| 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
cm for has been computed and considered in both zones. Although the 
resight new trajectory has discrepancies too, these are smaller. 
itricity 
LCR In some stereopairs of the selected pieces of the trajectory, 
some well-defined objects are identified using GEOMOBIL 
extraction software. The identified objects are corners of street 
flowerbeds, pavements and water outlets. There are 5 objects in 
zone 1 and 5 in zone 2. Once the coordinates have been 
computed, they are compared against the 1:1000 cartography of 
Sitges. Results of this comparison are summarized in table 3. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Zone | Zone 2 
RMS Mean c RMS Mean G 
Fasting 0.22 -0.11 0.21 0.50 0.38 0.36 
me Northing | 0.13 | 009 | 010 | 039 | 034 | 020 
H 0.26 0.26 0.04 0.48 0.47 0.07 
Table 3: Empirical accuracies in urban environment (units are 
pus in meters). 
us, i 
m has Notice the difference in precision and accuracy in height. 
Precision in height is directly correlated to the precision in 
s height of the computed van trajectory. Indeed Easting and 
S/IMU Northing coordinates are also affected by trajectory errors. 
to: the Differences in mean suggest different discrepancies between 
of the trajectory and map depending on the piece of trajectory. In 
- Up fo summary, the photogrammetric quality of the image subsystem 
ied in is good. However, to make the most of it, GPS/IMU trajectory 
IC IS determination must be improved. 
menn 5. LASER SCANNER INTEGRATION 
hing is Recently, ICC has acquired a terrestrial LIDAR system. The 
s-track. system has been successfully integrated in the GEOMOBIL 
ith the system. 
5.1 Boresight calibration 
Laser scanner is set in a configuration according to the goals of 
each mission. In particular, the laser can be set in two main 
configurations scanning direction upside-down or scanning 
direction right-to-left. Each configuration requires its own 
calibration. 
It is used the same calibration scenario that the used for the 
ifferent CCD camera boresight calibration. On the CCD camera 
results calibration wall 15 of the 60 measured points (see section 3.2) 
n is an were signalized with reflecting targets. The laser from different 
positions and azimuths scans the calibration wall. Targeted 
! points are identified automatically on each scan of the wall. A 
Bundle Block Adjustment is performed taken into account the 
laser observations of targeted points at each laser location, the 
vn near coordinates measured of the 15 targeted points and the positions 
n was J and attitudes computed by the orientation subsystem. In the 
ind the | adjustment there are determined the eccentricity vector and 
station : misalignment matrix (boresight parameters, an amount of 6 
parameters), which defines de relationship between the inertial 
reference frame (the one of the orientation subsystem) and the 
ch was laser scanner reference frame. 
n. Map | 
Some pilot missions have been carried out. Calibration and 
mission results are discussed deeply in Talaya et al, 2004b. 
es with 
. These | 6. FUTURE DEVELOPMENTS 
llowing 
ory had Further developments focus on two issues. The first one is the 
ew one integration of new digital color cameras pointing forward and 
267 
backwards. The second is to improve trajectory determination. 
In general, this improvement may be achieved by integrating 
new sensors (as barometers), which can help in the computation 
of trajectory. In urban environments with existing maps and/or 
aerial photography, some objects may be extracted and used as 
"ground control". These points will be used to improve 
trajectory computation. 
7. CONCLUSIONS 
Since the development of the GEOMOBIL system started, the 
ICC has successfully integrated and operated simultaneously 
two digital CCD cameras and a terrestrial laser range system. 
The ICC development consists of sensor integration, calibration 
and data extraction software. Current experiences prove that the 
GEOMOBIL system exhibits an excellent photogrammetric 
behavior. In fact, experiences confirm theoretical accuracies of 
the image subsystem (plotted in figure 3). It has also been 
proved that the calibration protocol obtains expected accuracy 
so that obtain the best performance of the image subsystem can 
be achieved. 
Despite the excellent results concerning photogrammetric 
capabilities and laser data orientation (Talaya et al. 2004b), in 
urban environments and projects demanding a high accuracy 
the GEOMOBIL system and LBMM systems require a more 
reliable direct orientation subsystem than the current one. 
REFERENCES 
Colomina,I.,Navarro,J., Termens,A.,1992: *GeoTeX: a general 
point determination system", International Archives of 
Photogrammetry and Remote Sensing, Vol. 29, Comm. III, pp. 
656--664, Washington D.C, USA. 
Talaya,J.,Bosch,E.,Alamüs,R.,Bosch,E.,Serra,A.,Baron,A., 
2004. “GEOMOBIL: the Mobile Mapping System from the 
ICC”, 4™ International Symposium on Mobile Mapping 
Technology (MMT'2004). Kinming, China. 
Talaya,J.,Bosch,E., Serra, A., Alamüs,R.,Bosch,E.,Kornus, W., 
2004b. “Integration of a Terrestrial Laser Scanner with 
GPS/IMU Orientation Sensors.”, International Archives of 
Photogrammetry and Remote Sensing, Istambul, Turkey.