urface. In
by angle
al factors
irfaces.
imited to
such as
However,
ucted for
n objects,
uctures.
diameters
eir point-
ied to the
fects will
the data
ie. overall
related to
he model
test field
thout any
iform for
normal in
s wind or
als were
cast iron,
diameters
with the
objects is
1 different
set up
ice from a
5 to 3.3m.
oints on à
to 178° on
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004
a diameter of 30mm to 100mm within an observation distance
of 2.9m. The scanning resolution was approximately 1.7mrad
(i.e., 5mm point spacing at a range of 3m) and the number of
points was between 30,000 and 220,000.
The first set of scanning data was acquired from stations 1, 2,
and 3 and the second set (stations 4, 5, and 6) was conducted
after applying a dulling spray on the surfaces.
Two scans were captured at each scanner station; the first
one had an approximate sample spacing of 5.0 mm x 5.0mm
and the second was 2.0mm x 2.0mm. The first scan was sparse
using a fast scan to examine the point cloud data to see if it was
reliable. This took less than a minute. For the purpose of the
experiment, the second data set involved a sample spacing of
2.0mm x 2.0mm. This set was dense and the data acquisition
time was less than 3 minutes per scan.
3. Analysis of Coverage Angles
3.1 Observation of Coverage Angles on different surface
materials
Point-clouds data from scanning stations 1 to 6 were selected
to observe coverage angles on the different surfaces. From the
scanning data, each cylindrical object was closely examined
and unnecessary points were eliminated for cylinder fitting. By
use of ‘fit-to-cloud’ option, cylinders were formed for each
object and their diameters were determined. Point-clouds data
were exported into text and “dxf” file formats to observe the
coverage angles and to conduct residual analyses.
Based on a diameter measurement obtained by a caliper, a
circle was drawn and fit into imported point-clouds data. Two
lines were constructed by connecting the center of circle and
one of the points, which represents the point on the surface.
The coverage angle was observed by intersecting two lines. As
shown in Figure 2, small squares in red and yellow are
represented point-clouds on the surface and circles and lines are
drawn in blue and three locations were arbitrarily selected, then
coverage angles were measured. From three observations the
maximum value was selected and calculated its percentage of
coverage (i.e. 360? is the total coverage of cylindrical object).
The diagram of maximum coverage angles for each object is
schematically shown' in Figure 3. In the case of PVC and
stainless steel, the coverage angles are 62.19? and 59.635,
which represent very poor percentage coverage of 17.396 and
16.6%, respectively. The cylinder fitting process with these
point-clouds was not possible due to a limited coverage area. In
the case of brass and ceramic, the coverage angles were
observed at 165.13? (45.996) and 126.49? (35.194), respectively
(see Table 1). From the experiment, it can be concluded that the
percentage of coverage angle should be at least 2096, but 3096
is recommended, in order to fit a cylinder model properly to
point-clouds.
The coverage angles of point cloud in PVC and stainless steel
were less than 20°, which was not useful to model objects
properly due to a lack of data points. In the case of PVC pipes,
the colour of the surface, which was black, was also one of the
factors that degraded the point cloud data.
Sufficient reflecting signal is a key factor for capturing proper
point data from field measurements. Since the energy of the
reflected pulse depends on physical properties of surface
material it is worth to investigate if any types of media could
improve surface reflectance without degrading accuracy. An
application of chalk dust or a light coat of emulsion has been
introduced to increase reflectance of material. In this
experiment, a dulling spray and masking tape were applied as
a medium for altering surface properties.
3.2 Application of the Dulling Spray
In order to acquire enough coverage point-clouds with
reliable accuracy, increasing the reflectance would be the first
method to be applied. A dulling spray was applied on the area
of the surface except brass and ceramic objects, which had
enough coverage angles. The thin clear coat was formed on the
surface and the coat was transparent. The advantages of this
application are 1) easy and fast application, 2) easily removable
by rubbing with or without water, 3) no effect on most
materials and, 4) no effect on measurement due to its thin clear
coating. Dulling spray was applied on the bottom half of the
objects to compare the differences with “as-is” condition.
A short period time was necessary for the spray to completely
dry. Once replaced, the object scanning process proceeded
under the same laboratory conditions as described in the
experiment setup.
From the point clouds, three observations were made to
measure a coverage angle within distribution of points on
surfaces. The impact of spray was quite extensive in terms of
increased coverage angles, which is critical to fit objects during
the modelling process. The reflectance difference after applying
dulling spray, as clearly shown in Figure 4, can be observed
visually in the point cloud data. The bottom half of the area was
coated with the dulling spray and had a wider coverage angle
than areas without spray by a factor of 61%.
The increase in percentage coverage was within the range
of 40% to 61% and the modeling process was now possible
with PVC and stainless steel, which had less than 20%
coverage angle without the spray.
The thin sprayed coat was easily removed by rubbing with a
cloth two days after the spray application was applied, and no
indication of g surface change was observed. However, heat
and high temperature of piping objects might be problematic in
practical field applications such as pipes in industrial process
plants. It should be handled carefully where open flame or
sparks exist. The effects of heat and high temperature have not
been studied at this time.
3.3 Application of Masking Tapes
Masking taping applied to the surface is another technique
to increase the coverage angles on certain surfaces such as
aluminum. The use of masking tape can alter the geometric
factors such as surface roughness and the angle of incidence.
To increase the surface roughness on an aluminum foil
surface masking tape was wrapped around the top half of
cylinder and scanning points were acquired as shown in Figure
5. A set of scattered red points appear on the left side of
aluminum surface and most of area is blacked out, which also
can be seen as shadow effects in other areas. The result
indicates that the laser beam did not return to the scanner head,
which means the laser energy was scattered due to the very
shinny and smooth (textureless) surface.
The point-clouds on masking taped area are shown
with yellow and green colors while it is mainly red in
aluminum surface area. Aluminum or galvanized steel is widely
used for insulation of piping system and the application of
