Ismael Colomina
6.2 Consider using IGS data and derived products
The mission of the International GPS Service “is to provide a service to support, through GPS data products, geodetic and
geophysical research activities.” Beyond that, the IGS is providing the general geodetic, mapping and Earth Observation
community, with an extraordinary, daily data set at no cost. There are many sources of IGS related information. The IGS
web site is worth the visit of anyone in the photogrammetric community.
The IGS products are: high accuracy GPS satellite ephemerides, Earth rotation parameters, coordinates and velocities
of the IGS reference tracking stations, GPS satellite and refernce station clock corrections, ionospheric information and
tropospheric information. All this products can be dowloaded from the IGS web sites and are formatted according to well
documented and stable formats. The IGS file formats are supported by all major GPS SW and HW manufacturers.
Before deploying an expensive ground GPS infrastructure, have a look at the IGS web site and see how much you can
“steal” from it. The same holds for reginal similar services like EUREF in Europe and for local services like those
maintained by the local geodetic agencies.
6.3 Fix a policy for sensor calibration
With INS/DGPS as the primary tool for orientation the calibration of the system becomes critical. Depending on external
and internal factors one could go for a small test field nearby the company headquarters and check the calibration peri-
odically. Or one could go for calibration blocks tailored to projects. One should have a policy for when (2-3 times per
season), where and how to check calibration.
6.4 Fix a policy for quality assurance
In the early stage and because of INS/DGPS is relatively new in this activity sector and because of the calibration un-
certainties, define and measure a small set of GCPs (Ground Control Points). The GCPs can be further entered into a
bundle adjustment involving just their images. (Note that with a suitable calibration model, a connected bundle network
is defined so the measurement of a few GCPs can be used for an optimal estimation of calibration parameters. Note as
well that this does not require any particular block geometry.)
6.5 Account for the second orientation step
This has been discussed in sections 2 and 3.
6.6 Develop and/or select the right set of tools
If one wants to stay on the safe side of the process, one needs a few more SW pieces than those offered by the man-
ufacturesrs. There is still a gap between the INS/DGPS SW packages and the photogrametric ones. This gap includes
calibration, geodetic reference frame transfers, and IMU-camera missalignment determination. These tools are only
available at universities and at some companies. The situation is even worse for creative users who might be interested in
implementing their own strategies like those proposed in this paper.
7 THAT ALL IMPORTANT SW FLEXIBILITY
Software development is a rather complex and expensive task. Manufacturers of software systems are forced to make
decisions on what to support and what not. It would be naive to ask that every “brilliant” idea is implemented. However,
a high degree of flexibility and efficiency is proven to be feasible (Navarro, 1999, Navarro, 2000). This development
at ICC (Institute of Cartography of Catalonia) is an example of how medium sized organizations can benefit from SW
development and integration. The Socet Set of LH Systems is another example of a flexible approach. On the other side,
there are stereo model set-up modules in photogrammetric workstations that do not allow for the introduction of exterior
orientation elements as observations. The same holds for some AAT packages. Both the systems and their users will
benefit from a slightly more open SW interfaces.
8 CONCLUSION
Much progress has taken place in geodetic positioning and attitude determination since the introduction of GPS in aerial
triangulation in the late eighties. In the next decade, the GPS modernization and the Galileo initiaves will bring addi-
tional accuracy and reliability to satellite RFR positioning. Progress has also taken place in the automatic generation of
terrain models by image matching —if not to produce acceptable contour maps at least to produce acceptable heights for
196 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000.