International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B6. Istanbul 2004
are polynomial functions of maximum power limited to 3 and
a; are the RFCs’ coefficients.
According to the purpose of FMAPP approach to provide a
mapping coverage at medium scale for developing countries,
where maps are aged or do not exist at all, we have chosen to
orthorectify the IKONOS image with the RFM approach
implemented in commercial SWs using the RFCs supplied with
the IKONOS imagery and a DEM of the area with a step
resolution of 50 m x50 m (see Figure 3).
Recently tests performed by the authors about QuickBird
orthoimage production using non-parametric models, showed
that using sequential polynomial geometric transformations,
such as the RFM and the affine transformation, it is possible to
obtain orthoimages with planimetric precision of about 1.0-1.5
meters using the RPCs supplied with image data and a very few
GCPs (5-6 in our tests).
Figure 3 — DEM of Lecco test site used for orthoimage
generation (step resolution of 50 m x50 m).
3.4 GPS data acquisition
GPS measurement concern two different purposes, i.e.
acquisition of GCPs for HRSI georeference and collection of
GIS data. Both tecniques must be used in relative mode to reach
a sub-meter accuracy To do this, a master station consisting in a
double frequency geodetic receiver must be setup. Moreover,
position of the master station could be computed with respect to
permanent GPS stations, in order to determine its coordinate in
a global geodetic reference system (in practise a realization of
ITRF).
Concerning measurement of GCPs, a rover GPS is moved over
all ground control points to be measured. As illustrated in par.
3.3, HRSI can be used to generate mid-scale orthophotos based
on a very small set of GCPs (usually 10-15 points per image,
covering an area of about 200 km“). The accuracy of their
measurement should be better than the accuracy of image
coordinate. Considering the ground pixel size of about 1 m for
IKONOS, usually GCPs are manually measured so that a
subpixel accuracy is very difficult to be reached. This fact leads
to an accuracy for GPS measurement in the order of 6,,=+40-50
cm, which can be obtnained with ease also by using LI GPS
receiver on long baselines with respect to the master station.
Kinematic and RTK techniques, which could be succesfully
exploited in developed countries, are not yet suitable to be
applied for the developing world. Firstly, the need for real time
measurement does not exist; secondly online communication of
differential corrections may represent a limitation, due to
40
possible long distances involved (for radio modem) or to the the
lack or the weakness of GSM signal; on the knowledge of the
authors, tests about using satellite mobile phones did not
yielded good results up today.
The second application of GPS in FMAPP is devoted to the
acquisition of vector and GIS information to integrate the
spatial DB. This task can be easily carried out by a GIS
datalogger palm receiver, which allows to collect georeferenced
features (points, lines, polygons) and to fill in their attribute
tables directly on the field. Different classes of these kind of
receiver exist, the most evoluted registering also phase
measurement. This fact result in the possibility of signal post-
processing, by differentiating it with respect to that acquired by
a master station. Accuracy in the sub-meter order for kinematic
points may be reached as far a distance of 30-40 km from the
master. Moreover, some receivers (e.g. Trimble Geoexplorer
CE XT) permits to determine also static points by registering
several epoques during the stationement on the same position.
This possibility, together with the an accuracy under 0.5 m,
might lead to the use of only GIS datalogger as rover receiver,
finalized to both purposes of GCP measurement and GIS data
collection. Practical exploration of this chance in a testfield
would be very interesting for development in FMAPP.
3.5 Cartographic reference system
The reference system of FMAPP must be linked to the use of
GPS measurement for georeferencing satellite imagery,
resulting in a given ITRF realization. This means that the
ellipsoid to be used is WGS84.
The cartographic coordinate system which is the most suitable
is undoubtely the UTM system, based on WGS84 ellipsoid
(UTM-WGS84). This selection would unify cartographic
reference systems used in different countries, following a trend
which in European countries does as well (e.g. in Italy).
Concerning transformation of points into own coordinate
systems of each country, an approximate mathematical relation
and one or more sets of suitable parameters should be provided.
However, this problem is not typical of FMAPP only, but
involves every mapping and surveying activity; in each context
this concerns should be analyzed in detail.
3.6 Vector information capture
The orthoimage derived from HRSI is a potential source of a
very huge information to be extracted. Nevertheless, barring the
work of Holland et al. (2002) and Holland & Marshall (2003), a
still insufficient effort has been carried out so far in analysing
which kind of vector data can be derived.
We performed a small test on this topic, considering a portion
of the orthophotomap generated from IKONOS image over the
area of Lecco, and trying to extract vector layers typical of a
1:10,000 map. In particular, layer representing roads and
buildings have been drawn.
In Figure 4 is reported a pacth from IKONOS image used for
this test, showing an area typical of a mid size town suburbs,
with cottages and intercity roads. In Figure 5 the same
orthoimage with superimposed some extracted vector layers
(buildings in red and roads in yellow) has been depicted.
Extraction of roads sounds to be the easier task, being these
objects well-identified in the image background. More difficult
has resulted the drawing of buildings, especially in case they
are very close to each other or are partially covered by
vegetations; however, in the second case the same problem
would sussist also in case aerial photogrammetry is applied.
The availbility of colour information would largely help in
