PFG 2015 / 5, 0357 – 0372 Article

Stuttgart, October 2015

Accuracy of Laser Scanners for Measuring Surfaces

made of Synthetic Materials
GRZEGORZ LENDA, URSZULA MARMOL & GRZEGORZ MIREK, Cracow, Poland

Keywords: terrestrial lidar, measurement error, material tests

Summary: Synthetics are a group of materials Zusammenfassung: Genauigkeit von Laser-Scan-

which, when observed using laser scanners, may nern bei der Oberflächenmessung von syntheti-
cause additional measurement errors. These errors schen Materialien. Synthetische Materialien kön-
result from optical permeability, which may vary nen bei der Erfassung von Oberflächen mit Laser-
depending on the structure of the synthetic materi- scannern zu zusätzlichen Messfehlern führen.
al. A laser beam which penetrates the material can Diese Fehler ergeben sich aus der optischen Durch-
be subject to numerous physical phenomena such as lässigkeit, die in Abhängigkeit von der Struktur des
absorption, dispersion, full or partial reflection, re- synthetischen Materials variieren kann. Ein Laser-
fraction at the interface between two media, dif- strahl, der in das Material eindringt, kann mehre-
fraction and interference that change the measured ren physikalischen Phänomenen, z. B. Absorption,
distance. As a result, additional measurement er- Streuung, vollständige oder teilweise Reflexion,
rors occur, independent of the known errors, when Refraktion an Grenzflächen zwischen zwei Me­
the beam is reflected from the outer surface of the dien, Beugung oder Überlagerung, unterworfen
material. Their occurrence is influenced by light sein, welche die gemessene Distanz verändern. Als
transmission, density and internal structure of the Ergebnis erhält man zusätzliche und von bisher be-
synthetic material, its colour and thickness. The kannten Fehlerursachen unabhängige Messfehler,
distance from the scanner, laser beam angle of inci- wenn der Strahl an einer aus synthetischem Mate-
dence, temperature, scanner type (carrier wave- rial bestehenden Oberfläche reflektiert wird. Diese
length), or even the distance between the synthetic werden von der Lichtdurchlässigkeit, der Dichte
material and its background are important as well. und der inneren Struktur des synthetischen Mate-
This study is devoted to examining these depend- rials beeinflusst. Die Distanz vom Scanner, der
encies. Currently there are no studies distinguish- Auftreffwinkel des Laserstrahls, die Temperatur,
ing and examining synthetic materials as a separate der Scannertyp (Trägerfrequenz des Lasers) oder
group of materials that may interfere with distance auch die Distanz der gescannten Oberfläche von
measurements due to the phenomena that occur in Objekten, die sich aus Sicht des Sensors im Hinter-
their structure. grund befinden, sind ebenfalls von Bedeutung.
Dieser Beitrag widmet sich der Untersuchung die-
ser Einflüsse. Bisher gibt es keine Studien, welche
synthetische Materialen als eigene Gruppe von Ma-
terialen unterscheiden, welche in spezifischer Wei-
se die Distanzmessung auf Grund ihrer inneren
Struktur beeinflussen.

1 Introduction
permeability, which may vary depending on
Synthetic materials (materials consisting of the structure of the synthetic material. Exist-
synthetic polymers) are a group of materi- ence of this phenomenon may be proved by
als which, when observed using laser scan- the following experiment. One can measure
ners, may cause additional measurement er- a distance to a synthetic element, for exam-
rors. These errors result from the optical ple 2 mm – 3 mm thick, consisting of white

© 2015 E. Schweizerbart'sche Verlagsbuchhandlung, Stuttgart, Germany www.schweizerbart.de

DOI: 10.1127/pfg/2015/0273 1432-8364/15/0273 $ 4.00
358 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

