The 
cene, 
fying 
ty. In 
hicle 
lobile 
der to 
used 
es, as 
d two 
digital 
erence 
d as a 
h new 
fferent 
ization 
supply 
ization 
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
The orientation subsystem is mainly based on the integration 
of a GPS receiver and an IMU sensor. However, due to the 
large number of GPS outages that occur in a terrestrial 
campaign, it also includes the integration of a DMI sensor 
(Distance Measurement Indicator), which is used to reduce 
IMU drifts during these outages. The orientation subsystem 
is completed with a GAMS system (GPS Azimuth 
Measurement System), an approach from Applanix based on 
the heading determination using two GPS antennas mounted 
on the top of the van, which allows a rapid heading 
correction after GPS outages. 
In order to correctly transfer the reference frame from the 
orientation subsystem to the CCD cameras, a rigid structure 
was designed (Figure 2). The reaction of the structure under 
different forces was modeled and the latest design showed 
deformations of less than 1 mm in distance and less than 70 
arcseconds in angle. The structure acted as a platform for 
integrating the sensors used for the orientation and any other 
sensor mounted on it. In particular, the integration of the 
digital cameras and the GPS/IMU orientation subsystem lead 
to very good results, photogrammetric points are determined 
with accuracies better than 5 cm in across track directions 
and 13 cm in along track direction at an average distance of 
18 m (Alamüs et al, 2004). 
The precise orientation computed by the orientation 
subsystem (GPS/IMU) can be transferred to any sensor (in 
particular to the terrestrial laser) mounted on the platform. 
Thus, the laser data can be directly oriented applying the 
same principle used for airborne lasers. 
   
Kant Sora 
31 1668-9 45% 3308-07 
Figure 2: Geomöbil integration platform (response to stress) 
3. INTEGRATION OF A LASER SCANNING 
The laser selected for the integration was a Riegl Z-210 that 
is able to collect up to 10000 points per second. For each 
point a distance measurement, an intensity value and RGB 
data is collected. The laser has a rotating mirror that allows 
taking vertical profiles while a  servomotor rotates 
horizontally the whole laser for scanning a static scene (see 
figures 3 and 4). For each laser point also the angle readings 
of the mirror and the scan encoders are obtained. These 
angular values, together with the distance measurement, are 
used to locate the measured point in a local laser reference 
frame. The raw data collected by a terrestrial laser are 
usually parameterized in a spherical coordinates frame, 
991 
denoting r the distance measured by the laser, ¢ the rotation 
angle of the mirror and ¢ the laser position angle during the 
scanning (see figure 3). 
  
c 
= eet 
|. 4 
= cA 
| (p 
+ | 
1 
Figure 3: Terrestrial laser spherical coordinate frame 
(courtesy of Riegl LMS GmbH) 
  
Figure 4: Static terrestrial laser scene coded with a 
combination of intensity and distance. The rotating 
mirror scans the scene in vertical lines while the 
laser rotates around a vertical axis covering the 
scene in the horizontal direction 
The transformation from the laser spherical coordinate frame 
to a laser cartesian coordinate frame is given by following 
equation: 
X=r-sing-cose 
y=r-sing- sing 
Z=7 C080 
Equation 1: Preliminary transformation from the laser 
spherical coordinate frame to a laser cartesian 
coordinate frame 
By construction the laser axes are not perfectly aligned, the 
mirror rotation axis (¢ angle) and the laser scanning axis (¢ 
angle) do not intersect, so this difference has to be corrected 
in order to transfer the initial laser spherical coordinates to a 
local laser cartesian coordinates frame. In equation 2 the 
formulas used by the laser software (Riegl, 2001) to perform 
the transformation are described. Assuming that the origin of 
the laser cartesian frame lies on the axis defined by the 
rotating mirror (@ angle), r, is an offset on the distance 
measurement and q, is the misalignment of the scanner 
rotation axis: