recently, tests have been performed using ADS40, IKONOS and Quickbird data. This paper reports 
on the performance comparison of the ATOMI system using different sensor data in two varying test sites. The test results were 
qualitatively and quantitatively analysed using accurate reference data. Visual analysis and quantitative measures of accuracy, 
correctness and completeness are presented, with typical completeness and correctness values of over 9096 and planimetric accuracy 
of 0.4 m to | m. The advantages and disadvantages using different sensor data for road network updating are also discussed. 
1. INTRODUCTION needed by most applications (see — http://www.ec- 
gis.org/etemii/reports /chapter1.pdf). 
In modern map production, a shift has taken place from maps 
stored in analogue form on paper or film to a digital database To cope with higher product demands, increase the productivity 
containing topographic information. A digital topographic and cut cost and time requirements, automation tools in the 
database is an essential part of a GIS. Recently, National production should be employed. As aerial images are a major 
Mapping Agencies (NMAs), especially in Europe, wish to source of primary data, it is obvious that automated aerial 
generate digital landscape/topographic models that conform to image analysis can lead to significant benefits. In addition, the 
reality and do not include map generation effects. In addition, development of new digital aerial sensors and high-resolution 
various existing and emerging applications require up-to-date, satellite (HRS) sensors signifies a revolutionary change in 
accurate and sufficiently attributed digital data, especially of ^ image acquisition and the possibility of fully digital processing 
roads and buildings, including car navigation, tourism, traffic from image acquisition to the generation of value-added 
and fleet management and monitoring, intelligent transportation products for various applications. At ETH Zurich, in 
systems, internet-based map services, location-based services, cooperation with the Swiss Federal Office of Topography 
etc. In 2002, two major European map providers and five car (swisstopo), we have developed a practical system for the 
manufacturers started the project NextMAP to identify and — automatic extraction of 3D road networks from imagery that 
evaluate the road database requirements for in-vehicle ITS integrates processing of colour images and existing digital 
(Intelligent Transportation Systems) and services applications, spatial databases, within the project ATOMI. Some reports on 
as well as the cost consequences involved for data capturing ^ the system performance can be found in Baltsavias and Zhang 
and data production techniques (2003) and Zhang (2003b). This paper reports on the 
(htip://www.ertico.com/activiti/proicets/nextmap/home.htm). performance of the ATOMI system using extensive areas with 
Also in 2002, twelve organisations from NMAs, road varying relief and landcover and images from different sensors. 
administrations and private sector key players of road data 
market submitted the HERDS (Harmonized European Road 2. BRIEF DESCRIPTION OF PROJECT ATOMI 
Data Solution) project proposal for EC funding. Furthermore, in 
the European Territorial Management Information 2.1 Aims of ATOMI 
Infrastructure project, roads are mentioned together with 
elevation and hydrography as the only objects, commonly 
agreed to be important enough to be defined as reference data, 
The aim of ATOMI is to update roads digitised from 1:25,000 
scale maps (part of the national VEC25 dataset) by fitting them 
741