C/PA 2003 XIX th International Symposium, 30 September - 04 October, 2003, Antalya, Turkey
helping to provide experience and understanding of the type
of large scale survey that may, in the future, be performed
using TLS. It also provided a basis from which to address
important data management issues. A total of 43 scans were
collected using a Cyrax 2500 TLS and over 50 control points
were used to register data to the local site grid. To ensure the
independence of the specification document from any one
particular system three different scanning systems were used.
In addition to the Cyrax 2500, a Zoller and Froelich Imager
5003 system was used to scan the presbytery (22 m x 14 m x
22 m in size) and, during the second survey, a Riegl LMS
Z320 system was used to scan several areas of the church.
Figure 1. The remains of the Priory Church (solid
shading represents standing walls) Hadcock, 1991.
The third survey will validate the specification document by
undertaking a “commercial’’ survey. A project brief will be
defined by English Heritage prior to the survey and data then
collected to the level described in the specification. The
survey data will then be delivered to English Heritage for
assessment and comment.
4. DISCUSSION
The following discussion is based on the data collected
during the first and second surveys in addition to consultation
with the steering committee. It outlines some of the
requirements and issues of the specification and illustrates
this with examples from the first two surveys.
4.1 Definitions
In order to ensure clarity throughout the specification
document a list of definitions was provided. A selection of
these definitions are provided below:
■ Data voids - Sections within the point cloud, more
than twice the point density of the scan in size,
which contain no data despite surface information
on the object itself.
■ Peripheral data - Additional scan data collected
during the scanning process not explicitly defined
as being required in the project brief.
■ Point cloud - A collection of XYZ coordinates in a
common coordinate system that portrays to the
viewer an understanding of the spatial distribution
of a subject. It may also include intensity or colour
information. Generally a point cloud contains a
relatively large number of coordinates in
comparison with the volume the cloud occupies.
■ Point density - The average distance between XYZ
coordinates in a point cloud. Density can be
represented by either a horizontal and vertical
separation taken from/to the centre of the footprint
at a specified range, or by using angular values.
■ Registration - The process of determining the
transformation parameters required to transform
point clouds onto a common coordinate system.
■ Resolution - The smallest discernable unit of
measurement.
■ Scan orientation - The approximate direction in
which the scan is made if the system does not
provide a 360 degrees field of view.
■ Scan origin - The origin of the arbitrary coordinate
system in which scans are performed. When the
scan origin is transformed onto the site coordinate
system it becomes the scan position.
■ Scan position - The location, in a known coordinate
system, from which a single scan is performed. If
the system does not perform a full 360 degree scan,
several scans may be taken from the same scan
position, but with different scan orientations.
■ Scanning artefacts - Irregularities within a scan
scene that are a result of the scanning process
rather than features on subject itself. These may be
geometric or radiometric in nature.
■ Terrestrial laser scanning system (TLS) - Any
ground based device that uses a laser to measure
the three-dimensional coordinates of a given region
of an objects surface automatically, in a systematic
order at a high rate in (near) real time (adapted
from Boehler and Marbs, 2002).
4.2 Data capture
Before any survey takes place the contractor must supply a
method statement detailing the techniques to be used. In the
case of TLS this must show the area to be scanned along with
details showing that the scanning system to be used is
operating correctly. It is considered vital that, for QA
purposes, such information is provided. However, as the
majority of manufacturers do not provide calibration
certificates some other standard is required. Details of a test
showing the system is operating to its supplied level of
precision and accuracy are required, although, as the exact
nature of such a test is likely to vary between users the
ultimate responsibility should be on the client to ensure
appropriate information is provided.
Accuracy in the recording of cultural heritage consists of
geometric accuracy, the closeness of a coordinate to its true
value, and narrative accuracy the “correctness” of a
particular attribute. From discussion with end users it was
clear that each type of accuracy must be specified for. In
TLS the two most important parameters for this are the
chosen point density and the geometric accuracy of the point
measurement. It is necessary to determine suitable values for
both. It is possible to assume that point density is mainly