Plexiglas or polystyrene, using any type of re- midity, temperature), and the properties of the
flectorless range finder, then measure it again scanned surface.
without changing the position of the element Considering the properties of the materi-
but with an identical element adhered to its al, which this article focuses on, the research
back surface. The observed distance will be work carried out by BOEHLER et al. (2003) is
greater due to laser beam penetration through worth mentioning. Their paper includes a se-
added layers of material. On the other hand, ries of tests for checking the quantitative and
covering the front surface with a sheet of pa- qualitative accuracy of points measured us-
per will prevent the laser beam from penetrat- ing nine laser scanners selected from differ-
ing and result in a shorter measured distance. ent manufacturers. Among other things, the
A laser beam penetrating the material may ini- relation between surfaces of varying reflectiv-
tiate a number of physical phenomena, chang- ity and the accuracy of the acquired informa-
ing the measured distance. As a result, addi- tion was tested. The following samples made
tional measurement errors occur, independent of various materials and in different colours
of the already familiar errors that occur when were used: white paint (90 % and 80 % reflec-
the beam is reflected from the outer surface tive), grey paint (40 % reflective), black paint
of the material. Even cursory tests of select- (8 % reflective), metallic paint, various kinds
ed materials demonstrate that in some cases of films. Measurements of different colour
these errors can appear much bigger than the surfaces yielded good accuracy results. For
assumed accuracy of the scanner. Due to the the white surfaces, a deviation of zero was
growing use of synthetic elements in various obtained. Also, the grey and black colours
branches of technology and construction, as brought zero errors for most scanners. The sit-
well as the high prevalence of scanning tech- uation was much worse in the case of differ-
niques for the measurements of all types of ent types of films, where errors from a few up
structures (BERENYI et al. 2010, HIREMAGALUR to tens of millimetres occurred. Contradictory
et al. 2007, HOLST et al. 2012, JOHNSON & JOHN- results were shown in BUCKSCH et al. (2007),
SON 2012, MONSERRAT & CROSETTO 2008, SA- where the bright surfaces gave much greater
LEMI et al. 2008), a need arises to investigate accuracy than the dark ones. The problem as-
the influence of these materials on errors of sociated with material reflectance has been
measured distances. analysed in CLARK & ROBSON (2004). The au-
In the literature on examining the accuracy thors selected materials of standardized col-
of laser scanners there have been no studies ours and textures, and scan them from vari-
that address measurements to synthetic mate- ous distances (4 m – 24 m) at the angles of
rials, although some of them indicate poten- 20°, 40° and 60°. The experimental results
tial problems associated with measuring ma- revealed systematic discrepancies in the re-
terials having a certain degree of transparency corded distances, depending on the type of
(EREN et al. 2009, SALEMI et al. 2008, VOEGTLE the scanned surface. Colours which exhibit
et al. 2008). Generally, on the basis of sever- low reflectance, such as black or red, lead to
al laser scanning studies (ABBAS et al. 2013, longer distance measurements than the refer-
EREN et al. 2009, GOTTWALD 2008, KAASALAI- ence distance. The authors proposed a correc-
NEN et al. 2011, KERSTEN et al. 2009, LEE et al. tion factor, reducing these systematic errors.
2010, LICHTI & JAMTSHO 2006, POLO et al. 2012, A similar problem was presented in PFEIFER et
SOUDARISSANANE et al. 2007) and those given al. (2008), where the authors focused on the
below, it is known that the measurement ac- signal energy returning to the receiver. It was
curacy is affected by several factors, includ- hypothesized that this energy can be deter-
ing: the type of a scanner (pulse-based, phase- mined based on the observed range and inten-
based), the precision of a scanning mechanism sity value. It was proven that it was possible
(laser footprint, accuracy and resolution of the to reproduce the reflectance of a given surface
horizontal and vertical angular encoders, the with the accuracy at the level of 6 %. Surface
eccentricity of the scanning system), scanning reflectance is also dealt with in Zámečníková
geometry (beam angle of incidence, scanning et al. (2014b). The authors used the standard
distance), external conditions (lighting, hu- white colour (spectralon), which is the mate-
Grzegorz Lenda et al., Accuracy of Laser Scanners 359

