1077
SURFACE RECONSTRUCTION OF LARGE COMPLEX STRUCTURES FROM MIXED
RANGE DATA - THE ERECHTHEION EXPERIENCE
S. El-Hakim *, J.-A. Beraldin, M. Picard, L. Coumoyer
Visual Information Technology (VIT), National Research Council (NRC), Ottawa, Canada - (Sabry.El-hakim,
Angelo.Beraldin, Michel.Picard, Luc.Coumoyer)@nrc-cnrc.gc.ca
KEY WORDS: Cultural Heritage, Acquisition, Laser scanning, Modelling, Registration, Multi resolution, Visualisation.
ABSTRACT:
We present the procedure and technologies used to modelling of the Erechtheion at the Acropolis in Athens, a large complex
structure with many challenges to data acquisition, processing, and visualisation. We used two scanners, one for medium range high-
resolution coverage (1-5 mm lateral data spacing) and one long-range scanner to capture some top sections occluded from the first
scanner. We also used high-resolution digital images for image-based reconstruction using Photogrammetric and image matching
techniques and for texture mapping. We focus in this paper on the geometric reconstruction from the range sensors. Several issues
had to be addressed. Our system must be able to work with data obtained at different resolutions and accuracies and acquired from
any viewpoint. The marble surfaces were of different age from one part to another due to decades of restoration. This resulted in
different reflectance properties and apparent laser penetration that varied from area to area. Another problem resulted from the huge
amount of data. The billions of generated 3D points by all sensors are currently beyond the processing capabilities of commercially
available software or hardware. Reducing the data to a manageable size without losing important details had to be addressed.
1. INTRODUCTION
The Erechtheion (figure 1), completed in 406 B.C., is a large
complex monument that consists of several different sections
and compartments. It is built on a slope with the south and east
sides about 3 meters higher than the north and west sides. The
entire structure is made of marble and remains impressive in
spite of the fact that it is only a remnant of what it used to be
and missing most of its art and decorations. For example, as
shown in figure 2, the friezes are missing their triglyphs and
metopes, and only very small parts of the pediments remain.
The famous porch of the Caryatids (Maidens) has replicas of
the original statues while the mouldings on the entablature and
the podium are mix of original and replicas.
Figure 1. The Erechtheion today (late 2007)
Figure 2. Missing architectural elements
1.1 Project Outline and the Main Requirements
This work is part of the project “Development of Geographic
Information Systems at the Acropolis of Athens“ (Moullou &
Mavromati, 2007). The goal of the project is the development
of GIS with the associated databases for documentation,
restoration management, and presentation of the monuments
and surrounding walls and landscape, starting with the
Erechtheion, which is so far the only fully restored monument
on the Acropolis. This requires the creation of a richly detailed
accurate 3D model. The resolution and visual quality of the
rendered model should ideally match what is perceptible by the
human eyes on a real visit, preferably when at close up range. A
movie with photo-realistic colour and lighting is also required.
To capture details needed for documentation and restoration,
lateral data spacing (spatial resolution) of the 3D model should
be about 5 mm in most parts and 1-2 mm on the highly detailed
parts. Local depth uncertainty and overall accuracy must be 1
mm and 10 mm, respectively. For interactive visualisation the
3D model must be viewable on a standard workstation. On such
large complex structure, it is very challenging to achieve all the
requirements throughout the modelling pipeline. The procedure
and techniques we adopted will be discussed next.
1.2 The Implemented Approach
Range sensors, such as laser scanners, can provide highly
detailed accurate representation of most shapes (Blais 2004).
Combined with colour information, either from the sensor itself
or from a digital camera, a realistic-looking model can be
created. On the other hand most scanners can be bulky, which is
a disadvantage on difficult terrain. The results are influenced by
surface light scattering and absorption properties. Since marbles
are made of translucent crystals, they are problematic (Godin et
al, 2001). One must also be careful in selecting a range sensor
for a project as they are intended for a specific range, thus one
designed for close range may not be suitable for medium or