04 
in 
be 
he 
ca 
ud 
of 
ial 
xd, 
he 
ler 
no 
he 
ed 
he 
ne 
ue 
ial 
nd 
tal 
ed 
he 
he 
us 
ne 
lar 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B6. Istanbul 2004 
  
2.5 Tecnology transfer 
Thanks to its flexibility and independence from local 
infrastructures, the FMAPP process is largely prone to be 
transferred to the operator in the country where the project is 
carried on. Almost all the stages of the production workflow, 
barring GPS measurement, consist on data processing 
operations, suitable to be accomplished by means of 
commercial SWs. These are usually provided by ease-to-use 
GUI, by help and tutorial (on-site or on-line via WEB). 
Furthermore, the FMAPP process should include an initial 
training stage, where personnel involved can be adequately 
formed. Also operations concerning data acquisition by GPS 
can be performed by not very expert operators, thank to the ease 
on use of currently available GIS datalogger. 
3. AN EXAMPLE OF FAST MAPPING 
A more detailed description of the process to yield maps based 
on the FMAPP approach will be proposed in the sequel by 
presenting a pratical example. 
3.1 Site selection 
Due to the location of Politecnico's laboratory in Lecco (lake of 
Como, Northern Italy), we have preferred to use an image 
acquired over this area. This fact results in four main 
advantages: 
1. a simplification in GCPs (and check points) 
measurement and in data acquisition on the ground by 
GPS; 
the terrain presenting a very large variety of different 
scenarios, so that urban, rural, hilly and mountain 
contexts could be imaged in the same data set; 
3. the large availability of other cartographic data to 
check the results, such as colour orthophotos 
(1:10,000), regional raster maps (1:10,000), vector 
map of urban areas (1:2,000); furthermore, new data 
acquisition are forthcoming and will be used for other 
comparisons; 
4. theavailability of a DEM. 
> 
3.2 Image data set 
A pair of adjacent IKONOS panchromatic images have been 
acquired over the interested area (see Figure 1) — Lecco 
testfield. These data have been collected in June 2001, and are 
stored in the Spacelmaging on-line archive. According to the 
purpose of FMAPP approach, the lowest price IKONOS 
product has been choosen, i.e. the CARTERRA Geo product, a 
rectification of the image to a plane with constant height. More 
details about the satellite structure and the delivered data can be 
found in Gerlach (2000). In Table 2 some important features of 
the images are reported. As can be seen, images present a 1,200 
m elevation range and could be considered as a very realistic 
test. 
  
| BERGAMO. — 
i. 
Figure 1 - IKONOS images over Lecco test field (from 
Space Imaging CARTERRA ONLINE). 
  
Image Sensor IKONOS 
  
Image Type Panchromatic 
  
Spectral Range 450 - 900 nm 
  
Processing Level Standard Geometrically Corrected 
  
Nominal GSD 0.84 m 
  
Interpolation Method Cubic Convolution 
  
Datum WGS84 
  
  
  
Map Projection UTM-32N 
  
  
Table 2 — IKONOS image main features. 
3.3 Orthoimage generation 
The orthorectification process converts imagery into map-like 
form by accurately removing all camera and terrain related 
distortions. In order to georeferenced images acquired with 
spaceborne sensors, two different approaches have been 
developed, based on a parametric (physical) and a non- 
parametric (generalized) model. 
Parametric models, based on the collinearity equations, are 
physical models that describe the physical imaging process. 
They need the knowledge of the sensor model and the position 
and the attitude of the sensor during the acquisition. 
Non-parametric models are generalized sensor models 
(platform independent) that use general functions to compute 
the transformation between the image and the ground reference 
systems. As mapping function, the Rational Function Model 
(RFM), based on the ratios of polynomials with different 
degree, is widely used (Tao & Hu, 2001; 2002). 
The sensor model for IKONOS images, as like as for the others 
HR satellites, is not available to users, but Spacelmaging is 
distributing the relation of the Geo-Image to the national 
coordinate system in form of rational functions coefficients 
(RPCs) - see Tao & Hu (2001, 2002). They do describe the 
scene position (r,, c) as the relation of a polynomial (RFM) as 
function of 3D ground coordinates (X,, Y,, Z,): 
| Pi(Xns Yn. Za) 
n " 
Pr Xp» Yn. Zp) n 
( 
P3 (Xp, Yn, Zn) 
en = 
P4(Xn> Yn> Zn) 
where: 
| m, m5 m, ed 
Pi(Xp: Y n) = X à ag X Y (2) 
i=0 j=0 k=0 
39