rial of almost perfect Lambertian reflectance, dence on the resulting accuracy was analyzed.
as well as cardboard boxes in varied colours. Considering the values of the mean square
The materials were moved on a specially con- errors, the following relation was noted: for
structed trolley in the range of 1.1 m to 29.7 m the perpendicular direction of the beam inci-
at every 7.5 cm. The authors found that the dence the errors are significantly larger than
obtained deviations depend not only on the for smaller angles. A metal plate (aluminium,
range, but also on the signal strength. high reflectance) also appears in Zámečníková
Another study of the properties of the et al. (2014b). The authors noticed that small
scanned objects was performed by VOEGTLE et changes in the scanning angle (0.1°) result in
al. (2008). This study used materials common- 20 % changes in the signal strength, and thus
ly applied on the façades of buildings, hav- to changes in the recorded deviations. The re-
ing different properties: colourful sheets and sults presented in that paper confirmed that a
shades of gray, different kinds of wood, metal higher reflectance results in larger errors for
elements, plaster of various granularity, trans- short distances, and for the distances exceed-
parent films and materials with varying de- ing 20 m their value decreases.
grees of moisture. The authors’ attention was In conclusion, it can be said that the research
drawn to the tests of partly transparent materi- hitherto conducted mainly analyzed the errors
als: errors ranged from 15 mm (for 35 % trans- caused by the reflection of a laser beam from
parency) to 34 mm (for 5 % transparency). the outer surface of a material. Synthetic ma-
It is not possible to analyse properties of the terials were neither tested nor distinguished
scanned material separately from the scan- as a separate group of materials which might
ning geometry. Both the angle of the beam distort the measurements of a distance due to
incidence and the distance have a significant their internal structure. Therefore, this work
effect on the resulting accuracy. In SOUDARIS- presents error analysis for distance measure-
SANANE et al. (2007) it was found that the ments of the elements of this type.
measurement noise for a fibreboard and ply- The authors initially studied measurements
wood painted white increases with increas- to synthetic materials in LENDA & MARMOL
ing angles. The effect of the scanning angle (2010) by testing the distance measurement
on the resulting accuracy was also analysed deviations of reflectorless range finders for
in SOUDARISSANANE et al. (2009) and SOUDARIS- a few selected materials. It enabled the first
SANANE et al. (2011). An increase in the meas- conclusions on the factors influencing the dis-
urement noise with the increasing scanning tance measurement errors of synthetic materi-
angle was identified. The authors proposed a als, as well as to plan appropriate methodolo-
model which optimized the point position and gy for the implementation of research studies,
corrected the influence of the angle. There taking into account a wider range of materials
was a significant improvement in the stand- for laser scanners. The studies carried out in
ard deviation, from the value of 3.25 mm to this paper present a wide range of issues. From
2.55 mm. The influence of the scanning an- the point of view of the users, this paper al-
gle on the final accuracy was also presented lows for a sufficient assessment of the impact
in Zámečníková et al. (2014a). For the angles of synthetic materials on the laser scanners’
100° – 65° and 50° – 45° the measured dis- distance measurements.
tances were longer than the reference dis-
tance, while in the range of 65° – 50° they
were shorter. The authors did not explain the 2 Research Methodology
specific character of this phenomenon. Hav-
ing eliminated the systematic factor, using a A laser beam in the structure of a synthetic
polynomial model, the values of standard de- material may be subject to a series of phe-
viations were significantly reduced. nomena, such as absorption, dispersion, full
From the point of view of the research re- or partial reflection, diffraction and interfer-
sults presented in our study, experiments with ence, and refraction at the interface between
a metal plate were interesting. In VOEGTLE et two media (FEYNMAN et al. 2001, FOWLES 1989,
al. (2008) the effect of the beam angle of inci- REES 1990, TRAGER 2007, WANDACHOWICZ
360 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

2000). Depending on the type of the material, was favourable from the point of view of the
the influence of each of these properties var- range and accuracy of the reflectorless meas-
ies with respect to the errors of the measured urements. Plexiglas had a specified degree of
distance. On the basis of the tests carried out light transmission. This parameter was un-
in LENDA & MARMOL (2010), it was observed known for the other materials. The tests as-
that the magnitude of these phenomena is af- sociated with different thicknesses of the syn-
fected mainly by factors such as the degree of thetic materials were performed for two mate-
light transmission, type of the synthetic ma- rials: with high density, i.e. for Plexiglas, and
terial (chemical composition), density, thick- with low density, i.e. for Styrofoam.
ness and the temperature of the material. The Temperature also influences the results of
colour of the material is of some importance the reflectorless distance measurements. All
as well, which, however, was not included in materials were tested at two temperatures: at
the previously conducted research. According about 15 °C and at about 40 °C. The tempera-
to these observations, a group of 16 materials ture of 15 °C (cold) was ambient temperature,
was selected for testing. The analysis of 13 of and the temperature of 40 °C (warm) was
these (Tab. 1), due to the similar properties of reached by heating the materials with a porta-
the materials and obtained results, turned out ble heater. The temperature measurement was
to be sufficient. carried out using a Testo 830-T2 pyrometer.
Plexiglas was the material with the big- The results of the reflectorless measure-
gest changes in measured distances during the ments are also affected by the incidence angle
range finder tests. Therefore, for this particu- of the beam on the target. To assess the sig-
lar synthetic material, the most detailed tests nificance of this factor, basic tests were per-
were planned, taking into account its colour, formed at the most favourable, normal inci-
thickness and light transmission. Other high dence of the laser beam on the target (90°). For
density materials included: HIPS, Polysty- comparison, measurements were also carried
rene and Polyolefin (rubber). Materials with out for the beam incidence at an angle of 45°.
low density were represented by foam syn- Materials with partial transparency placed at
thetic materials: Styrofoam and Polyurethane an angle increase the distance which the beam
foam. All tested materials, except for the col- can travel within the structure of the materi-
oured Plexiglas, were white or gray, which al. Potentially, this could affect the increase in

Tab. 1: The materials tested.

