CIPA 2003 XIX th International Symposium, 30 September - 04 October, 2003, Antalya, Turkey 
270 
The orthophotos produced by different methods and 
different software have to be compared. The 
purpose of this didactic project was to orient the students 
to experiment theory and practice of terrestrial 
photogrammetry. Inspired by the 3x3 Rules, see (8, 
Waldhausl / Hanke / Ogleby / 6, Herbig), some years 
ago, we in Brazil are very concerned about the 
divulgation of projects with simplified ground control 
and low-cost software, such as PhotoModeler. We require 
that this IME-Project gives some useful information to the 
community, like the project for the Documentation of 
Monuments and Architectural and Archeological Sites: 
The Tower Castle Garcia D’ Avila see (3, IPHAN - 
Fundaçâo Ricardo Franco / IME). 
2. PROJECT PLANNING 
In the initial stage of the project only the Digital Camera 
KODAK DC 265 with the following characteristics was 
available: 
- focal distance: 38 mm 
- approx, format : 26mm xl 8mm 
Considering the camera characteristics and also the large 
dimensions of the object, short camera-to- object 
distances, obstructions like trees, parked cars, etc., it was 
decided: 
- The main façade had to be divided into three 
parts, to be processed independently in 
PhotoModeler. 
- the mean orthogonal image scale should be 
about 1: 1000 
- the scale of the orthophoto should be 1: 200 
- the tolerances for the coordinates of the 
ground control should be : 
- in x- and z-direction : ± 15 mm 
- in y-direction : ± 45 mm 
furthermore the approximate camera stations were 
selected. 
3. GROUND CONTROL 
3.1 Distribution of the targets 
As we were not very familiar with PhotoModeler 
software, we expected some difficulties during processing 
regarding absolute orientation of models. Therefore we 
fixed the rather large number of 34 targets at the object, 
distributed in lines and rows. 4 of these fell down during 
the interval between the taking of images and the 
topographic measurements, (see Appendix) 
3.2 Coordinate systems 
In accordance with the methodologies adopted in 
Industrial Measuring Systems we used two different 
coordinate systems with different characteristics for the 
ground control: 
- the object system , defined in the object 
- an arbitrary local survey system, approximately 
parallel to the object. Both systems have the same origin 
in altitude. Therefore the transformations between the two 
systems can be treated as plane transformations. 
3.2.1 The object system 
(see figure 1). 
Figure 1 - Definition of the Object Coordinate System 
see (7 Luhmann, page 28 ) 
Exploiting the geometric properties of a vertical plane in 
plane objects like façades and walls, the object system 
offers considerable vantages such as: 
- points in vertical lines : x = constant 
- points in horizontal lines : z = constant 
- points in the same vertical plane : y = constant 
- approximately horizontal distances : = Ax 
In many cases of small objects the above mentioned 
coordinate system is sufficient for a simplified ground 
control, which permits the absolute orientation of the 
model and therefore the generation of an orthophoto: 
- some horizontal distances between targets 
measured by a tape define the scale of the model. 
- one vertical line between two targets defines the 
z-axis 
- the y-planes are defined in the object 
- some altitudes of approx, horizontally distributed 
targets, obtained by a level , define the horizontal 
x-axis. 
- some more distances between targets (diagonal, 
vertical and horizontal) improve the geometry of 
the system. 
In this way we performed some projects such as : 
- The Empress Manor, Rio de Janeiro, CIPA 1999 
(stereo), see (5, Gomes). The ground control was 
completed later by densification (see 3.2.2) 
- The Chapel Santa Bárbara, Santa Cruz, RJ, CIPA 
1999, (PhotoModeler), see (4, Gomes). 
- About 15 isolated façades of the Tower Castle 
Garcia D’ Avila, Praia do Forte, Bahia, 2001, 
(PhotoModeler), see (3, IPHAN - Fundaçâo 
Ricardo Franco / IME)