OR 
jased on 
rational, 
sists of a 
includes 
nents for 
mance is 
e market 
) a wide 
ned with 
y 1980s 
| trend is 
"remote- 
Maguire, 
rried out 
up with 
latanova, 
nces by 
collector 
vith the 
nd close- 
are these 
of non 
d already 
ven low- 
is even 
rocessing 
initiative 
913. “This 
ay using 
; already 
of a web 
enmeyer, 
nentioned 
evelop e- 
| Service 
xample a 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
  
  
Figure 1 : System architecture 
The last part of our work is related to 3D modelling and the 
management of 3D information obtained from photogrammetric 
processing. This is essentially based on 3D reconstruction 
methods from image sequences using computer vision 
techniques, like projective reconstruction and self-calibration. 
Important works and research were undertaken in this field 
(Faugeras, 1995; Debevec, 1996; Koch, 1999; Leymarie, 2000; 
Pollefeys, 2000). Web 3D applications, not only for visualisation 
but also for data entry, have been the subject of research 
(Zlatanova, 2000) using VRML as a basis. An example of 
implemention was done by Pharus s.rl. (www.int3d.com). 
Concerning City models, they are being widely used in some 
European cities and some degree of automation has already been 
implemented (Fórstner, 1999). City model building will certainly 
benefit from the development of telephotogrammetry. 
The implementation of a photogrammetric service will further 
require the development of software architecture on the network. 
Web service developed for this purpose will help establish a 
communication pool in the network and deliver spatial data 
between various systems (Parsons, 2003). 
2. CONCEPTION 
This chapter will explain more in-depth the concept elaborated 
about telephotogrammetry. 
2.1 System architecture 
Telephotogrammetry relies first on a communication network 
used to exchange information (data) and to allow interaction 
between the user and the machine performing the processing 
(Figure 1). 
The Mobile and Geographic Imager (Mobile Geolmager or MGI) 
has the task to acquire coordinate and image data and to 
synchronize them. The data are stored temporally locally but then 
transferred to the Processing Center (PC) using mobile 
communication and internet network. The PC receives and stores 
the acquired data in a Database Repository (DBrep) and wait for 
user commands to carry out given high computational intensive 
tasks. 
One or several MGI may exist in the network, so that multi-user 
tasking is possible within a defined project. Tasking may also be 
triggered from an Office Computer (OC) when it cannot be 
GSM 
  
Mobile and 
Geographical Imager 
(MGI) 
  
efficiently done on outdoor conditions. In this context, the 
network has to support heterogeneous systems and a common 
language should be used to harmonize the work flow. Such an 
organisation has however the advantage to speed-up workflow. 
2.2 Available functions 
Data received from the MGI is processed according to the user 
commands. High computational processing will take place on the 
PC. Several levels of processing are identified: 
1) No processing: the image is stored in the database 
2) Mosaicing: sequence of images are combined together 
Relative and absolute orientation: coordinates of 
camera point of view are retrieved. 
4) Measurement: sequence of images is used to perform 
direct measurements 
5) Modelling: sequence of images is used to build 3D 
models interactively 
6) Texturing: texture extracted from images are applied 
to 3D surfaces 
7) External texturing: Texture obtained for external data 
are applied to 3D surfaces 
8) Precise orientation using textured 3D city model 
2.3 Data visualisation 
The data contained in the DPrep can be visualised on the MGI or 
on the OC. During visualisation the level of processing is taken 
into account and on-the-fly processing is performed using 
parameters from the processing stage. Various visualisation 
interfaces are used for either 2D or 3D data. 
2.4 Virtual processing 
The acquired data stored on the DBrep is not altered by the 
processing. Only the parameters computed to perform the 
process is recorded. This helps maintain data quality as best as 
possible and reduce the volume of data. When desired, function 
to bring data to production will be available. 
3. IMPLEMENTATION 
This chapter will bring more details about technological solution 
or solutions to satisfy the concepts described in the previous 
chapter. 
    
  
     
INTERNET 
    
Processing Center 
(PC) 
Office Computer 
(OC)