ISPRS Commission III, Vol.34, Part 3A „Photogrammetric Computer Vision“, Graz, 2002 
  
  
  
  
  
  
  
  
  
  
  
  
   
  
      
  
  
  
  
  
  
  
  
  
           
  
  
  
  
     
  
  
  
  
  
  
  
  
  
  
  
  
   
   
       
  
   
      
    
  
   
  
   
   
Conditions of Object-based in I Then, (S2) simulates spectral emitted and reflected fluxes. 
SYRIEN representation Ernie rules Emitted flux is predicted with the help of equation (3) and the 
User INPUT User INPUT User INPUT User INPUT ; : : . . : 
m (S1) output; ancillary data, like the viewing direction of the 
| 
Sa) EE scene is required here. Reflected flux computation requires the 
computation modelling knowledge of all incident spectral fluxes, especially the sun 
| ( spectral radiation and the atmospheric spectral radiation (Poglio 
Temporal Facet-based 1 i - 
| Toon | | Ed | et al., 2001c). The output of the simulator (S2) is the 3-D scene 
] = wherein the emittance of each element is known for the viewing 
Shadow maps Facet heat direction (figure 3). 
computation conduction 
Y 
Element-based zb Element visibility 
representation computation Temperature Spectral database Conditions of 
representation ( p. t, £) synthesis 
Y (S1) OUTPUT DataBase User INPUT 
Conductive Radiative 
neighbourhood neighbourhood 
Spectral radiative 
computation fluxes computation 
  
  
  
  
  
Wind disturbance Element heat ) ( Form-factor } 
  
  
  
  
  
computation conduction computation 
Emission . Spectral 
for each element incident fluxes 
; ; Conductive Form-factor 
Wind velocity | neighbourhood | | matrix 
  
  
   
Reflective 
fluxes computation 
  
  
P | | m 
Element-based representation 
Y 
( Addition of 
OUTPUT of the pre-processor (S0) Sontributions 
    
  
   
       
. Reflection 
for each element 
  
  
  
  
  
  
  
Figure 3: the detailed architecture of the simulator (SO). (order 7) 
    
  
  
; Global flux 
| Iterative (order k) ; 
All outputs and inputs of (S0) are inputs to (S1). (S1) predicts ve A 
the surface temperature for each element defined by (SO). The | NT mis net reached | 
simulator operates in an iterative way. It predicts the heat | ostlestion | 
  
exchanges, the relative humidity, and the depth-dependent 
temperature for each element and for each time step (Poglio ef 
al., 2001c). The output of (S1) is the surface temperature for 
each element constituting the landscape (figure 4). 
| 
Element-based representation with radiance 
OUTPUT of the radiative spectral module (S2 
  
  
  
  
  
  
  
  
Figure 5: the detailed architecture of the simulator (S2). 
  
  
Element-based In-depth material Thermal and optical Conditions of 
representation constitution database synthesis 
(S0) OUTPUT User INPUT DataBase User INPUT 
computation 
Physical processes 
modelling (I) 
  
  
  
  
  
  
  
  
  
  
  
  
The fourth primary simulator (S3) generates an image for a 
given viewing angle. Depending on the user preference, this 
primary simulator generates an image as it would be seen by the 
sensor, or an image that is only a visualisation of the 3-D scene. 
The first option requires another simulator to model the 
acquisition system, such the AS?-I simulator of Alcatel Space 
Industries. The second option offers a visualisation of the 3-D 
scene. It permits to see the scene without any alteration due to 
the acquisition system. 
  
     
   
  
  
   
  
    
  
   
In-depth element 
mesh computation 
In-depth element 
mesh representation 
—— Initialisation T— Tabulated values 
State of the 
element at f, 
      
  
  
  
The set of these four primary simulators constitutes the 
simulator OSIRIS. The user of such a simulator can be 
interested in not only the final images, but also in other 
additional outputs. The element-based representation of the 
scene, the surface temperature of each element and the emitted 
  
   
  
    
  
  
  
  
  
    
   
  
  
State of the Physical processes 
   
     
    
Differential element at / modelling (1I) 
equation | while flux or the reflected flux for example, are such outputs. They 
lving f s 
nes Fun can be used for the training of future users, and to relate each of 
  
   
      
Differential 
equation solving 
If t=t 
simu 
   
Flux balance at t 
| and humidity the inputs and initial conditions to a change in the simulated 
image. 
  
Element-based representation with temperature and humidity 
OUTPUT of the thermal module (S1) 5. RESULTS AND DISCUSSION 
  
  
  
  
  
  
  
The simulation presented here takes place in Amiens in France 
(latitude: 49.54 N, longitude: 2.18 E). Houses are made of 
construction concrete, and an internal thermal insulation made 
Figure 4: the detailed architecture of the simulator (S1). 
A - 243