C er-A
Cyb
Arch
haeologyy
E
Edited
by
b
M rizio Fortee
Maur
B
BAR
Innternaationall Seriees 21777
2010
Published by
Archaeopress
Publishers of British Archaeological Reports
Gordon House
276 Banbury Road
Oxford OX2 7ED
England
bar@archaeopress.com
www.archaeopress.com
BAR S2177
Cyber-Archaeology
© Archaeopress and the individual authors 2010
ISBN 978 1 4073 0721 3
Printed in England by 4edge, Hockley
All BAR titles are available from:
Hadrian Books Ltd
122 Banbury Road
Oxford
OX2 7BP
England
www.hadrianbooks.co.uk
The current BAR catalogue with details of all titles in print, prices and means of payment is available
free from Hadrian Books or may be downloaded from www.archaeopress.com
On-Site Digital Archaeology 3.0 and Cyber-Archaeology:
Into the Future of the Past – New Developments, Delivery and the
Creation of a Data Avalanche
Thomas E. Levy, Vid Petrovic, Thomas Wypych, Aaron Gidding, Kyle Knabb, David
Hernandez, Neil G. Smith, Jürgen P. Schlulz, Stephen H. Savage, Falko Kuester, Erez
Ben-Yosef, Connor Buitenhuys, Casey Jane Barrett, Mohammad Najjar,
Thomas DeFanti
University of California, San Diego, Center of Interdisciplinary Science for Art, Architecture, and Archaeology, California
Institute of Telecommunications and Information Technology, U.S.A.
Corresponding author: tlevy@ucsd.edu
Abstract
Over the past 3 years, with the establishment of the new Center of Interdisciplinary Science for Art, Architecture and Archaeology
(CISA3) at UCSD’s California Institute of Telecommunication and Information Technology (Calit2), a collaboratory framework has
been established facilitating joint research between archaeologists, computer scientists and engineers. We report here on a cyberarchaeology field recording system that feeds into a cyberinfrastructure delivered over the Mediterranean Archaeology Network
(MedArchNet) on a Google Earth platform. A field test of the new system was carried out in 2009 at Khirbat en-Nahas (KEN), an Iron
Age (ca. 1200 – 900 BCE) copper production center in Jordan.
Key words: Cyber- Archaeology, LiDAR, Balloon Photography, Artifact Informatics, Cyber-infrastructure
Introduction
In 1999, when we made a commitment to ‘go digital’
to record all our field measurements on excavations in
Jordan related to the role of ancient metallurgy on social
evolution, we had no idea that our 20th c. data would be
‘pre-adapted’ to the growing field of 3D visualization.
This first application may be referred to as on-site digital
archaeology (OSDA) 1.0(Levy et al. 2001). Being based in
far away San Diego with only periodic access to the field
and artifact collections left behind in the Middle East at
the end of each excavation season; we wanted to develop
a recording system that would enable us to take our entire
dataset home with us for analyses. In this case, as Plato
(ca. 427 BC – 347 BC) wrote in The Republic, necessity
was indeed the ‘mother of invention’ and the driving
force for abandoning the old analogue paper recording
system we had used for over 25 years (Levy 1987). Over
the years, as computers have become more portable and
more powerful, OSDA 2.0 emerged, that like Ver. 1.0,
has Geographic Information Systems (GIS) at its nexus
(Levy and Smith 2007) to facilitate the spatial analyses of
archaeological data.
Here we summarize the most important new
developments in OSDA 3.0 that make it a much more
versatile system (Figure 1) that takes advantage of
both off-the-shelf technologies and also includes
new computer programs and hardware developed
specifically to solve archaeological/cultural heritage
problems that face researchers working around the
world today. As a field science, archaeology depends
on precisely documenting the x, y and z coordinates
of excavation and cultural heritage data. By acquiring
this kind of metadata for material culture, it is possible
to thread together an array of different kinds of spatial
and analytical data recorded in the field that ultimately
relates to the larger theoretical and historical questions
that drive interest in world cultural heritage.
OSDA begins with the mapping of all realms of material
culture or any spatial information of relevance to the
archaeological research. For archaeological field survey
work, we depend on differential GPS, whereas excavation
work depends on Total Stations. Both methods have a
high degree of accuracy, typically around 1-2 centimeters.
All artifacts, features and topographic points are stored
in a GIS database that can be quickly and easily accessed
for analysis or transferred to a variety of 3D computer
modeling programs. Many of the other technologies
we use rely on the data taken with the total station.
For example, most of the imaging techniques we use
are geo-referenced using control points taken with the
total station. Balloon photographs, described below, are
imported into ArcGIS and geo-referenced. This allows
us to create detailed site plans of architecture and
other features. For a more detailed description of how
Total Stations and GPS are used in the field readers are
referred to Levy and Smith (2007). Figure 2 shows how
geo-referenced aerial photographs taken with the new
helium balloon system are used for creating publishable
maps at KEN. These data are subsequently used in the 3D
visualization environments discussed below.
135
Cyber-Archaeology
Figure 1. Flow-chart illustrating the On-Site Digital Archaeology 3.0 system with new elements highlighted and
discussed in this paper: LiDAR mapping, helium balloon airborne photography, StarCAVE, NexCAVE, Artifact
Informatics, and cyber-archaeology represented by the Mediterranean Archaeology Network (MedArchNet).
Figure 2. An image captured by the UCSD balloon system is geo-referenced and
the rocks are ‘traced’ to create a detailed architectural plan. Shown here is part
of a building complex from Area W, KEN viewed in ArcMap.
136
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
Making more out of X, Y, and Z: New Imaging Techniques
Used in OSDA 3.0
The new OSDA 3.0 system integrates a number of imaging
techniques including a specially designed helium balloon
platform for taking vertical and oblique photographs
of archaeological/cultural heritage sites, GigaPan
photography that enables multi-gigapixel panoramas to
be taken of sites with DSLR cameras, and terrestrial LiDAR.
a) Airborne Balloon-Based Imaging
Surface level photography provides a record of excavations,
however to place the site in its more general context for
larger scale analyses, aerial photography is needed. Here
we describe improvements on the ’boom’ system used
in OSDA 1.0 and 2.0. The early method involved placing
a small digital camera on a ca. 7 m long wooden stick,
struggling to lift it and capturing an image covering only a
ca. 5x5 m excavation square (Levy and Smith 2007). This
was time consuming, awkward and involved long hours
of ‘stitching’ geo-referenced photos together to cover
one excavation area. These problems were a catalyst to
design our airborne helium balloon system in the summer
of 2009 by UCSD CISA3/Calit2 undergraduate students
that was deployed in Jordan in September of that year. It
consists of a helium-filled balloon with a sail appendage
mounted on a stable aluminum platform equipped with
two 15 megapixel digital SLR cameras monitored with a
live-feed (Figure 3). The balloon (Kingfisher™ Aerostat
from Southern Balloon Works) is tethered to an operator
on the ground that can position it over a specified
excavation and cover areas up to 700 m2. The system
can capture images up to a height of ca. 200 meters at
high resolution and excellent stability that insure image
quality. The system is versatile enough that it can be
brought down to a lower height for even higher resolution
image capturing.
The vertical high-resolution images of site architectural
features are geo-referenced and aid in the creation of
publication quality maps using ArcMAP in the ArcGIS 9
suite. As all photographs are shot as stereo-pairs that
canbe used to create digital elevation models (DEMs).
For oblique views, the cameras can be manually rigged to
supplement more interpretive analyses.
Another requirement for this system was quick
deployment on a daily basis. Thus complicated rig systems
were avoided in the design of the aluminum platform.
Figure 3. Top left: Helium balloon and photography platform in-flight; Top right: View of
monumental Iron Age building at Khirbat en-Nahas (excavation area shown, ca. 15 x 25 m);
Bottom: Image composite of KEN (ca. 10 hectares; square fortress = ca. 73 x 73 m). Shot from
altitude of ca. 200 meters.
137
Cyber-Archaeology
Figure 4. GigaPan 360o view of Khirbat en-Nahas, Jordan.
