ws the 
stems: 
:ontrol 
image 
pration 
rame). 
mance 
asured 
of the 
1OBIL 
uch as 
orm is 
ms for 
ioning, 
ty and 
mobile 
  
lowing 
'al and 
ase for 
> frame 
tration, 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
eo Image Acquisition: handles the exposure 
synchronization and control, image acquisition 
and storage. 
e Synchronization: creates the temporal reference 
frame coherently transferable to all sensors. 
e Power and Environment Control: guarantees 
power supply and stabilizes the operational 
environment conditions for all the sensors. 
Apart from the subsystems installed on the vehicle that run 
during data (orientation, synchronization, image and laser 
ranging) acquisition, the GEOMOBIL system is made up of a 
sensor calibration procedure and data extraction software. 
2.1 Orientation subsystem 
The orientation subsystem is responsible for georeferencing the 
photographs taken by the GEOMOBIL. Thus, it provides the 
coordinates (position) and the angles (attitude) of their 
projection centers. This subsystem is based on an Applanix 
system, which is specially designed for land vehicle 
applications and is integrated in the GEOMOBIL. This system 
is basically composed of: 
e An IMU (Inertial Measurement Unit), a sensor 
that provides measurements of accelerations and 
angular velocities. 
e Two sets of GPS antenna-receivers, one of double 
frequency to provide observations of the position 
and velocity, and the other one of single 
frequency to improve the heading angle 
determination. This system is called GAMS (GPS 
Azimuth Measurement System). ) 
oe A DMI (Distance Measurement Indicator), a 
sensor directly installed on one of the vehicle’s 
rear wheels which provides information of the 
distance traveled. 
e APOS Computer System, which contains the 
core of the system, IMU and DMI interfaces, two 
GPS receivers and a removable PC-card disk 
drive where data is stored. 
e A POSPac, software for processing GPS data and 
integrating the GPS solution with the 
observations of the other sensors. 
Like any system that combines inertial and GPS observations at 
a high level of integration, GPS derived trajectories are used to 
correct and calibrate the drifts of IMU gyros and accelerometers 
so that the position and velocity errors derived from inertial 
sensors are minimized. However, the main drawbacks for 
terrestrial navigation are the presence of obstacles on the road, 
like bridges or tunnels, which interrupt totally or partially the 
acquisition of GPS observations during some time interval, and 
the existence of areas where most of the GPS satellites signals 
are blocked by terrain conditions, like urban areas with high 
buildings, forest zones, etc. In these areas without GPS 
coverage or with a very poor constellation, position and 
velocity are calculated from IMU observations, whose errors 
only depend on the distance traveled since during the GPS 
signal outage DMI observations are used. 
In order to obtain the position and attitude of photographs from 
the position and angles provided by the orientation subsystem, 
it is important to fix the relation between all the reference 
frames of the orientation process. For this reason, the 
relationship between the IMU, cameras and GPS must be totally 
stable. 
2.2 Integration Platform 
The integration platform is the structure where the different 
sensors are mounted for their operation. This platform must be 
sufficiently stable for the precise transference of reference 
frames. Two basic requirements must be considered, namely 
that the platform must have a maximum physical space at the 
top of the van, and the geometry of the platform must be totally 
stable in order to transfer the global reference frame (computed 
from the GPS/IMU data) to any sensor installed on the 
platform. This implies high immunity to deformations. The 
design of the platform was studied [3] and several options were 
analyzed. The optimal solution is based on an irregular mesh 
system with diagonal reinforcements, as can be seen in figure 2. 
This structure is equipped with equidistant anchorage points so 
that different sensors can be easily distributed. 
  
Figure 2: Integration Platform simulation with the diagonal 
reinforcements. 
As explained above, the biggest constraint in the design of the 
platform and the anchorage system for the sensors has been the 
stability requirements. The maximum deformations tolerated 
between the reference center of the absolute frame (IMU) and 
the reference center of the relative frame (Camera) are 1mm in 
displacement and 70 arc seconds in rotation. 
2.3 Image sensor subsystem 
The subsystem design has been driven by two main 
requirements: to acquire images of at least 1Mpix and to get 
10m stereoscopic overlap at a 10 m distance from the van 
(about 100 m?). The selected image size is a compromise 
between image resolution and data storage and management. 
The stereo overlap requirement is conditioned by two factors: 
getting the maximum stereoscopic overlap free of obstacles 
(between the vehicle and the objects of interest) and preserving 
a B/D ratio (stereoscopic base — object distance) as good as 
possible (see figure 3). Table 1 summarizes the image sensor 
subsystem characteristics. 
  
  
  
  
  
  
  
  
  
  
  
No. Píxels 1024x1024 
Pixel size 12 um 
Focal length 10.2 mm 
FOV 62.13? 
IFOV 3 min. 38 sec. 
Stereoscopic overlap @10 m 10:55 m 
Precision@10 m (across-track) 0.8 cm 
Precision@10 m (along-track) 5:6 cm 
  
Table 1: technical features of on-board image sensors. 
Figure 3 shows some significant photogrammetric figures. 
Notice the dependency of the along-track photogrammetric