No. Material type Colour Thickness Transparency

(mm) (%)
1 Plexiglas (Polymethyl methacrylate) red 5 30
2 Plexiglas green 5 30
3 Plexiglas blue 5 30
4 Plexiglas white 5 30
5 Plexiglas white 5 70
6 Plexiglas black 5 30
7 Plexiglas white 10 30
8 HIPS (High impact polystyrene) white 5 -
9 Polystyrene white 5 -
10 Styrofoam (Extruded polystyrene foam) white 5 -
11 Styrofoam white 10 -
12 Polyurethane foam grey 10 -
13 Rubber (Polyolefin) grey 5 -
Grzegorz Lenda et al., Accuracy of Laser Scanners 361

the error values. On the other hand, due to the and then the distance between them was com-
partial reflection of the beam from the surface pared, taking into account the thickness of the
of the material placed at an angle, the pene- cardboard. The compared planes were created
tration of the interior of the material can be from two different point clouds, so they were
limited, which will reduce the value of the er- not exactly parallel. The determination of the
rors. For these reasons, it is difficult to find an distance between them was carried out as fol-
analogy with other studies at varying angles lows: The regression planes have been fitted
of the beam incidence, contained in the exist- in the reference cloud and the test cloud. The
ing literature. dense point grid with a resolution of 1 point
Since our previous studies of reflectorless per 3 mm² has been interpolated on the test
range finders presented the effect of a distance plane. Then the distances from all points of
on the obtained errors associated with greater the sample of the reference plane have been
penetration of the material by the laser beam on determined, along the normal to that plane.
shorter distances, tests of the scanners were Then the average distance has been calculated
carried out on the bases of three lengths: 5 m, and taken as the distance between the planes.
15 m and 50 m. Such a method of determining the distance
The procedure of determining the distance allowed eliminating inaccuracies associated
measurement deviations for each of the mate- with a slight rotation of both planes. The dif-
rials requires an explanation. We built a tar- ference was the distance measurement devia-
get of 20 cm × 30 cm for each material. Those tion of the synthetic material with reference to
targets were split into two halves of 20 cm × the white cardboard, and, consequently, to the
15 cm each (Fig. 1). For the cardboard, which reflective tape. The deviation of the distance
served as a reference, the distance deviation measurement of synthetic materials observed
in relation to the original Leica reflective tape in this way (ΔD) is not an overall error of dis-
target had been predetermined – it equals tance measurement. It is, however, an addi-
0.6 mm – taking into account the difference tional contribution, which should be added to
in thickness of the cardboard and tape target the standard errors of a laser scanner, assum-
(LEICA 2015). It was decided not to stick the ing reflection of a beam from the outer sur-
tape target directly on the tested synthetic ma- face of the material. In addition, the precision
terials, due to a possibility of it being trans- (mp) of fitting the regression plane into a set
mitted by the laser beam, which would en- of points representing the synthetic material,
able the penetration into the structure of the expressed by the RMS­error of the distances
synthetic material, distorting the results. The of the points from the regression plane was de-
tested synthetic materials adhered directly to termined.
the opaque background. The planar target was The tests were performed using two laser
scanned at a high resolution (2 mm density). scanners with similar accuracy, but with a
Points near the boundaries of the two halves different carrier wavelength: Leica C10 and
of the targets were removed from the cloud of Riegl VZ­400. Basic data of these scanners
points in order to obtain homogeneous sets of with reference to the subject in question were
points. Based on the filtered data, a regression presented in Tab. 2. The carrier wavelength is
plane was determined for each of the fields, important especially when measuring targets
of varying colours since spectral curves may
differ even in the case of very similar materi-
als (Toś 2014). The colour of the material may
influence the absorption of beams with differ-
ent wavelengths. It can also affect the meas-
urement accuracy of synthetic materials in a
varied manner, depending on the size of the
inhomogeneity in their structure.
In the production process it is not always
possible to obtain a uniform chemical compo-
Fig. 1: Experimental setup. sition and repetitive structure of synthetic ma-
362 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

Tab. 2: Selected parameters of the tested laser scanners.

Leica C­10 Riegl VZ­400

Type pulse-based pulse-based
Accuracy of length measurement 4 mm (up to 50 m) 5 mm (up to 100 m)
Range up to 350 m up to 600 m
Carrier wavelength 532 nm 1550 nm
Beam divergence 0.24 mrad 0.35 mrad

terials. For this reason, comparative tests were 3 Results of Test Measurements
performed for several materials (Plexiglas
white, 5 mm thickness and 30 % transparen- The results obtained from the conducted tests
cy (hereinafter briefly stated with x mm/Y %), are shown in Tab. 3 (Leica C­10 scanner) and
Polystyrene 5 mm, and Styrofoam 10 mm), Tab. 4 (Riegl VZ­400 scanner). Due to high-
taking into account the measurement of three ly ambiguous results (scattered data) of the
samples of the same production lot, as well measurements performed at the angle of 45°
as the measurement to three samples of un- at the distances of 15 m and 50 m, they are
known origin. The tests were performed using not presented here. The diagrams presented
the Leica C­10 scanner, at the baseline of 5 m, in Figs. 2 and 3 were drawn based on these
the distance between the laser scanner and the Tables. The measurement results of different
test material. samples of the same materials are included in