This imagery will be used in various 3D visualization environments.
The balloon was kept inflated throughout the 2-month
long expedition season with little helium loss. It was
sheltered each night in an on-site tent-like ‘hanger’. In
the 1930s, the Megiddo project similarly had to construct
an on-site ‘hanger’ for their photography balloon – one
of the earliest such applications (Guy 1932). OSDA 3.0
cameras were positioned on the balloon chasse system
at the beginning of each day, allowing it to be ready for
shooting within ten minutes.
Rapid recoverability of the balloon is possible by being
tethered to the operator with a high strength, lightweight,
Spectra line. This allows the operator to position the
balloon over a specified area for shooting and recove
it using the tether. Operation iss simple - the balloon is
‘walked’ until positioned over the area of interest. A
live-feed from the air-borne cameras to a laptop on the
ground allows a visual check to determine if the system is
over the target.
Desert conditions required that the balloon system be
r ugged, thus, there are few electronic components.
The balloon itself was rugged and able to withstand
UV radiation during its long deployment. Over 13,000
images were successfully captured amounting to over
200 Gigabytes of data during the 8-week season (Table
1). We are currently designing a more rugged aluminum
platform, remote control to facilitate automated camera
movement for both vertical and oblique photography, and
stereo-video photography to enable 3D fly-through and
fly-over of archaeological and cultural heritage sites.
b) GigaPan
GigaPanTM enabled us to use our Digital Single Lens
Reflex (DSLR) cameras to produce multi-giga-pixel images
at KEN not only to capture high-resolution panoramic
views of the site (Figures 4,5) but also as a supplemental
research tool. To enhance data acquisition and analysis,
GigaPan photographs can be overlaid on the point
clouds of archaeological features acquired digitally by
LiDAR. Individual excavation areas were imaged using
the GigaPan (Areas M, R, T, and W) and will then begeoreferenced with the billion-point LiDAR scans made with
the LiDAR scanner.
The GigaPan imagery supplements the LiDAR data
by providing more accurate, higher-resolution color
information than delivered by LiDAR scanning. The
process enhances the positional accuracy for the highdensity three-dimensional LiDAR point clouds with color
and texture information of comparable fidelity. The
combined record is both photographically representative
and three dimensionally precise down to the millimeter.
Figure 5. GigaPan view (180o) of the interior of a room in
the Area W Iron Age building complex, KEN will used to
supplement areas not covered by terrestrial LiDAR scans
in the field.
c) LiDAR for Archaeological Field Work vs. Cultural
Heritage Conservation
Recent developments in on-site scanning technologies add
an important new field tool to the OSDA 3.0. Terrestrial
LiDAR scanning was introduced in 2009 at KEN to augment
recording, analysis and conservation efforts. The scanning
was carried out with a Leica ScanStation 2™ that uses
laser light to capture a collection of 3D points sampling
the geometry and color of objects within its field of view.
We acquired over 1.75 billion points in space comprising a
high-resolution spatial record of the ancient fortress walls,
gatehouse, residences and some 100 ancient unexcavated
buildings visible on the site surface. The application of
LiDAR scanning for documenting on-going excavations
is relatively new: To date, most archaeology LiDAR
applications have focused on recording sites as a means
of ancient monument conservation (Barton 2009) and as
a reconnaissance tool (McCoy and Ladefoged 2009). In
the UK, relatively high-resolution LiDAR (ca. 1 – 2m) was
used for site prospecting that produced results as good
or better than aerial photography (Bewley, Crutchley,
and Shell 2005). Our goal is to use LiDAR as a heuristic
device to investigate archaeological data collected during
the course of excavation as well as provide an accurate
conservation record of a site.
i) LiDAR On-Site – Recording a 10 hectare archaeology site
with sub-centimeter accuracy
The application in OSDA 3.0 application in Jordan
exploited the full potential of sub-centimeter accuracy
provided by LiDAR scanning for site recording and
analyses. First, points were sampled every centimeter
(compared to the 5-centimeter resolution of previous
scans our team carried out in the Anza Borrego desert,
138
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
California, in preparation for the Jordan project – see
Fox 2008a). LiDAR scanning at KEN produced ever-moredetailed scans, determining points approximately one
millimeter apart. The LiDAR scans also yielded color and
intensity information for each point, providing some
insights about the properties of the material.
Successfully scanning a large archaeological site using
terrestrial LiDAR technology poses a set of unique
challenges, including proper scanner and target setup in
and around sensitive archaeological artifacts dispersed
over a site that in this case extends over 10 hectares. At
KEN, line-of-sight distance from the scanner was limited,
with many surfaces requiring scans from oblique angles
adversely effecting the achievable range, speed and scan
resolution. These challenges were further aggravated
by electrical power requirements imposed by extended
acquisition runs as well as environmental conditions such
as dust and heat that impacted both the scanner and the
control laptop. Finally, solid boundary conditions for highprecision geo-referencing of the acquired data had to be
provided.
At least half a dozen LiDAR-based surveying devices
using either time-of-flight or phase-based techniques (to
determine the distance between the scanner and a target
within line-of-sight of the scanner) are now commercially
available. Most of these devices support a spherical scan
envelope that is swept out one point at a time during the
acquisition process. Most commonly, the laser is directed
via a rotating or pivoting mirror, covering approximately
270° vertically (the remainder is commonly occluded
by the system’s tripod) and a rotating head allowing for
360° horizontal sweeps. Common sampling rates range
between 15,000 and 150,000 points per second with a
sample spacing that may be as small as 1mm for systems
with an average laser spot size of approximately 4mm.
can be used to align data collected from multiple scan
positions. To drive the scanner, power is required as well
as a standard Ethernet connection to a control terminal
(limited by specification to ca. 100 meters). Electrically,
the world-compatible power supply draws up to 400W
for the scanner, although in practice the scanner averages
100W power, plus 60W average for the control laptop.
Field Deployment
A scanning sampling strategy is needed when tackling
a large site as not all areas are of equal interest. The
foremost concern when deploying the scanner is its
placement relative to the object or panorama of interest.
Uniformity of coverage is of utmost importance not all
positions are visible from the scanner’s head location
can be represented resulting in voids in the final scan.
Careful planning to cover a particular feature or surface
from multiple angles is required to capture surface
characteristics that would otherwise be represented
as shadows with insufficient coverage. This involves a
‘walk-through’ with the excavation director to identify
significant areas for detailed scanning.
The main excavated areas on the site were scanned
at very high resolutions (1mm or 2mm). These high
resolution scans have provided an impetus for developing
new processing and visualization techniques capable of
handling these large datasets. For key excavation areas,
such as the fortress gatehouse, the following scanning
goals were met:
1. Provide a clear and unobstructed view of features
of interest.
2. Select locations where scanner is equidistant to
features of most interest.
3. Provide multi-angulecoverage of non-planar
surfaces – commonly, several scans were needed
to avoid occlusions. In practice, this required
changes of scanner elevation as well as position.
4. Solid geo-referencing support with limited
interference from surveying targets (markers).
5. Limiting required scan distance to control errors
introduced by variations in temperature and
humidity.
6. Minimize the number of scan positions needed to
achieve target objectives in respect to coverage
and resolution.
System Specification
The Leica ScanStation 2TM used at KEN is theoretically
capable of sampling rates close to 50,000 points per
second and acquisition densities of under 1mm2 over
distances of up-to 300 meters for objects with a 90%
albedo. Under Jordan field conditions sample distances
of 100 meters tend to be more realistic. The scanner
weighs approximately 19 kg, the battery pack 12kg and
its surveyor’s tripod 9kg for a total of 40kg for the base
system, excluding the acquisition laptop. When combined
with a ruggedized transport case the overall weight more
than doubles that requires an experienced operator and
one assistant. The ScanStation2 has a maximal scan range
of -45° to +45° vertically, requiring a two-pass sweep
and 360° horizontally. User specific scan windows can be
flexibly defined within this. In specifying a solid angle of
interest, a physical resolution is derived using an average
distance to target. With these two parameters in place,
the physical resolution is computed in terms of samples
per degree – in effect creating a variable spatial resolution
based on the varying distance from scan surface to scanner.
