CIPA 2003 XIX th International Symposium, 30 September - 04 October, 2003, Antalya, Turkey 
potentially influential variables. The results will allow for a 
discussion concerning aspects such as the adaptation of the 
software to the type of geometry, which factors can limit the 
DSM precision and possible improvements or corrections in 
order to advance in the automatic architectonical generation of 
DSM. 
2. MATERIAL AND DATA 
2.1 Material 
The photographs were done with Agfapan APX 100 
Professional (Agfa, 100° ASA) and Technical Pan 6415 
(Kodak, 25° ASA) film. A semi metric Rolleiflex 6008 with a 
60 mm objective was used for the exposure. The negatives were 
digitalized with an UltraScan 5000 (Vexcel) scanner. The 
objective focal length was determined in relation to the space 
available in front of the façade. The two films were selected in 
order to check for possible differences due to different 
sensitivity and grain size. 
The photogrammetric operations have been carried out with 
Erdas Imagine® 8.5 with OrthoBase Pro (Leica Geosystems). 
This application is the only commercial program that allows the 
adjustment of convergent images and the use of automatic 
stereo matching for DSM construction. The error calculus and 
subsequent analyses were carried out with the SIG Arc View 
3.2a (ESRI) and the extensions with Spatial Analyst 2.0a and 
3D Analyst 1.0. 
2.2 Data 
This work was carried out on one of the churches of the 
historical centre: San Mateo Church, sited in Câceres (Western 
Spain, city declared World Heritage by UNESCO in 1986). The 
Southern façade has a complex design due to the presence of 
one tower and some relieves that frame the main gate. All these 
elements make this church ideal for our study (Figure 1). 
Figure 1. Southern façade of San Mateo Church. 
3. METHODS 
3.1 The taking and pre-processing of data 
The taking and preparation of data can be divided into the 
following stages: 
3.1.1 Geodetic network. The form and size of the signals 
were designed for an optimum recognition in relation to the 
characteristics exposure and digitisation. The point size is 
31x31 mm, which corresponds approximately with a matrix of 
17x17 pixels (see 3.1.3 and Figure 2). 
Figure 2. Signal designed for an oblique exposure. 
The façade was marked with signals that determined the 
position of a total of 12 control points and 75 check points 
(Figure 3). The positioning of these points in the local system of 
coordinates was calculated through simple or multiple direct 
intersections with a reflectorless total station TCRM 1102plus 
(Leica Geosystems). 
3.1.2 Photographic exposures and developing. The points 
of photographic exposures form a convergent network of 11 
images. The central image is perpendicular to the façade and 
the rest of them are taken with an increment of 5° at both sides 
up to a maximum of 25° (Figure 4). From each point, two 
pictures were taken with the two films we are working with 
(APX 100 and Pan 6415). 
Figure 3. Positioning of the control and check points on the 
façade. 
The developing of the photographic film was carried out with 
the best products according to the manufacturer’s