Tab. 3: Distance measurement deviations of the Leica C-10 scanner (ΔD = distance measurement
deviations relative to the comparative model, mp = precision of fitting the regression plane for a
synthetic material, tr. = transparency).
Grzegorz Lenda et al., Accuracy of Laser Scanners 363

Tab. 4: Distance measurement deviations for the Riegl VZ-400 scanner.

Tab. 5. During the tests with the VZ­400 scan- Tab. 5: Distance measurement deviations of
ner, it was noticed that synthetic materials the samples for selected materials: A1–A3 –
were not directly adhering to the background, sample material from the same production lot,
B1–B3 – sample material from unknown pro-
but they stood out approximately 2 cm – 3 cm duction lots, Leica C-10 scanner, distance 5 m.
behind it. This resulted in a significant deteri-
oration in the distance measurement deviation Plexiglas Polystyrene, Styrofoam
for some materials. The measurements were white, 5 mm, 5 mm 10 mm
repeated and, for comparison, Tab. 6 shows 30%
the results for the background adhering and transparency
(mm) (mm) (mm)
not adhering to the synthetic material.
The precision of fitting the regression plane A1 8.5 4.8 3.3
for the reference cardboard was similar for A2 8.1 4.6 3.9
both scanners and it was ± 1.2 mm for 90° A3 8.8 5.1 4.2
beam incidence angles, and ± 0.8 mm for the
45° incidence angles, with the scatter for these B1 8.5 4.8 3.3
values reaching a maximum of ± 0.2 mm. B2 6.9 4.7 5.2
B3 11.8 6.4 2.8
364 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

Fig. 2: Distance measurement deviations for the materials 1–6, ΔD – distance measurement de-
viations relative to the comparative model.
Grzegorz Lenda et al., Accuracy of Laser Scanners 365

Fig. 3: Distance measurement deviations for the materials 7–13, ΔD – distance measurement
deviations relative to the comparative model.
366 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

Tab. 6: Comparison of the distance measure- 4 Analysis of the Results

ment deviations of the VZ-400 scanner for
various locations of the synthetic material The assessment of the results was divided into
background. Test 1: The background adhering
to the rear surface of the synthetic material. several categories, resulting from the prop-
Test 2: The background placed approximately erties of the material itself, the conditions in
2 cm – 3 cm behind the synthetic material. which it was found, and the tested instrument.
Comparison for the 5 m distance (tr. – trans- The analyses included the following catego-
parency).
ries: type of material (light transmission, den-
sity and internal structure, colour, thickness),
distance, angle of incidence, temperature, type
of scanner (carrier wavelength), and distance
of the synthetic material from the background.
Some phenomena were especially noticeable
for the shortest measurement distance (5 m), at
low temperature of the material (15 °C), which
is explained later in the text. Therefore, dur-
ing the comparisons, we will frequently refer
to the results obtained under such conditions
as the “basic” conditions.

4.1 Light Transmission

This is the primary factor affecting the dis-

tance measurement of synthetic materials. If
the laser beam is not able to penetrate into the
structure of the synthetic material, there are
no changes in the measured distance in com-
parison with the reference cardboard. Such a
situation can be observed only for one of the
tested materials, i.e. opaque rubber (Polyole-
fin). Regardless of the conditions, the distance
measurement error for this material did not
exceed the value of 2 mm, usually remaining
at the level of about 1 mm, unlike other mate-
rials, where a certain degree of transparency
allowed for the occurrence of additional phe-
nomena affecting the distance measurement.
The changes in the distance observed for them
ranged from about 2 mm to 15 mm. The influ-
ence of transmission was clearly visible when
comparing the two samples of Plexiglas white,
with a transmission of 30 % and 70 %. Mean
deviations for the material with lower trans-
mission were approximately 7 mm (C10) and
8 mm (VZ400), and 12 mm for larger trans-
mission (C10 and VZ400). For both scanners,
taking into account comparable materials
(5 mm thickness, white colour), the largest de-
viations occurred for Plexiglas, which had a
significant light transmission. Other materials
were measured with mean deviations within
Grzegorz Lenda et al., Accuracy of Laser Scanners 367