In the course of each scan, stationary surveyors’ markers
are scanned and enumerated as a spatial reference that
In practice, handling ancient rooms with interesting
features is relatively straightforward: the LiDAR
technician can “walk” the scan lines, observe occlusions
and make concessions of coverage between prospective
scan positions. A digital camera and diligent analysis of
occluded spatial regions also helps in planning. Constraints
such as the suggested minimum distance from corners
and surfaces helped to converge prospective scanner
placement between multiple candidate sites.
To achieve greatest wide-area coverage of the site, a
number of specialized scans directed exclusively at lowresolution panoramic coverage from an elevated position
139
Cyber-Archaeology
are best. Covering wide areas uniformly is best achieved
by choosing a set of physically elevated candidate
sites with the greatest visual coverage of the site. Colocating these wide area scan sites with areas of interest
For background scans, resolutions of 1cm at 50m are
considered acceptable.
Empirically, the time flow for a scanner placement is
estimated as follows:
•
•
•
•
•
•
5 minutes tripod deployment
5-10 minutes cabling and scanner set-up
8 minutes to boot up scanner
10 minutes to acquire background photography
5 minutes to acquire surveyor’s markers
60 minutes for each 30 million points to be scanned
(in practice)
Based on these parameters, it was important to minimize
the number of scanner placements. For example,
interior scans inside a building can be around 30 million
points. Thus, the deployment cost noted here governs
the time spent acquiring the scan. The only limitation
for prospective scanner placement is direct line-of-site
visibility to at least two separate surveyor’s markers and
those markers should be clearly visible and within 100
meters.
Geo-Referencing
Included in the scanning field deployment are four
surveyor’s markers. These markers are placed strategically
to be seen from the majority of the scan locations, so that
scan location can see at least two markers. The purpose
of registering placement relative to the markers is so that
several scans from different physical positions can be
stitched together as one uniform scan. These stationary
markers serve as the ideal reference points for chaining
together a collection of scans.
With careful placement, four markers are sufficient for all
but the most complicated and obscured areas of interest.
Complications arise when multiple sites of interest are to
be scanned and co-registered when one of two conditions
is encountered:
1. Four target sites are insufficient to cover both
areas of interest
2. The areas of interest are separated by more than
100 meters
In the course of the scanning campaign at KEN, both of
these conditions occurred. To avoid the propagation of
error likely to occur if targets were moved, they were
never moved until all scanning was complete. Likewise,
revisiting a site using a distant marker as a common
registration link was ill-advised as error could easily be
propagated in false or noisy measurements to the distant
marker. To solve both problems, a strategy was established
of systematically moving the scanner along the scan path
in one direction through the site – thus containing error
propagation by avoiding re-visitation of scanned areas.
Trying to manage the tradeoff and process of marker
placement, scanner placement, context of scans
and ordering of on-site scan work-flow proved to be
formidable. In practice, the scan site and all regions of
interest were thoroughly surveyed visually for an entire
day before any equipment was deployed.
ii) Post-Excavation LiDAR Processing
The product of a scanning campaign is a collection of raw
point clouds, one per scanner setup. These individual
point clouds subsequently have to be merged, cleaned
and geo-referenced (assigned latitude, longitude, and
elevation), turning the data collection into an accurate
representation of the field-site and excavation, suitable
for visualization, analysis and co-registration with other
digital data assets.
The scanning campaign at KEN yielded some 1.75 billion
scanned points that were acquired following the scanning
procedure outlined above. The process ensured that
several GPS-referenced markers were visible in each cloud,
establishing a common reference coordinate system
for the entire collection and insured that scans could
be accurately merged. Once merged, the overall data
collection is cleaned to remove redundant, undesirable
or extraneous points, for example, those of inadvertently
scanned persons or equipment. Of particular importance
is the treatment of points covering objects visible in
multiple clouds, since not all objects are scanned equally
well in every cloud: some are near the scanner, with high
point density and precision, some are far away, or at an
awkward angle with low density and precision. Consider
a part of a wall that is visible in multiple clouds, with a
differing scan quality in each. If the wall is covered well
enough by points from a single ‘best’ cloud, the overall
quality of the dataset can be improved by removing the
lesser-quality wall points from the other clouds. The
size of the resulting datasets have historically presented
a major obstacle: just bringing the points to the screen
quickly is a technical challenge. To address this, we have
developed a novel system that allows billions of points
to be interactively visualized on commodity hardware—
however, the algorithms involved are not presented here.
Once the overall point cloud is available, a baseline
record for the site is established that captures its spatial
characteristics at a specific moment in time. With this
record in place it is then possible to virtually explore the
site, flexibly and freely, overcoming on-site limitations
in respect to how easily particular locations may be
accessible. When paired with intuitive interaction and
display devices, visualization of this point data allows for
the exploration of spatial relationships, correlations and
measurement of the site and its artifacts (Figures 6-7).
The use of room-sized display walls such as HIPerSpace,
and virtual reality environments such as the NexCAVE
and StarCAVE (e.g., Figures 21-22), provides a means to
140
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
collaboratively study a vast amount of geo-referenced
data of various types within an appropriate topographic
context. At the same time, laptop-centric visualization
provides a means to analyze data in the field, at a level of
interactivity previously only seen in videogames. This may
include analysis of the locations of artifacts, correlation
of radiocarbon samples, and augmentation of close-up
photographic records or references in the geo-referenced
context provided by the point cloud.
We have also experimented with more structured
analytical paradigms. One example is a grid system
that partitions the site into fixed-size cells. The grid
provides an organizational structure for annotations
and measurements, and enables a constrained cell-bycell navigation mode that complements the freeform 3D
inspection interfaces. In addition to letting us view the
site as a camera would—in perspective—the visualization
system can also be used to cut out and draw arbitrary
slabs and slices of the site orthographically, overlaid with
a measurement grid and the real-world scale. This mode
allows us to perform both qualitative and quantitative
analyses. For example, we can quickly measure length
visually by positioning and scaling the viewing slab
appropriately. We can also easily obtain ground plans
(Figure 8) or vertical sections (Figure 9) by positioning the
viewing slab parallel or perpendicular to the ground, and
selecting the appropriate slab thickness. Note that these
types of analyses are made possible by the performance
and interactivity of the underlying visualization system—
the subject of interest is interactively manipulated into
the desired position on the measurement grid, with
continuous visual feedback, simply by moving through the
virtual environment.
Figure 6 Viewed from above, a visualization of the LiDAR record resembles an
aerial photograph. Note the differing levels of scan coverage corresponding to
areas of greater and lesser interest: areas with minimal coverage appear as gaps
or shadows between the more important regions scanned with a high point
density. Shown here is the Iron Age fortress (ca. 73 x 73 m) in Jordan. Compare
with Figure 3 (Image: CISA3/Calit2).
Figure 7. Viewing an area of interest up close reveals the detail acquired by the
LiDAR scanning technique. A structure with interior rooms can be seen in full
detail in the foreground, with sparser coverage of other areas and the background.
Shown here is the gatehouse of the fortress at KEN (Image: CISA3/Calit2).
141
Cyber-Archaeology
Figure 8. The point visualization tool can be used to
present spatial relationships faithfully and to scale. The
point cloud data is a precise record of the scanned area or
objects, and can be used to perform measurements and
other analyses. For example, we can obtain a floor plan of
the Iron Age gatehouse with minimal effort by selecting
a horizontal slice of points to view (Image: CISA3/Calit2).
Portable analytical tools – XRF in the field (EB-Y, TEL)
Portable high precision analytical tools are rapidly allowing
researchers to bring the geoarchaeology laboratory to
the field (Katz et al 2010).
In recent years, the field
of x-ray fluorescence (XRF) has been revolutionized with
the development of portable devices. Instead of bringing
samples to the laboratory, the researcher can now
measure the bulk chemical composition of materials in the
field and get results on the spot. Pioneering applications
of this type of research have been carried out on early
metallurgy sites in Israel (Yekutieli et al 2005; Vardi et
al 2008) The portable XRF has various applications and
more and more publications demonstrate the important
role of integrated chemical analysis in field research.
