CIP A 2003 XIX th International Symposium, 30 September - 04 October, 2003, Antalya, Turkey
translation vector is the scanners position and the
column vectors (r^ r 2 j r3j) T are the directions of the local
coordinate axes in PRCS.
Figure 2: Example for the PRCS, GLCS, and a number of SOCS at
a site for scanning a building.
Each scan position holds the scan data taken at this
scan position, stored in the scanner’s binary data
format with extension 3dd. Furthermore, each scan
position holds it’s SOP information. In order to
transform data from SOCS into PRCS the data points
are simply multiplied with the SOP matrix (M S op) of
the scan position.
In case a data point P have to be transformed from a
specific scan position into GLCS, a sequential
multiplication has to be carried out, first with the Msop
matrix of the scan position to get into PRCS, then Mp 0 p
which transforms from PRCS into the GLCS.
The process of registration of the various scan
positions in the PRCS is basically the determination of
the respective M S op- This process is based usually on
tie points within RiSCAN PRO. Tie points are
managed by tie point lists (TPL). Tie points are
usually defined by retro-reflective targets showing up
clearly in the intensity data of the scan data and which
can also be accurately localized by the use of total
stations. The tie point itself is commonly the center of a
reflective target.
Every project can hold one tie point list in the project
coordinate system TPL (PRCS). Usually the data are
gained by measuring the position of the tie points with
a total station or by other means, e.g., DGPS. If the tie
point coordinates are available in global coordinates,
they are entered into the global tie point list, TPL
(GLCS), and are subsequently transferred into the TPL
(PRCS) by defining an appropriate matrix M> 0 p to
fulfil the requirement of single precision
representation.
In order to register a single scan position into the
project coordinate system, a tie point list in the SOCS
system have to be acquired, which s done by fine-
scanning the retro-reflective targets visible from the
specific scan position. RiSCAN PRO extracts retro-
reflective targets from a so-called overview scan or
panorama scan and supports the automatic subsequent
sequential fine scan of the targets. Once sufficient tie
points have been gained M S op can be determined
automatically and the scan data is transferred into the
project coordinate system as well as into the global
coordinate system.
3 Use of high-resolution images
within RiSCAN PRO
RIEGL 3D imaging sensors can be equipped with an
optional high-resolution digital camera. The images
can be used to assign a color to the vertices of the point
cloud data or to apply the images as a texture to the
meshed surface generated from scan data. Before being
able to apply the image information to the scan data a
camera model has to be selected and the camera has to
be calibrated.
RiSCAN PRO makes use of an 10-parameter model for
the internal camera calibration (Intel 2003). The
internal parameters of the camera model can be
determined within RiSCAN PRO in different ways and
are described below in the calibration tasks section. In
order to make use of the image data calibration data
holding the internal parameters of the camera have to
be available. The calibration data are contained in the
calibrations section of the project file. The calibration
data have to be determined for every camera to be used
and for every setting of the camera’s focus and aperture
to be used.
In order to utilize the image information also the
external orientation of the camera has to be known. The
definition of the external orientation differs for the two
different kinds of images managed in RiSCAN PRO:
Type 1 Images: Images taken when the camera is
mounted on top of the scanner
Type 2 Images: all other images, i.e., when the camera
is NOT mounted on the scanner
Type 1 images are stored in the folder corresponding to
the scan position at which the images have been taken.
As the camera is firmly mounted on top of the scanner
and the orientation of the rotational platform is well
known, the orientation of the camera within the
scanner's coordinate system is well defined after