the range of 1 mm – 3 mm (C10) and 1 mm – dex will be subject to changes. Comparing the
5 mm (VZ400). For the C10 scanner, Plexiglas measurements results performed with the C10
with 70 % transmission caused a slight dete- scanner to various samples of materials from
rioration in the precision of fitting the regres- the same production lot (Tab. 5), a stable level
sion plane relative to other materials (up to ca. of deviations for Plexiglas white 5 mm / 30 %
± 2 mm). For the VZ400 scanner, the fitting – (8.1 mm – 8.8 mm) and Polystyrene 5 mm /
precision for this material decreased in some 30 % – (4.6 mm – 5.1 mm) is noticeable. For
cases to ca. ± 3 mm. the foamed material, 10 mm Styrofoam, the
relative differences are larger – (3.3 mm –
4.2 mm). When comparing samples of these
4.2 Density and Internal Structure materials originating from different produc-
tion lots, possibly from different manufactur-
The studies covered two groups of materials ers, the results look somewhat different. Dif-
with large differences in density and inter- ferences for 5 mm Polystyrene increased to
nal structure. The first one included materials 4.7 mm – 6.4 mm. For other materials, the rel-
with a continuous structure (Plexiglas, HIPS, ative differences were nearly doubled: 10 mm
Polystyrene, Polyolefin), and the second one Styrofoam: 2.8 mm – 5.2 mm, Plexiglas white
included foamed materials (Styrofoam, Poly- 5 mm/30 %: 6.9 mm – 11.8 mm. Therefore,
urethane Foam). In the foamed materials the the type of a material and its internal structure
beam was subject to changes, alternately in have a significant impact on the level of the
fragments of the material and in air cham- measurement deviations. At the same time,
bers. Foamed materials were measured with this influence is difficult to determine accu-
smaller deviations in relation to the thickness rately due to variations in the production.
of the material than their continuous counter-
parts. This could be assessed by comparing
the measurement deviations of Polystyrene 4.3 Colour
and Polystyrene Foam (Styrofoam) (5 mm) in
basic conditions: Polystyrene: 4.8 mm (C10), The colour of the synthetic material signifi-
6.3 mm (VZ400) and Styrofoam: 2.1 mm cantly influences the measurement deviations.
(C10), 2.9 mm (VZ400). The foamed materi- A sample Plexiglas was tested under basic
al achieved deviations of approximately half conditions using the same parameters regard-
that size. It is interesting that for some mate- ing thickness and permeability. Then, the de-
rials, especially for Plexiglas, the measure- viations resulting from the changing colour
ment deviations were larger than the thickness of the material were within 4.2 mm – 9.1 mm
of the material itself. Examples under basic (C10) and 7.1 mm – 10.5 mm (VZ400). For
conditions were Plexiglas white 5 mm / 30 % both scanners, the differences between the ex-
– 8.5 mm (C10), 9.4 mm (VZ400), Plexiglas treme values of the deviations had a factor of
white 5 mm / 70 % – 13.7 mm (C10), 15.2 mm 1.5 to 2. As only a few selected colours were
(VZ400). During the tests, the materials ad- tested, a discrepancy may be even larger un-
hered directly to the opaque background, so der real conditions. The colour of the material
that the beam, having passed through the ma- strongly influenced the absorption of beams
terial, did not have a possibility of incidence with different wavelengths. Due to the fact
on other targets. The values of the deviations, that two scanners emitting laser radiation of
therefore, must have resulted from additional different carrier wavelengths, 532 nm (C 10)
phenomena associated with the beam propa- and 1550 nm (VZ400), were selected for this
gation in the structure of the material. They experiment, the differences in the values of
might be related, among others, to the repeated the deviations for the same colours were clear-
refraction or the reflection of a beam. In addi- ly noticeable. Under basic conditions, for the
tion, the refractive index depends on the inter- C10 scanner, the smallest deviations were ob-
nal structure of materials. If it is not homoge- served for the following colours: blue: 4.2 mm,
neous, and the inhomogeneity has dimensions black: 4.4 mm, and red: 5.3 mm. It is interest-
greater than the wavelength, the refractive in- ing that, at the same time, for those colours
368 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