For example, archaeologists can measure and map soil
composition (metalliferous pollution or other chemical
variables of interest) spatially on site or along profiles in
excavated sections, correctly identify objects already in
the preliminary stage of research (bronze vs. copper vs.
iron artifacts, gem stones, beads etc.), obtain typological
‘chemical signatures’ of artifacts as part of the sorting
and cataloguing process in the field, etc. One of the
fundamental advantages of portable device is the ability
to measure materials that cannot be carried to the
laboratory due to governmental rules or the large size of
some cultural material.
The output information of the portable XRF is a complete
list of chemical elements found in the sample, with their
relative or absolute quantities (depending on instrument
calibration and type of material). The measurement
itself is non-destructive, rather quick (usually up to 300
seconds), and in most cases without the need of special
sample preparation. The fast measurement process
results in an exponential growth of digital data that
should be organized and correctly linked to the sample
information and excavation contexts. A major caveat of
XRF application is that each measurement represents
only a limited portion of the sample on the scale of a few
millimeters (in diameter and in depth, depending on the
specific structure of the device, its settings and the type
Figure 9. Alternatively, structures can be isolated and compared, such as the above common wall with obstructions
removed. Note the automatic grid-line and scale overlay; the grid spacing and scale legend updates dynamically as
the view is changed by the researcher. Here a north-face section has been made through the LiDAR data to illustrate a
section illustrating two walls that delineate the doorways to two guard rooms (Image: CISA3/Calit2).
142
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
of material). This should be taken into account in the case
of heterogeneous, corroded or patinated samples, and
such metadata regarding sample characteristics must
complement the XRF data.
Portal Science, Cyber-Infrastructure and
Cyber-Archaeology
At KEN, we integrated data from routine XRF
measurements with the master GIS-based database of
the excavation. More than 600 artifacts were measured
with a portable Bruker XRF device, including soil samples,
scarabs, copper ore, ceramics, slag and metal artifacts.
The field measurements helped, for example, to identify
high impurities of iron in the raw product of primary
smelting and the lack of tin in most of the copper objects
from the site; we also obtained elemental composition
of hundreds of artifacts that had to stay back in Jordan.
One such artifact was an arrowhead (Figure 10), that was
found to be made of pure iron (now corroded, Figure 11).
The XRF reading of this artifact (.PDZ file format in our
system) is linked to its EDM number , and by it to its spatial
coordinates, context (locus, basket), digital photography
and other information included in the relevant Access
database. The XRF database is currently being integrated
into the visual analytics methodology described in this
paper.
To share ancient settlement pattern and other
archaeological data with as large an audience as possible,
a cyber-infrastructure is needed that can promote the
sharing and analyses of data in a communal manner. The
NSF GEON “Cyberinfrastructure for the Geosciences”
project (http://www.geongrid.org ), whose Information
Technology component is led by Chaitan Baru, serves
as model for how the Earth Science community uses
a cyberinfrastructure with “data portals” to facilitate
delivery, discovery, access, and integration of distributed
heterogeneous data sets (Baru in press). Our group
has built a similar cyberinfrastructure to unify the many
digital datasets and methods described in this paper. This
‘portal science’ application is called the Mediterranean
Archaeological Network (MedArchNet, Figure 12) and is
envisioned as a series of linked archaeological information
or atlas nodes, each of which contains a regional database
of archaeological sites that share a common database
structure in order to facilitate rapid query and information
retrieval and display within and across nodes in the
network. To date, one digital archaeology atlas node is
fully functional and utilized by hundreds of researchers.
MedArchNet is a signature project of UCSD’s CISA3/Calit2
and the Geo-Archaeological Information Applications
(GAIA) Lab, Archaeological Research Institute at Arizona
State University. The ultimate vision of MedArchNet
is to develop a network of archaeological sites (from
remote prehistory to the early 20th century – see http://
medarchnet.orgMedArchNet currently contains one
active archaeological information nodes -the Digital
Archaeological Atlas of the Holy Land (DAAHL) at http://
daahl.ucsd.edu
Figure 10. Arrowhead from Area W, a weapon made from
iron, 2009 UCSD Jordan expedition (EDM# w09f2918;
Photo: A. Gidding, UCSD Levantine Archaeology Lab).
Fig.ure 11. XRF reading of the arrowhead (Figure 20)
in initial examination mode showing that the artifact
contains almost exclusively iron (Fe).
a) MedArchNet – DAAHL
The MedArchNet cyberinfrastructure provides secure and
reliable storage of data from the field to the central data
storage facility. It will provide authenticated, portal-based
access to data, derived products, analysis, visualization,
and GIS tools, collaboration spaces, etc, including
provision of “publish/subscribe” interfaces for data, to
enable a large user community to gain access to data and
derived products. The cyberinfrastructure will manage
heterogeneous archaeological data from a variety of
sources, and support a community of contributors as
well as users of the information, using a comprehensive
authentication and authorization system to control access
privileges of different classes of users.
The MedArchNet approach to archaeological site data
envisions our data nodes as “switchboards” that contain
top-level site and project metadata, plus bibliographic
references and extensive use of linked resources outside
the MedArchNet data structure. It is not our goal to corral
every bit of data about every site in the Mediterranean—
an enterprise that would clearly be impossible even if
it were desirable. Rather, the MedArchNet approach
is designed to let researchers and the public easily find
143
Cyber-Archaeology
Figure 12. The MedArchNet website which highlights the most active node – Digital
Archaeology Atlas of the Holy Land (http://medarchnet.calit2.net ).
archaeological sites based on location and other attributes
such as site type, features, time periods, etc., provide a
mechanism for creating substantive maps linked to the
various MedArchNet nodes, and then point the user to
the locations of substantive research on the site, whether
it be on- or off-line. The MedArchNet project serves to
highlight the research of the archaeological community,
rather than subsume it under the MedArchNet umbrella.
Each data node maintains a table of data donors, including
contact information and primary web sites, and each site
contributed by a donor will be “branded” with the donor’s
information. Whenever a contributed site is displayed,
the donor information is also shown, so the links to the
donor’s website are clearly shown, along with specific
external resources for individual sites.
The MedArchNet project actively cooperates with
research organizations and government agencies to
develop new data nodes and applications. Each of our
current nodes has received significant sponsorship. The
Digital Archaeological Atlas of the Holy Land (DAAHL;
Figure 13) is a sponsored project of the American Schools
of Oriental Research, the flagship organization that
coordinates North American archaeological research
in the Levant (http://www.asor.org ). As MedArchNet
develops additional data nodes, we look forward to
expanding our cooperative efforts with additional data
donors, research organizations and government agencies.
The MedArchNet databases are UTF-8 encoded, so they
support multinational character sets; moreover, the
Google Translation tool is included at the bottom of every
MedArchNet web page, so the output can be translated
into any available language at the touch of a button.
Figure 13. DAAHL’s On-line Virtual Museum georeferences artifacts collected and recorded during
excavations and displays them over a GoogleMaps
platform. Shown here is a 9th c. BCE pottery sherd found
at KEN hovering over a GoogleEarth image.
The DAAHLsite illustrates some of the content-rich
methods it uses to disseminate data drawn from its
database. Efforts to harvest archaeological site data from
Israel and Palestine are currently underway. Another
highly innovative feature is the DAAHL’s Virtual Museum
(Figure 13), which displays interactive, 3D objects at their
original find locations through a Google Earth API—the
user can manipulate the object in all three dimensions as
well as the map itself.
144
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
MedArchNet is already having a significant research and
education impact by providing easy online access to
archaeological data and information. The MedArchNet
hub and its data nodes are deployed to provide access
to information contributed by each member of the
MedArchNet “Virtual Organization”. Via the hub, users
are able to navigate back to the original member sites and
databases to access the full information and related data
from the respective site. Each data donor receives full
recognition and credit for their contribution. Individual
portals have been developed initially for ASOR (DAAHL)
and other partnering groups. The network of linked
portals support collaborations among users and provide
a platform for initiating and sustaining discussions related
to cross-site thematic areas of study.
