-36- 
Figure 2. Test target showing the 5 specific positions 
With help of the Trimble total station, capable of measuring 
reflectorless, all five positions on all test targets were 
determined. The positions were over-determined, and from all 
the measurements, an averaged value was given to each 
position. 
Cyclone was later used to evaluate the scanner data from all the 
systems. Data from the Optech system was converted in to an 
ASCII text-file exchange format, the .ptx file format, and 
thereafter imported to Cyclone. The .ptx file format includes the 
X, Y, Z and intensity value for measured points. For the Riegl 
system, the data was imported by using the Riegl .3dd file 
format. 
In the Cyclone software, each test target was modelled into a 
plane. A best fitting algorithm for a plane, available in Cyclone, 
was used. For the best fitting of a plane to each target, measured 
points on the test targets, from all ten measurements, were used. 
From the planes, at the five specific position spots that had been 
measured with the Trimble, point-to-point distances between 
the different planes, for each system, were measured. 
By fitting a plane to the test target the precision for each system 
in measuring towards the test targets was given. Furthermore, 
using the measurements from the Trimble as true values the 
accuracy for each system was tested. 
2.4 Representation of reality - scanning of purposely built 
test objects. 
Two test objects have been specially designed and built for 
evaluating laser-scanning results for problematic areas in the 
built environment. One of the test objects consists of parallel 
mounted plates, approximately 60 by 40 centimetres in size, 
with fixed distances between, and with cut-outs of known 
geometrical figures in the front plate. The plates are made of a 
high reflective material and all measures, including the cut-outs, 
are within sub-millimetre accuracy. To be able to evaluate 
different angle of incidences and surface structures a one cubic 
meter box has been built. Painted on the sides are different 
surface materials and the box is mounted on a centre axis to 
allow the possibility to rotate it. In addition, the top of the box 
is equipped with points of attachment for specific registration 
spheres. A third object was constructed out of two thin metallic 
plates and used for testing the problem of measuring and 
modelling small opening angles between surfaces. The plates 
were position either with the plate angle opened (>) towards the 
scanner or closed (<) towards the scanner. All this special 
objects were placed together so they could be measured all at 
the same time. The measurements of the objects were done with 
all three scanners more or less at the same time. The scanners 
were placed with a very small offset to each other, having all 
scanner heads within a box of approximately 3.0x0.5x0.5m size. 
Scans were made with different resolutions, at different ranges, 
and with different positioning of the objects between each 
scans. The box was rotated around the central axis into eight 
different positions turning it round a full circle. When 
positioning the box the 60 by 40 cm plate was also positioned 
into its different angle of incidences before each scan. The thin 
metal plates were also placed in position and the opening angle 
was measured before each scan. 
These objects were used when evaluating the behaviour of the 
different scanner results in detecting edges, scanning into small 
openings and resolve small objects. The influence from the 
different materials on the box was also to be evaluated. 
2.5 Measurements on targets with known reflectivity 
A target with four areas with different reflectivity was used 
provided by the Laser Systems department. The reflectivity 
ranged from 20% black to 90% white and included two grey- 
levelled areas on the target with know reflectivity. 
The measurements of the objects were done with all three 
scanners more or less at the same time. The scanners were 
placed with a very small offset to each other, having all scanner 
heads within a box of approximately 3.0x0.5x0.5m size. 
The test target was scan with a point-to-point spacing to give to 
give acceptable amount of points on the target. 
The target was placed in eight different positions, and in all 
positions being almost perpendicular to the systems scanning 
direction. The positions went from 10 to 80 meters with IO 
meter increment and with the meter range being scanned twice, 
as the first and the last scan. 
Possible linearity in recording reflectivity, possible drift over 
time and possible problems in register intensities on same 
reflectivity material will be evaluated. 
3. RESULTS 
3.1 Systems range precision and accuracy 
For all the systems, a specific point-to-point spacing at a certain 
distance was set. See following tables. Ten repetitive scans of 
the object area was done for all the systems. For each system, 
including the points from all ten scans, all points for each target 
were modelled into a plane. Only points well inside the targets 
were included in the modelling. Number of points gives a figure 
of how many points that were included in modelling each test 
target plane. 
Cyrax 30 mm point-to-point spacing@60m 
Standard dev. 
Max error 
Number of 
points 
Target 1 
3 
15 
59459 
Target 2 
3 
11 
14664 
Target 3 
2 
9 
1690 
Target 4 
3 
14 
632 
Target 5 
4 
17 
250 
Table 3. Modelled precision for Cyrax 2500.