the lowest precision of fitting a regression be recognized with other synthetic materials
plane were obtained: blue: ± 2.8 mm, black: measured with the VZ400 scanner, e.g. HIPS
± 3.0 mm, red: ± 3.9 mm. Larger deviations and Polystyrene. The dependencies related
were observed for white (8.5 mm) and green to the thickness of the synthetic material are
(9.1 mm), with better precision of fitting the therefore significant, nonlinear, and for better
plane: white ± 1.2 mm, green ± 1.4 mm. The assessment it would be advisable to carry out
VZ400 scanner demonstrated the smallest de- more extensive tests.
viations for red (7.1 mm) and green (7.5 mm),
and the largest for the blue (10.2 mm) and
black (10.5 mm) colours. Therefore, for green, 4.5 Distance
blue and black, the dependencies observed
for the C10 scanner were inverted. In basic Dependencies resulting from the variable dis-
conditions, the precision of fitting the regres- tance of the target are easiest to evaluate from
sion plane were at a low level for all colours the diagrams in Figs. 2 and 3. For most materi-
(± 1.0 mm – ± 1.6 mm). For the test configu- als, the deviations are reduced with increasing
rations other than the basic ones, the devia- distance. The differences between the results
tions for red and blue increased. The colour of for the distances of 5 m and 15 m do not exhib-
the synthetic material in combination with its it this phenomenon clearly, i.e. sometimes the
light transmittance is therefore a very impor- deviations even increase. However, the differ-
tant factor determining the value of the dis- ences between the results at the distances of
tance measurement deviations. 5 m and 50 m exhibit a certain tendency and
reach the values of up to 3.5 mm. The increas-
ing tendency of the deviations was observed
4.4 Thickness for two foamed materials: Styrofoam (10 mm)
and Polyurethane Foam for the VZ400 scan-
The effect of the thickness of the synthetic ma- ner. Deviations decreasing with distance are
terial was assessed based on two samples of contrary to current experiments for reflector-
materials (Plexiglas 30 %, 5 mm and 10 mm, less instruments. Moreover, the equipment
and Styrofoam, 5 mm and 10 mm). Under basic manufacturers themselves inform about in-
conditions, the measurement deviations using creasing deviations with increasing distance.
the C10 scanner for Plexiglas ranged from However, this phenomenon can be explained
8.5 mm to 11.6 mm, and for the VZ400 scan- by a reduced ability of radiation to penetrate
ner from 9.4 mm to 13.3 mm. For Styrofoam, the material which, at the larger distance, has
the deviations for the C10 scanner ranged from a lower energy density at the surface of the
2.1 mm to 3.3 mm, and for the VZ400 scanner synthetic material. In this way, the scale of the
from 2.9 to 5.2 mm. The deviations were dif- phenomena described above is limited. Due
ferent for each sample and scanner, and were to the fact that the largest deviations were ob-
within the limits of 36 % – 79 %, at 100 % served at short distances, the 5-meter baseline
change in thickness of the material. Greater was set as the “basic” condition. With the in-
differences were observed for the foamed ma- creasing distance there was no significant in-
terial of low density, i.e. for Styrofoam 57 % crease in the precision of fitting the regression
– 79 %, while for Plexiglas they amounted to plane.
36 % – 41 %. As Styrofoam in practice had a
thickness greater than 5 mm or 10 mm, this
could translate into relatively large measure- 4.6 Incidence Angle
ment deviations. In view of the fact that colour
has also a strong influence on the deviations of Studies related to the varying incidence angles
the measurements on Plexiglas the deviations of the beam yielded inconclusive results. The
could reach values higher than the thickness tests were carried out at the angle of 45° with-
of the material itself. This happened in more out the materials being heated. The results for
than half of the cases for the C10 scanner and the distances of 15 m and 50 m did not demon-
all of the cases for the VZ400. This could also strate any systematic order, the measurement
Grzegorz Lenda et al., Accuracy of Laser Scanners 369

deviations just increased or decreased in a deviations of fitting the regression plane from
random manner. This was associated with sig- about ± 1.2 mm to ± 0.7 mm on an average.
nificant decrease of precision of fitting the re- Similar observations were made in VOEGT-
gression planes with up to several millimetres. LE et al. (2008), where it was also found that
The ambiguous test results at greater dis- with a decreasing incidence angle, the devia-
tances could have resulted from the variable tion decreased. However, in SOUDARISSANANE
reaction of fragments of the materials to the et al. (2007), for some materials an opposite
illumination at an angle. At a greater distance tendency was observed. It was also confirmed
where the density of the radiation energy at by the previous research of the authors associ-
the sample surface is smaller, the non-homo- ated with the tests of reflectorless range find-
geneity of the material could have determined ers (LENDA & MARMOL 2010). Similar relations
the phenomenon of partial reflection from were observed for the instrument VZ400,
the surface of the inclined sample, changing where the deviations decreased by more than
the degree of the material penetration by the 4 mm. However, for some materials (Plexiglas
beam. This is particularly visible for the blue green and blue, 10 mm Styrofoam), a slight,
Plexiglas measured with the C10 scanner at a about 1 mm increase in the deviations oc-
distance of 15 m (Fig. 4). The left figure pre- curred. The precision of fitting the regression
sents a sample scanned at the angle of 90°, plane were also subject to a decrease an av-
and the right one at the angle of 45°. Signifi- erage of about ± 1.2 mm to ± 1.0 mm. In our
cant losses in the point cloud (front view, right experiments, at the angle of incidence of 45°,
side) together with highly scattered individual the materials Plexiglas red and black were im-
points (top view, right side) are visible in the measurable for the Leica scanner. Based on
latter case. the obtained results and the results of other re-
For this reason, only the results for the searchers, it was apparent that the angles of
distance of 5 m are presented, because they incidence affect the measured distance. The
seemed not to be subject to blunders. With the tests exhibited a decrease in the deviations
decreasing angle of the beam incidence (from at more acute angles of the beam incidence –
the standard 90° to 45°), for the C10 scanner, which may be the result of reduced penetra-
a decreasing measurement deviation of up to tion by laser beam due to partial reflection
about 3 mm in almost all the cases was ob- from the surface of the sample at incidence
served. This was related to the decrease in the angle of 45°.