In an era of rapidly expanding population and urban
development, a system like MedArchNet can provide
mechanisms to monitor archaeological site conditions
over time and lessen the impact on cultural heritage
resources by careful planning and significantly enhance
site preservation and development potential in
the Mediterranean basin. Furthermore, by uniting
archaeological site metadata from many disparate datasets
and organizations, the MedArchNet cyberinfrastructure
will dramatically improve the ability of researchers to ask
large-scale, cross-border questions of the archaeological
data, providing fresh new insights into some of the most
culturally meaningful regions on Earth.
Figure 14. Undergraduate Caity Connolly uses
NextEngine™ 3D scanner to image small Iron Age ceramic
vessel in CISA3/Calit2 cyber-archaeology lab. This
instrument was taken into the field to Jordan in fall, 2009
(Photo: Eric Jepsen, UCSD/Calit2).
b) 3D-Artifact Scanning
For much of its history, post-excavation archaeological
research has relied upon manual drawings of artifacts
that are often time-consuming and expensive to produce,
subjective in detail and limited in their two-dimensional
scope. Often, those drawings depict mere fragments
of artifacts, making it difficult for archaeologists to
recreate complete objects. To create the necessary 3-D
algorithms that reflect the chronological and cultural
‘address’ of ancient potters in Jordan, our team uses
NextEngine™ laser scanners to obtain triangulated
meshes of the potsherds which can later be converted
into any 3-D vector format (2-D images generate raster,
or dot-based images, but vector, or shape-based images,
are more amenable to mathematical manipulation). Since
all the 3-D scans done by our team can also be imported
into MATLAB (a numerical computing environment and
programming language), we can compare and analyze, in
the same format, both the 3-D scans and 2-D vectors of
images taken from archaeological publications (see Foxb
2008).
The portable and relatively inexpensive NextEngine 3D
scanners are“field operable” units.. They consist of two
components - a turntable and a data capture device. The
turntable allows for the object scanned to rotate 360o
in front of the data capture device at specific intervals
allowing for the scans to stitch together more accurately
The data capture device consists of two primary
Figure 15. 3D scan of Iron Age ceramic cup from Area R,
KEN. The 3D image file is used in the Pottery Informatics
program described here (Image: CISA3/Calit2).
components, a laser array and a digital 3mp camera in
order to capture texture data. This combination allows us
to create photorealistic models of the artifacts as portable
as the computers that we used to scan the artifacts
(Figures 14-15).
In 2009 we took the 3D scanners to the field for the first
time. This wasan experiment to test how they would
cope with the harsh desert conditions experienced
within our field lab. In the CISA3 lab there is total climate,
environment and light control that allow us to ensure that
the artifacts are in the optimal conditions for high levels
of scan accuracy (Guidi 2008). The problem in Jordan was
how to adjust to far more challenging conditions while
in the field. The ambient temperature was a particular
problem, restricting us to be able to scan mostly in the
early morning when the day was coolest. There was
also the issue of the effect of ambient light affecting the
accuracy of the scans. To solve this problem we simple
built an inexpensive enclosure out of cardboard to create
the most controlled environment possible in our field lab.
With these minor adjustments we were able to achieve
quality resultant scans on the field.
145
Cyber-Archaeology
An area we are currently investigating is how to further
streamline the 3D scanning process. One method to
circumvent the unavoidable time expense required for 3D
scanning was to purchase several scanners thus reducing
scan time by a third. We are also trying to automate
a number of the processes so that less student time
is required to scan and process these data. In general,
now that we have begun 3D scanning of artifacts, the
NextEngine scanner has become a standard tool in our
archaeological ‘digital tool box.’
c) Artifact Informatics: Pottery Informatics Queryable
Database – PIQD
Beginning in 2008 we began to develop an informatics
database for the processing of diagnostic Iron Age
(ca. 1200 – 500 BCE) pottery sherds. The goal of this
project was to develop a method of digitizing artifacts
collected in the field within an analytical framework
similar to that used in bioinformatics for the analysis of
protein and DNA sequences (Altschul et al. 1990; see
below). The increased precision achieved through these
analyses exposed a number of drawbacks to the use of
only digitized 2D illustrations of ceramics. One of the
most prevalent problems being the relative subjectivity
of the professional illustrators attempt to draw a 2D
profile representative of the whole sherd, its stance, and
measured diameter. In order to achieve the most accurate
representation of the diagnostic pottery sherds recovered
from excavation, we started in this season to scan pottery
using the NextEngine™ 3D scanner (see Figures 14-15),
which allows us to confidently attain accuracy within 0.12
mm for each sherd scanned. Once the sherd is scanned
it is imported in Matlab where its proper stance, rim
diameter, and profile are extracted at this same level of
precision. In addition, we used the 3D scanner to scan
other artifacts that are even more geometrically complex.
Although there has been increasing interest among
archaeological projects aimed at digitizing various
aspects of their archaeological datasets, there has been
no satisfactory solution for integrating these different
projects’ data for cross-regional comparison and analyses.
Most researchers still must rely upon printed publication
reports to conduct any form of regional study. In general,
this medium is very limited in its ability to inform the
reader of the nuances of the material culture collected at
the site, such as architecture, stratigraphy, ceramic, and
other artifact assemblages.
In order to circumvent a number of these drawbacks, a
comprehensive online queryable digital database called
the Pottery Informatics Query Database (PIQD) was started
for the Iron Age ceramic assemblages of the southern
Levant (http://daahl.ucsd.edu/PIQD/PotteryInformatics.
php ). This project was begun in 2008 between Smith and
Levy at UCSD CISA3/Calit2 and Avshalom Karasik and Uzy
Smilansky at the Weizmann Institute of Science-Hebrew
University in Israel. It is a new online tool designed to
enable researchers to test their own interpretations and
models against the ever-expanding digital medium of
ceramic datasets in ways that conventional print data
cannot provide. Where the PIQD differs from other online
archaeological databases that may archive published 2D
vectorized images or 3D models of ceramics, is in queryability – in particular, with objective mathematically based
algorithms of artifact (ceramic) profiles. This project uses
recent technological advances developed by Karasik and
Figure 16. Mathematically extracted 2D profile of 3D scanned bowl from the
Iron Age site of Khirbat am-Malayqtah, Jordan and creation of a pottery drawing
suitable for publication .
146
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
Figure 17. The Pottery Informatics Query Database GUI Page
Smilansky (Karasik 2008) to mathematically encode and
store the ceramic profile data as complex algorithms
(Figure 16). Three mathematical representation functions
(radius, tangent and curvature) measure various scales
of differences in vessel form, stance, and rim diameter,
which can be used to determine in an objective manner
the statistical difference between ceramic shapes. This
technique combined with several methods of cluster
and discriminate analysis has been used to construct
objective mathematically based typologies and ceramic
prototypes. Specifically for the PIQD, the three functions
enable the rapid search of a whole database of digitally
stored vessels in an objective mathematically grounded
approach. In this sense, these queries are similar to online
BLAST searches (Altschul et al. 1990) developed in the
field of genetics in being able to rapidly associate large
quantities of digital vessel profiles to each other based on
similar morphological traits.
The PIQD has been designed using MySQL, PHP, Ajax,
and Javascript with an imbedded Google Maps API to
provide a fully queryable spatial environment for the user.
In essence, the PIQD is an open source GIS. GoogleEarth
and Google Maps function as a real-time spatial display
for all the ceramics’ coordinate information. MySQL
functions as the server database to organize all the stored
ceramics profiles, curvature functions, 3D scans, and
metadata, while PHP and Javascript are used to query the
database.
Currently the PIQD is being designed to run autonomously
in its incorporation of new vessel profiles and the
computation of its mathematically based typology query
system. Users are able to directly upload their raw data
and have the PIQD properly store it, insert the metadata
into tables, and run the needed cluster analyses. A set of
automatic error checking mechanisms were developed
to insure that only error-free data are accepted into the
PIQD for study. By this means, the PIQD is able to expand
exponentially as multiple researchers can contribute their
ceramic assemblage data to the overall PIQD without
the need for a technician to oversee every new entry.
