- 105 - 
LASER SCANNING AND TRADITIONAL SURVEY INTEGRATION TO BUILD 
A COMPLETE 3D DIGITAL MODEL OF 
“SAGRESTIA DELL’ARCHIVIO DI STATO A MANTOVA” 
Roberto Cantoni a , Giorgio Vassena a , Carlo Lanzi a 
d University of Brescia, Civil Engineering Dep. 
Via Branze, 38 - 25123 Brescia Italy Tel. +39.030.3715516 Fax. +39.030.3715503 
E.mail: robypost@mail.tele.dk vassena@bsing.ing.unibs.it lanzi@bsing.ing.unibs.it 
KEY WORDS: Laser scanning, digitai model, cultural heritage recording 
ABSTRACT: 
The main tasks of this work consisted of the look for an integration between methods of traditional survey and laser scanning 
technique in the 3D architectures survey. 
The intentions to build a model that is able to satisfy the demands of precision, for example, to planning an intervention of recovery 
and therefore to editing architectural plates, but also those of aesthetical quality and representative effectiveness, they have 
conditioned the choices that have driven this work. In other words has been estimated the effectiveness and the potentialities of this 
new survey method , which in this case was supported and assisted with traditional methods that have also allowed to test the model 
obtained from laser scanner. 
During this work we have realized how this new method of survey, that allows a quick acquisition of million points, that allows the 
description of even very complex geometries, imposes many evaluations beginning from the fitter tool until the choice of target 
points and about the method of textures mapping, which are very important for the quality of final digital model. 
1. INTRODUCTION 
The Sacristy of State archives in Mantova, is an architecture 
that allows to appraise the effectiveness of this integration: 
inside the Sacristy there are many decorations and plasters that 
would require a long and hard-working traditional survey. The 
Sacristy has an octagonal plant, 15 meters high and about 12 
meters wide. 
There are many ways of creating 3D model of a building and 
many technologies can be employed to record the spatial and 
visual complexity of real word environments. The work 
described in this paper, wants to underline the necessity to 
integrate these different techonolgies to optimize the costs, 
accuracy and visual quality of the final 3D model. Another 
important thing is the approach to create a complete 3D not 
limited to an internal or anyway partial reconstruction but a full 
3D composed from an internal end external surface to create the 
as completed as possible representation. In this way is also 
possible to give a more realistic perception about the building 
volumes. 
The 3D model obteined is not only a photo-realistic 
representation of the reality building but is also dimensionally 
rigorous so it has also an important metric value which makes it 
a powerful tool of investigation and study of the architecture. 
Is important to underline how from such product is possible to 
obtain a series of graphic useful results for the planning and 
design of interventions to the building safeguard. 
2. LASER SCANNING SURVEY 
The laser scanning survey was made with Cyrax 2500 laser 
sensor, that has single point accuracy equal to ± 6mm. The 
scanning procedure took up 1 working day, using 15 different 
viewpoints. In some of these the scanner was mounted on a tri 
pod while for the other it was just leaned on the floor. The 
scanning survey has been articulated in the follow steps. 
• Planning the distribution of laser viewpoints. 
• Laser data acquisition. 
• Images acquisition by digital camera. 
• Scans merging and orientation in an absolute reference 
system. 
• Meshing to produce a triangulated 3D. 
• Editing operations on the 3D model. 
• Texturing the photograph images on the geometric model. 
In the most cases it is not possible to have complete 3D 
representation from data acquired at a single viewpoint. Indeed, 
to resolve occlusions in the scene or to reconstruct large scenes, 
it is required to have 3D data acquired from multiple 
viewpoints. So it was necessary to locate the sensor at several 
positions inside the sacristy, since all surfaces may no be visible 
from a single point or will data be acquired at sufficient 
resolution. 
The problem of determining the next capture point depends on 
how much a priori information about the scene is available. The 
selection of the next capture point should consider both the 
resolving occlusions in the already extracted model and the 
acquiring additional views of the non-modelled environment. 
The latter are required to have a sufficient overlapping region 
between two scans so to aim their integration. Is easy to see that 
the quality of the final model depends highly from the planning 
strategy and in particular from the distribution of the viewpoints 
inside the architecture. Another important thing to consider is 
that the number of these points, from where the data is acquired, 
should be minimised to reduce the time needed for a complete 
scene reconstrucion. These considerations are important also for 
the points where photograph images are taked. The problem to 
chose the viewpoints position represent an important phase of 
the survey, in particular for a complex architecture like this one. 
Recently have been proposed some algorithms that allows to 
determine the best viewpoints position considering many