International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
hand, knowledge about the object can be considered: planarity 
of surfaces, continuity of curvature, symmetries, identities and 
other knowledge can be used by an operator in a semi- 
automatic process to create the DSM. This is provided by the 
measurement program FAMES (Facade Measurements) and the 
interpolation and editing tool FASED. 
FAMES allows the semi automatic measurement of profiles. 
Only a few of these profiles are necessary for the generation of 
DSM if the fagade is regularly structured and organized in 
window axes and partially identical floor levels. Some profiles 
may be duplicated, due to their identity. 
This semi automatic profile measurement is done with a 
motorized and reflectorless working total station (tested with 
Leica TCRA 1100 series) under control of a mobile computer 
(Fig. 1). The semi automatic process (Juretzko 2002) cares for 
the correct leveling during the measurement of horizontal 
profiles and the control of the correct location for vertical 
profiles. In the first run points are measured in a predefined 
metric or angular distance. In a second step an automatic 
process locates significant depth jumps in the profiles (Fig. 2). 
From a second station the occluded parts of the profiles may be 
filled if they are visible from this new point of view (Fig. 3). 
Even regularly structured façades need some additional 
measurements of polygons around individual features (e.g. 
chimneys, balconies, protrudes) and a façade surrounding 
polygon (Fig. 4). 
3. DERIVATION OF DSM FROM PROFILES 
With FASED the façade area is divided into rectangular fields 
with the edges at the non static positions of the profiles and the 
assumption of a continuous curvature in the fields. In the first 
step the fields touched by a profile are filled with the 
corresponding height value. In a second step all adjacent fields 
are filled step by step based on similarity to already filled fields 
and other knowledge about the façade. Finally the additional 
measurements of individual features are added by deriving their 
height from their surrounding polygons. 
  
Fig. 5: Digital Surface Model (left), Digital Orthoimage (center) and derived CAD data (right) 
The generation of the DSM is done under supervision of the 
operator. Helpful is also a wide mashed raster (e.g. 1 m grid 
width) measured over the facade to confirm the DSM derived 
from the profiles, to solve ambiguities and to detect irregu- 
larities and deformations in the fagade. 
The DSM is stored as a regular raster grid in a sufficient grid 
width, usually as 8 or 16 bit image file (e.g. TIFF). This offers 
the opportunity to visualize and edit the DTM with a wide 
variety of software tools. The result of this semi-automatic 
process is a DSM to fulfill the requirements for the generation 
of high quality true orthoimages for architectural applications. 
4. PRODUCTS 
The quality of the DSM can be checked by doing differential 
rectifications of images from different points of view. The 
differences in the orthoimages result from differences in the 
DSM, errors in the orientation process or obstacles in the 
foreground. The bundle block adjustment (e.g. with the 
Archimedes3D tool IMBUN) provides quality control data for 
the orientation data. Differences between the orthoimages from 
different points of view indicate areas that require special 
attention. The erroneous parts of the DSM can be located, 
delimited and improved — using knowledge about the facade, or 
additional measurements at the site or in the available image. 
Archimedes3D also provides tools to generate different types of 
documentation requested in heritage documentation: image 
mosaics (with MOSAIC) or 2D or 3D line drawings (Fig. 5) by 
mono or 3D plotting (with IMDIS). Under consideration of the 
DSM the 2D CAD data acquired by plotting in one image can 
be extended to 3D data. The rectified images and mosaics can 
also used directly for geometric correct further 2D analysis 
(Lerma 2002). 
5. CONCLUSION AND OUTLOOK 
Archimedes3D allows the documentation of architectural 
  
  
  
  
  
  
  
  
  
  
  
   
Internat 
BES EE 
objects 
hardwar 
designe 
and mid 
line bas 
work h: 
limited 
tools m: 
and la 
www.ar 
We ack 
program 
who su] 
thank J 
student: 
engager 
Bóhler, 
Investig 
Int. Sy 
Internat 
Sensing 
Dequal, 
Archite 
Sympos 
Internat 
Sensing 
Hemml 
and 
Photog 
Photogi 
Conser 
Remote 
Schwec 
Juretzk 
Geomei 
Monum 
Potsdar 
for Ph 
XXXIV 
Lerma, 
Rupesti 
XVIII. 
2001. I 
Sensin; 
Schout: 
(2002): 
Conser 
Potsdar 
for Ph 
XXXh 
Wehr, . 
and La 
Int. Ar