At present, there are over 10,000 ceramic figures and
their associated metadata from Iron Age Edom that have
been incorporated into the database (Figure 17). The
immediate goal is to achieve complete coverage of the
Iron Age for all of the Southern Levant, which will reach
into the 100,000’s.
Finally, two daughter programs were also developed
to further facilitate the original contributions of
archaeologists and their publication. These programs
plug directly into the PIQD. First, the PlateMaker program
enables publication quality plates of queried ceramics to
be auto-generated and manipulated on the fly (Figure 18).
Tables are dynamically linked to the displayed figures on
the plate so that either the reordering of the table or the
plate always remains synced. Second, the MasterTable
is a dynamic spreadsheet that can expand or contract
hundreds of fields and rows of data stored within the
PIQD based on a simple user interface. It can immediately
update changes made to the data through direct user input
or barcode scanners. Finally, a series of Archaeological
functions were developed to further facilitate common
but complex manipulations of archaeological data sets.
These functions can be accessed for any future daughter
programs related to archaeology.
147
Cyber-Archaeology
Figure 18. The PlateMaker: A daughter program of the PIQD designed for the rapid deployment
of publication quality ceramic plates
The PIQD is not limited to only the Iron Age but can be
adapted for any ceramic period. For example, we are also
developing a mirror of the Iron Age pottery informatics
database for the Early Bronze Age in the Southern Levant.
In a period such as the Early Bronze Age, where variation
in the pottery is not especially significant through time
(Dever 1973), this digital technology can provide an
important new way to make inferences about the material
within the site itself.
The PIQD has the potential for
revolutionizing how ceramic assemblages are analyzed,
typologies constructed, and how regional analyses are
conducted. As the PIQD grows the dependence on the
printed medium will subside as equivalent data can
be found on the PIQD but by a much more rapid and
accessible manner. The PIQD, itself being an online tool
part of the MedArchNet, means that it can be utilized
for data analyses not only in the Southern Levant but
wherever a portal has been established. Moreover, the
PIQD is not limited by the Iron Age period or ceramics but
was designed to eventually be adapted for various forms
of material culture and archaeological periods.
Toward a Cyber-Archaeology:
techniques and visualization:
Data
acquisition
Multi-spectral imaging holds great promise for the
transformation of traditional dirt archaeology into its
digitally-enabled form. Visual records in the form of static
images and videos in the visual range provide a baseline,
capturing the spatial and temporal characteristics of a site.
These can be further augmented with thermal images
capturing sub-surface characteristics, photogrammetry
techniques associating physical dimensions with image
data, 3D topological scans yielding high-resolution point
cloud collections via light detection and ranging (LiDAR) or
stereo photography techniques, all of which are backed up
by “traditional” on-site surveying techniques. While it is
still possible to use old fashion dumpy-levels and measuring
tapes, the trend is for research oriented projects to employ
some form of on-site digital archaeology recording of the
excavation process (Bunimovitz and Lederman 2009, Daly
and Evans 2005, Levy 2010, Levy and Smith 2007). The
compelling reasons for the transition to Digital Enabled
Archaeology are accompanied by a set of daunting
challenges associated with the exponential growth of
data that has to be properly recorded, processed, fused,
analyzed, archived and preserved. Some of the required
workflows, visualization and analysis techniques as well
as underlying infrastructure are extensive and illustrated
here.
Visual Analytics and Instruments
a) HIPerSpace – A Visual Analytics Cyber-Collaboratory
The complexity and amount of data that we are confronted
with as a result of the digitally enabled archeology
paradigm, creates unique challenges for the accurate
and efficient analysis of data in formats appropriate
for the tasks at hand. Data records acquired with the
148
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
acquisition tools described here are massive in size and
multidimensional in nature. Moreover, comprehensive
data analysis usually requires simultaneous access to
multiple data sources represented in domain specific as
well as synthesized formats, requiring an environment
that co-locates domain specialists and data assets while
enabling interactive, visual data analysis and reasoning.
To address this challenge, our team has developed the
concept of visualization portals, scalable, high-resolution
tiled display environments, operating at tens to hundreds
of megapixels resolution, while enabling intuitive,
information rich and rapid visual analytics, capitalizing
on multiple human senses to convey higher-dimensional
content. The “gold-standard” for these environments has
been HIPerSpace, a 1/3 gigapixel resolution visualization
environment, with the ability to collocate vast data
collections in real-time, for a broad set of 2D and 3D
formats (http://hiperspace.calit2.net ). HIPerSpace is
powered by a scalable and hardware agnostic visualization
middleware called CGLX (Doerr and Kuester 2010), which
was designed to provide a common visualization platform,
supporting networked, scalable, multi-tile 2D and 3D
visualization environments. With this infrastructure in
place, it has been possible to also transform the traditional
data analysis workspace into its digital equivalent,
where data in the form of images, videos, 3D models,
publications, web references, simulations, etc. can be colocated in one room-sized collaborative digital workspace.
Figure 19shows the co-location of multiple different
artifacts from UCSD excavations in Jordan that can be
manually or automatically sorted, clustered, segmented
and filtered, effectively providing immediate control over
how data is being represented and explored. Figures 19
and 20 show and interactive visualization of a largescale point cloud data set obtained via LiDAR scanning,
which provides an accurate 3D view of a field site and
means to explore the site, topological artifacts and spatial
relationships. HIPerSpace and other environments similar
to it, described below, are a portal into vast archeological
data collections and allow research team to harness both
data and even more importantly the human assets, the
researchers on the forefront of new discoveries.
Figure 20. Interactive visualization of large-scale HD
composite aerial photograph, providing an georeferenced
view of KEN, and means to explore the site, topological
artifacts and spatial relationships (Photo: Falko Kuester,
Calit2) .
The HIPerSpace has proven to be a critical element
for rapid visual comparison of theory with massive
experimental data collections, enabling transdisciplinary
teams of scientists to swiftly validate and comprehend
theory and practice (Figure 19-20). With visualization
as a unifying language anchored in mathematics and
physics, OptIPortals (http://wiki.optiputer.net/optiportal/
index.php/Main_Page ; DeFanti 2008b ) provide a
unique mechanism to communicate information in a
universal format that allows hard domain problems
to be approached by co-located or spatially separated
interdisciplinary research teams.
The StarCAVE
Archaeological data is especially suited for building
scientific visualization paradigms and creating virtual
environments for Cyber Archeology. At Calit2, our group
has focused on four main approaches to help create cyberinfrastructures for archaeology (Knabb, Schultze, and
Levy in press). The areas of archaeological visualization
research center on use of the 3D immersive visualization
environment called the “StarCAVE” (Defanti et al 2008a,
Levy et al 2008 [Figure 21].:
1. The StarCAVE is used to display compelling visual
imagery using data collected in the fieldincludeing
spatial data, digital images, and site reports. This
is an important and efficient way to disseminate
complex information in a manner that is
straightforward and comprehensible (Forte and
Siliotti 1997). Furthermore, researchers are able
to investigate each excavation area or unit and
the data collected from these units in threedimensions.
2. After visualizing an archaeological site in the
StarCAVE it is used as a heuristic tool. One is able
to revisit the site and data again and again without
ever going back to the field. In this manner we can
Figure 19. Interactive visualization and co-location of
different artifacts from Jordan on HIPerSpace.
The
HIPerSpace is made with 72 ‘off-the-shelf’ Dell 3007WFPHC, 30” displays or tiles (Photo: Falko Kuester, Calit2) .
149
Cyber-Archaeology
investigate the topological and spatial relationship
between artifacts, features and other areas of the
site and how these changed through time.
3. The StarCAVE is also used as a virtual reality
Geographic Information System. GIS programs,
such as ESRI’s ArcView, allow the user to browse,
query and manipulate the database, but these
cannot handle three-dimensional data. Virtual
reality technology such as the StarCAVE does,
however, operate in a three-dimensional virtual
world, and is sophisticated enough to simulate a
‘real’ environment. The precise location of each
recorded artifact, feature and locus is, in a manner
of speaking, put back together again. Additionally,
the incorporation of GIS databases into the virtual
reality model imparts the archaeologist with the
ability to perform spatial and statistical analyses
similar to the tools available in standard GIS
programs.
