O- 
to 
on 
of 
nd 
of 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
topographic cartography (high resolution digital 
camera or a photogrammetric scanner. Thus, the 
idea was to test medium resolution scanner to 
produce digital images for which the restitution 
result will be confronted to a number of points 
determined by GPS. 
2. MATERIAL AND METHODS 
2.1. Used material 
The material used in this study is composed of : 
1. A computer with the following caracteristics : 
e Speed 800 MHz ; 
e SDRAM 128 Mb ; 
e Graphic card Labs Oxygen de 32 Mb ; 
e Screen ADI 21 inchs with a maximal resolution 
of 1920x1200 pixels ; 
e Fastbus; 
e 3D measuring device ; 
e Hard disk of 10 Gb. 
2. A flat scanner EPSON GT 12 000 with the following 
caracteristics 
e Black and white with a photoelectric device : 
color CCD array ; 
* Effective resolution: 9760 dots per 13760 
pixels at 800 Dpi ; 
e Digitizing resolution : 50 Dpi to 3200 Dpi per 
interpolation ; 
e Color separation : RGB color filters on CCD 
array. 
3. 3D display system (Nu Vision Model # 60GX) 
composed of crystal liquid focusing glasses. 
4. Digital photogrammetric software VirtuoZo integrating 
the following modules : 
° Aerotriangulation ; 
e DTM derivation ; 
e  Orthophos production ; 
* Digital restitution either using IGS module or 
Bentley micro-station software. The latest is the 
| one used in this study. 
| 5. Two stereoscopiques couples covering two different 
| zones, one flat area and a rough muntainous area (Table 
| 
| 
  
  
  
|). 
6. Three GPS receivers (ASHTECH). 
| Area Photos Date Average Topography 
| scale level 
| diffrence 
(m) 
Al 1/40000 | February 300 Rough 
Hoceima 2000 terrain 
| Sidi 1/40000 May 10 Flat 
Slimane 1991 
  
  
  
  
  
  
  
Table 1. Used couples caracteristics 
2.2 Methodology 
The purpose of this research is the study digital 
restitution accuracy for the needs of topographic mapping 
at scale of 1/50 000 and lower. 
Taking into account that the available maximal scanner 
resolution is 800 dpi and that the dimension 30 um is 
acceptable in the majority of aerotriangulation tasks, 
  
DTM, orthophotos orthographies, etc... (Kasser, 2001), 
the selected resolutions for this study are : 300, 400, 600 
et 800 dpi. After scanning the photographs at the above 
resolutions, the results were converted from TIFF format 
to the VirtuoZo software format. The general followed 
methodology during this research is provided by the 
flowchart of figure 1. 
  
  
Left photo Right photo 
  
  
  
  
  
  
  
  
  
  
V 
Photos scanning 
  
  
Y 
Saving under TIFF 
format 
  
  
  
  
Y 
Conversion to VZ 
format 
  
  
  
Vv 
Digital interior, 
relative and absolute 
orientations 
  
  
  
  
Digital restitution 
  
  
  
  
Y 
  
  
  
verification « GPS Points 
  
  
  
  
  
  
Figure | : Flowchart of the methodology 
2.3 Used formulas 
Due to the fact that coordinates are obtained from the 
process which integrates scanning, interior, relative and 
absolute orientations and restitution, their accuracy 
depends on mesearing tasks and consequently on the 
dimension of the pixel. This latest is a function of 
geometric resolution and photographs scale. In this 
context, we present the used formulas for the purpose of 
computing the theoric accuracies : 
- Pointing error on the scanner coordinates is about half 
pixel. It is given as follows : 
O70, 1/2 (1) 
where t the pixel dimension 
- Relative orientation is realized automatically by means 
of images correlation. Its accuracy depends on the 
number of measured points by correlation and on the 
pixel dimension L'orientation relative. Hence the formula 
is : 
Kxt (2)