2 
L$ 
N 
to 
of 
nd 
ial 
VS. 
“FAST MAPPING” FROM HIGH RESOLUTION SATELLITE IMAGES: A 
SUSTAINABLE APPROACH TO PROVIDE MAPS FOR DEVELOPING COUNTRIES 
M. Gianinetto ?, A. Giussani?, G.M. Lechi *, M. Scaioni? 
* Politecnico di Milano, Dept. II.A.R., Remote Sensing Lab., P.zza L. da Vinci 32, 20133 Milano, Italy 
Politecnico di Milano — Polo Regionale di Lecco, Dept. IL.IA.R., via M. d'Oggiono 18/a, 23900 Lecco, Italy 
e-mail: {marco.gianinetto, alberto.giussani, giovanmaria.lechi, marco.scaioni}(@polimi.it 
KEY WORDS: Developing countries, QuickBird, SPOT, IKONOS, EROS, High Resolution, Mapping, Sustainable Development 
ABSTRACT: 
Imagery coming from high resolution sensors seems to become in the near future a tools to derive maps, comprehending large scales 
as well. At the current state-of-the-art, this dream is still infeasible: the main reason is the unavailability of stereo-pairs (barring few 
exceptions). Results of different researches on this topic has stated the usefulness of this kind of data to yield cartography in those 
countries where production based on traditional methods cannot be really afforded. This impossibility is due to the high skilfulness 
of operators involved in the process, to the expensive hardware required (stereo-plotters or DPW), to the availability of a company 
which may provide aerial photos over the interested area. 
On the contrary, the use of satellite imagery is simpler: data can be purchased via Internet, choosing between different kinds of 
sensors and resolutions, processing can be completely performed by commercial SWs, control points can be measured by GPS, 
operations are simple and can be easily standardized. Furthermore, the technology transfer process from researchers to operators is 
sustainable. 
Obviously the cartographic product that might be obtained in this way is different with respect to a numerical 3D map derived from 
aerial photogrammetry. The third dimension of object cannot be computed, due to the unavalability of stereo-pairs, and the pixel size 
does not allow to derive maps larger than scales up to 1:10,000-1:5,000. Due to difference from the traditional cartography, either in 
the production process and in the content of maps, we have termed such a product *Fast Mapping". 
1. BACKGROUND infrastructure required to collect and process them is not 
available, the use of HRSI to derive topographic maps and 
One of the topic challenge which have engaged the Earth orthophotos is an excellent opportunity at the present as well. In 
Observation community in the recent years can be summarized this context, satellite imagery can provide a rapid and high- 
by the following question: which is the practical use of high quality data source for the production of different kinds of 
resolution satellite imagery (HRSI)? In spite of the large maps, such as vector maps, orthophotos, thematic maps and the 
volume of research outputs yielded on this subject, the use of like. 
HRSI-derived products is relatively low. In this paper an approach to derive a such kind of cartography is 
According to some addresses given by several studies carried discussed, giving some addresses to set up an operational 
out at national and over-national levels (Holland er a/., 2002; process termed as "fast mapping" (Caprioli & Tarantino, 2001). 
Devriendt er al, 2003; Holland & Marshall, 2003), the spatial As illustrated in the sequel, the proposed issue is not restricted 
resolution of this kind of data allows to reach the level of detail to an operational concern. Recently, many papers have been 
to resolve individual objects in the landscape, in a similar way reported to tackle typical problems about HRSI: orientation 
than the airborne data does. As consequence, HRSI might be (Biiyiiksalih ef al., 2003; Fraser ef al., 2002a, 2002b, Fraser & 
used as source for extraction of detailed, object-related Yamakawa, 2003, Tao & Hu, 2001, 2002, Valadan Zoej & 
information and for the production of large-scale maps. The Sadeghian, 2003), orthophoto production (Jacobsen & Passini, 
pixel size of IKONOS images is corresponding to the 2003), image classification (Neubert & Meinel, 2003). On the 
information content of aerial images with scale 1:80,000, that other hand, analysis of which features can be effectively and 
of QuickBird to aerial photos at scale 1:50,000, resulting in the correctly derived from high resolution data is still to be further 
possibility of deriving topographic maps up to a scale in the considered. Important results have been published by the 
range 1:6,000+1:10,000. Concerning orthophotomaps, IKONOS ~~ Ordnance Survey of UK, which have tried to extract different 
images could be used up to scale 1:8,000 and QuickBird up to kinds of features from IKONOS images (see Holland ef a/., 
1:5,000 (Jacobsen & Passini, 2003). 2002, Holland & Marshall, 2003). 
In traditionally well-mapped countries, such as West Europe In order to better explain the topic issues of "fast mapping" 
and North America, where high-quality maps already exist, (FMAPP) production process, an example of geocoded vector 
aerial photogrammetry is generally available and the layers extracted form an IKONOS image will be presented 
infrastructure to process airborne imagery is well established, along the paper. 
then the actual cost of high resolution space data is preserving 
its use in operational environments. The high interest in 
carrying on researches on this field is however motivated by the 2. FAST MAPPING GUIDELINES 
fact that the price of HRSI will lower in the years to come, as 
more suppliers of this kind of data will enter the market. | The goal fo FMAPP is to provide mid-scale regional maps to 
On the contrary, in countries where experience in mapping and countries of the so called “developing world”. According to this 
aerial photogrammetry is not developed, and furthermore the purpose, the selection of source data, techniques for 
37