4. The StarCAVE contributes to cultural heritage
preservation. The unfortunate consequence of
archaeological research is we destroy that which
we study. Now, incorporating the many sources
of data we collect into a virtual reconstruction of
the site preserves a record of what was destroyed
during the course of excavation in a manner that is
more compelling to a large audience than a twodimensional representation.
Our 3D virtual reality visualization application is designed
to run within COVISE. COVISE allows us to design software
applications at the desktop and then run them in a large
variety of virtual environments, including Power Walls,
CAVEs, and tiled display walls, including the StarCAVE at
Calit2. The StarCAVE is a 5-walled, rear-projected, 360
degrees virtual reality device. It uses 34 high definition
(1080p) projectors to generate passive stereo images
on 15 screens and the floor. We use head tracking and
a 3D input device to navigate and interact with virtual
environments. Our application uses COVISE’s VRML loader,
as well as the previously described artifacts plugin. We
converted the Google Sketchup model to a VRML file in
order to load it directly into COVISE. Once that happened,
we were able to navigate around the terrain and to all the
excavation sites modeled, as well as get a bird’s eye view
of the entire area. In addition to the navigation around
the scene, the users can scale the size of the excavation
area so that they can display the 3D structures life size,
which makes them appear as if the users were on site.
Alternatively, the size can be scaled down so that the
area looks like a small model where everything is within
arm’s reach. The ability to change the scale helps point
to specific locations, and will help in the artifact display
mode to select larger volumes of artifacts than the person
can comfortably reach when displayed life size. However,
the ability of the display to convey size like in the real
world allows the user to not only perceive objects at their
original level of scale, but also to measure distances and
sizes with their hands, or a virtual measuring tape.
The current implementation allows the user to optionally
display or hide the artifacts or the Sketchup model, so
that one can focus on either without cluttering the screen
with the other. This functionality is selected directly from
within the virtual environment using a 3D menu. The 3D
menu API we use is part of the COVISE framework and
allows the programmer to add buttons, check boxes,
sliders, dials, submenus, and custom dialog windows.
Although the menus are mostly 2D and resemble menu
systems in desktop systems, they can be moved around
Figure 21. Jürgen Schultz and Kyle Knabb demonstrate Iron Age building and
excavation section through associated slag mound from KEN in the StarCAVE. The
StarCAVE is a 360o total immersive 3D VR environment with additional imagery on
the floor (Photo: Eric Jepsen, UCSD-Cali2).
150
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
freely in the 3D environment so that they are not in the
way when exploring the virtual world.
NexCAVE
Calit2 virtual reality researchers offer archaeologists
a 3-to-21-panel, 3-D visualization display made from
newly available synchronized 3D HDTVs. The technology,
dubbed “NexCAVE,” was designed and developed by
Calit2 Research Scientists.. The NexCAVE technology was
developed at the behest of Saudi Arabia’s King Abdullah
University of Science and Technology (KAUST), which
established a special partnership with UC San Diego
last year to collaborate on world-class visualization and
virtual-reality research and training activities (Figure 22).
When paired with polarized stereoscopic glasses, the
NexCAVE’s modular, micropolarized panels and related
software ill make it possible for archaeologists to visualize
massive datasets in three dimensions, at unprecedented
speeds and at a level of detail impossible to obtain on
a typical computer display. The NexCAVE’s technology
delivers a faithful, deep 3-D experience with great color
saturation, contrast and very good stereo separation. The
JVC panels’ xpol technology circularly polarizes successive
lines of the screen clockwise and anticlockwise and the
glasses you wear make you see, in each eye, either the
clockwise or anticlockwise images. This way, the data
appears in three dimensions. Since these HDTVs are
very bright and high-contrast, 3-D data in motion can be
viewed in a normally-lit environment, even with the lights
in the room fully on.
The NexCAVE’s data resolution is superb, close to human
visual acuity (or 20/20 vision). The 10-panel, 3-column
NexCAVE has a ~6000x1500 pixel resolution, while
the 21-panel, 7-column version built for KAUST has
a ~15,000x1500-pixel resolution. The NexCAVE’s LCD
screens are scalloped “like turtle shells,” which allows
the screens’ bezels (frames) to be minimized because the
screens are tucked behind one another. This works well in
3-D because the virtual reality illusion is so strong that you
don’t even see the screens and bezels as “windows,” just
the 3-D images in motion and stereo.
NexCAVE’s specially designed COVISE software
(developed at Germany’s University of Stuttgart) and
CGLX software (developed at UCSD) combine the latest
developments from the world of real-time graphics and
PC hardware with high-end Nvidia game engines. The
Calit2 and KAUST NexCAVEs are connected via 10 gigabit/
second networks, which allows researchers at KAUST to
collaborate remotely with UCSD colleagues. NexCAVEs are
being designed and built for several new partners around
the world. For the opening of the KAUST celebration in
September, 2009, archaeological data from the UCSD
Levantine Archaeology Lab excavations in Jordan that had
been originally modeled for the StarCAVE was featured in
the NexCAVE. We are now building a portable NexCAVE
prototype that will help answer the problem of ‘data
avalanche’ now facing field archaeologists.
Figure 22. The 21-Panel Xpol LCD stereo NexCAVE with
UCSD – Department of Antiquities of Jordan excavation
data from KEN on display at the opening of the King
Abdullah University of Science and Technology (KAUST),
September, 2009. Shown here are Tom Levy, UCSD (left)
and Sami Almagouth, KAUST, with the NexCAVE demo in
Saudi Arabia (Photo: T. DeFanti, Calit2).
Conclusion – the data avalanche
This paper has presented an integrated system of
on-site digital archaeology coupled with an active
cyberinfrastructure for Levantine archaeology with data
that is capable of being viewed with a variety of 3D
visual instruments. The discussion of data acquisition
techniques revolves around the problem of how to acquire
digital data in the first place. We refer to this as creating a
‘digitally enabled archaeology.’ Here the type of data and
how it can be obtained is outlined in relation to a number
of imaging techniques used by the UCSD CISA3/Calit2
team. These include Total Station, GPS, LiDAR, balloon
based airborne imaging, and 3D artifact scanning. How
these data are synthesized is also discussed. To achieve a
truly “Cyber-Archaeology” the global networks that portal
science cyberinfrastructures such as the Mediterranean
Archaeology Network (MedArchNet) and its Pottery
Informatics Queryable Database are described. Finally, a
number of visualization paradigms and environments for
Cyber-Archaeology that take advantage of these datasets
are also described. The ones used by the CISA3/Calit2
team include the HiPerSpace, StarCAVE, and NexCAVE.
Based on the Jordan field work discussed here, our team
has experienced a ‘data avalanche.’ This is highlighted in
Table 1 that outlines the range of digital data collection
instruments used in the field in 2007 versus those used in
2009. The exponential increase in digital data (from 172
GB in 2007 to 1,373 GB in 2009 is astounding. How will
archaeologists deal with this data avalanche in their field
projects in the future? Our team is working on a portable
NexCAVE that will have software, computing power
and portability needed to thread together all these rich
sources of data in a manageable manner. One of our goals
is to equip our helium balloon platform with stereo digital
video cameras to allow our x,y and z archaeology data to
be fully integrated into the 3D visualization instruments
currently evolving.
151
Cyber-Archaeology
Data Collection Instrument
2007 Excavation Season
2009 Excavation Season
3D NextEngine Scanner (Artifacts)
0
168 GB
High Res Digital Artifact Photography
70 GB
253 GB
High Res Digital Site Photography
10 GB
52 GB
Balloon-based Stereo Digital Photography
0
253 GB
Gigapan Panorama Photography
0
15 GB
Terrestrial LlDAR Scanning
0
500 GB
20 GB
30 GB
GIS Data
HD Video
72 GB
120 GB
TOTAL
172 GB
1,373 GB
Table 1. Data avalanche from 2007 to 2009, UCSD CISA3/Calit2 expeditions in Jordan.
Acknowledgements
Much of this project was funded by two UCSD Chancellor
Collaboratory grants awarded for the 2008/2009 and
2009/2010 academic years and the CISA3/Calit2 CyberArchaeology Laboratory at UCSD. Thanks especially to
Ramesh Rao, Director, and the staff at Calit2, San Diego
Division for support of the balloon project. Thanks also to
Dr. Fawwaz Al-Khraysheh, Director General, Department
of Antiquities of Jordan for hi support.
References
Altschul, S. F., W. Gish, M. Webb, E. W. Myers, and D. J.
Lipman. 1990. Basic Local Alignment Search Tool.
Jorunal of Molecular Biology 215:403-410.
Barton, J. 2009. 3D laser scanning and the conservation
of earthen architecture: a case study at the UNESCO
World Heritage Site Merv, Turkmenistan. World
Archaeology 41:489 - 504.
Baru, C. in press. “Portal Science Today: Building a Web
Portal System for Archaeology,” in Portal Science and
Archaeology - Views from the Mediterranean Lands.
Edited by T. E. Levy, S. H. Savage, C. Baru, and O.
LaBianca. London: Equinox.
Bewley, R. H., S. P. Crutchley, and C. A. Shell. 2005. New
light on an ancient landscape: Lidar survey in the
Stonehenge world heritage site. Antiquity. 79:636647.
Bunimovitz, S., and Z. Lederman. 2009. The Archaeology
of Broder Communities - Renewed Excavations at Tel
Beth-Shemesh, Part 1: The Iron Age. Near Eastern
Archaeology. 72:114-142.
Daly, P., and T. L. Evans. Editors. 2005. Digital Archaeology
- Bridging Method and Theory. London: Routledge.
DeFanti, T., Dawe, G., Sandin, D., Schulze, J., Otto, P.,
Girado, J., Kuester, F., Smarr, L., and Rao, R., 2008a
“The StarCave, a third-generation cave and virtual
reality Optiportal,” Future Generation Computer
Systems/The International Journal of Grid Computing:
Theory, Methods and Applications, Elsevier B.V., Vol
25, Issue 2, Nov. 2008. Also: http://ivl.calit2.net/wiki/
index.php/StarCAVE_Measurement_Diagrams.
DeFanti, T., Leigh, J., Renambot, L., Jeong, B., Verlo, A,
Long, L., Brown, M., Sandin, D., Vishwanath, V., Liu,
Q., Katz, M., Papadopoulos, P., Keefe, J., Hidley, G.,
Dawe, G., Kaufman, I., Glogowski, B., Doerr, K.-U.,
Girado, J., Schulze, J., Kuester, F., Smarr, L., 2008b
“The OptIPortal, a Scalable Visualization, Storage, and
Computing Interface Device for the OptIPuter,” Future
Generation Computer Systems/The International
Journal of Grid Computing: Theory, Methods and
Applications, Elsevier B.V., Vol 25, Issue 2, Nov. 2008.
Doerr, K. U., and F. Kuester. 2010. “CGLX: A Scalable,
High-performance Visualization Framework for
Networked Display Environments,”. IEEE Transactions
on Visualization and Computer Graphics 99.
Eiteljorg II, H. 1994. “Using a Total Station,” in CSA
Newsletter, vol. VII.
Forte, M., and A. Siliotti. Editors. 1997. Virtual
Archaeology: Re-creating Ancient Worlds. New York
Harry N. Abrams.
Forte, M. 2008. Virtual archaeology: communication in
3D and ecological thinking. BAR International Series
1805:21-35. Oxford: Archaeopress.
Fox, T. 2008a. “CISA3 Researchers Analyze Native
American Site in Push for Digital Archaeology,” in
Calit2 Newsroom, vol. 2008, pp. http://www.calit2.
net/newsroom/article.php?id=1284.
Fox, T. 2008b. “CISA3 Archaeologists Develop 3-D Images
of Artifacts for Digital Pottery Informatics Database,”
in California Institute for Telecommunications and
Information Technology Newsroom, vol. 2010. San
Diego: UCSD Calit2.Access date: May 7; http://www.
calit2.net/newsroom/article.php?id=1357
Guidi, G., F. Remondino, G. Morlando, A. Del Mastio, F.
Uccheddu, and A. Pelagotti. 2008. “Preformances
evaluation of a low cost active sensor for cultural
heritage documentation.”
Guy, P.L.O. 1932 Balloon Photography and Archaeological
Excavation. Antiquity 6:148-155
Haiman, M. 1996. Early Bronze Age IV Settlement Pattern
of the Negev and Sinai Deserts: View from Small
Marginal Temporary Sites. Bulletin of the American
Schools of Oriental Research 303:1 - 32.
Karasik, A. 2008. “Applications of 3D technology as a
152
On-Site Digital Archaeology 3.0 and Cyber-Archaeology
research tool in archaeological ceramic analysis. ,” in
Beyond Illustration: 2D and 3D Digital Technology as
Tools for Discovery in Archaeology. Edited by Bernard
Frischer and Anastasia Dakouri-Hild. pp. 103-116.
British Archaeological Reports International Series
1805. Oxford: Archaeopress.
Katz, O., Cabanes, D., Weiner, S., Maeir, A. M., Boaretto, E.,
and Shahack-Gross, R. 2010. Rapid Phytolith Extraction
for on-Site Analysis of Phytolith Concentrations and
Assemblages: An Application at Tell es-Safi/Gath,
Israel. Journal of Archaeological Science 37(7): 1557–
63.
Knabb, K., J. Schultze, and T. E. Levy. in press. “StarCAVE:
3D Virtual Reality Environments for Archaeology”
in Portal Science and Archaeology – Views from the
Mediterranean Lands. Edited by T. E. Levy, S. Savage, C.
Baru, and O. LaBianca, pp. 16 pages. London: Equinox.
Levy, T. E. Editor. 1987. Shiqmim I: Studies Concerning
Chalcolithic Societies in the Northern Negev Desert,
Israel (1982-1984). Oxford: BAR International Series
356.
—. 2010. “The New Pragmatism: Integrating
Anthropological, Digital, and Historical Biblical
Archaeologies,” in Historical Biblical Archaeology and
the Future – The New Pragmatism. Edited by T. E. Levy,
pp. 3-44. London: Equinox.
Levy, T. E., J. D. Anderson, M. Waggoner, N. Smith, A.
Muniz, and R. B. Adams. 2001. Interface: Archaeology
and Technology - Digital Archaeology 2001: GISBased Excavation Recording in Jordan. The SAA
Archaeological Record 1:23 - 29.
Levy, T. E., T. Higham, C. Bronk Ramsey, N. G. Smith, E. BenYosef, M. Robinson, S. Münger, K. Knabb, J. Schulze,
P., M. Najjar, and L. Tauxe. 2008. High-precision
radiocarbon dating and historical biblical archaeology
in southern Jordan. Proceedings of the National
Academy of Sciences 105:16460-16465.
Levy, T. E., and N. G. Smith. 2007. “On-Site Digital
Archaeology: GIS-Based Excavation Recording in
Southern Jordan,” in Crossing Jordan – North American
Contributions to the Archaeology of Jordan. Edited by
T. E. Levy, M. Daviau, R. Younker, and M. M. Shaer, pp.
47-58. London: Equinox.
McCoy, M., and T. Ladefoged. 2009. New Developments in
the Use of Spatial Technology in Archaeology. Journal
of Archaeological Research 17:263-295.
Vardi, J., S. Shilstein, S. Shalev, and Y. Yekutieli. 2008.
The Early Bronze Age IV Chipped and Ground Stone
Assemblage of ‘En Yahav and its Relation to Copper
Smelting Activities. Journal of the Israel Prehistoric
Society 38:1-20.
Yekutieli, Y., S. Shalev, and S. Shilstein. 2005. ‘En Yahav - A
Copper Smelting Site in the ‘Arava. Bulletin American
School of Oriental Research 340:35-55.
153