Fig. 4: Point clouds obtained from blue Plexiglas with the C10 laser scanner at 15 m distance.
Sample scanned at 90° incidence angle on the left and 45° incidence angle on the right; above:
front views of the materials, below: top views of the materials.
370 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

4.7 Temperature of the Material transmitted through the material, can be re-
flected by the background, causing significant
The influence of the temperature is easiest to changes to the obtained results, of up to sev-
assess from the diagrams contained in Figs. eral centimetres. For some synthetic materi-
2 and 3. A decrease in the deviations for the als it does not really matter (HIPS, Polysty-
non-foamed materials (except rubber) relat- rene, Styrofoam, Polyolefin), for others these
ed to the temperature growth is quite clear, changes may be small (Polyurethane foam).
as well as its lack for the foamed materials. For some of the materials (Plexiglas, espe-
Non-foamed synthetic materials, when heat- cially with small thickness) it may, however,
ed, generate deviations that are smaller by result in very significant discrepancies. The
about 1 mm – 3 mm (C10) and by about 1 mm observed increase in the deviations for dif-
– 4 mm (VZ400) than cold ones. Continu- ferent colours of 5 mm Plexiglas, under ba-
ous materials, having no air bubbles in their sic conditions, ranged from 71 % (white) to as
structures, exhibited susceptibility to heat. much as 345 % (red), reaching the values of
This phenomenon could be explained by the up to 32 mm. It was combined with a slight
thermal expansion of materials and the related increase in the plane fitting precision. The de-
lower density of synthetic materials, reducing pendencies associated with the colours were
the refractive index value. However, since the quite different from the ones observed for the
change in the volume of the synthetic materi- background adhering to the synthetic materi-
als when heated was negligible, other factors al. The loss of accuracy related to a complete
related to temperature, unnoticed by the au- passage of light through the synthetic mate-
thors, must have affected the reduction of the rial may thus result in obtaining significant,
measurement deviations. Due to the fact that unpredictable values. Further studies, taking
larger deviations were observed for cold tar- into account various distances between the
gets, low temperature was adopted as another synthetic material and the background, may
“basic” conditions. foster interesting results.

4.8 Scanner Type 5 Summary

In this category, the carrier wavelength of the The paper analyses a number of factors affect-
laser, different for the two scanners (532 nm ing the distance measurement of synthetic ma-
(C10) and 1550 nm (VZ400)) should be distin- terials. The most important of these is optical
guished. The light of various wavelengths is permeability, which induces other phenomena
absorbed in a different manner by the colours occurring within the material structure. Some
and materials of a different size and ordering of them may have a significant impact on the
of the molecules. Taking into account all 13 obtained results, e.g. degree of permeability,
tested materials, almost all of the tests demon- type of material, colour in combination with a
strated slightly higher deviation values for the carrier wavelength of the laser radiation, and
VZ400 scanner. This dependency could have thickness. The effects of other factors seem
been expected when comparing the precision to be less significant, but noticeable and sys-
parameters of both instruments. tematic such as distance and temperature of
the material. The angle of the beam incidence
is also relevant. Synthetic elements may there-
4.9 Distance between Synthetic fore exhibit shifts of several, sometimes over
Material and the Background a dozen millimetres in the point cloud relative
to other objects. Especially when they not ad-
The location of the background of the tested here to the other materials, shifts in some cases
material is of great importance to the meas- (Plexiglas) may increase to several centimetres.
ured distance of some synthetic materials. In some cases, measurement deviation of the
If it is close, just a few centimetres from the synthetic material can affect the accuracy of
synthetic material, a laser beam, after being the entire scanned structure. This may happen
Grzegorz Lenda et al., Accuracy of Laser Scanners 371

when spatial markers in the form of Styrofoam mation Sciences, Commission V Symposium 38
balls are used to connect neighbouring scans. (5): 80–85, Newcastle, United Kingdom.
As it was demonstrated by our experiments, Boehler, W., Bordas, v.m. & m arBs, a., 2003: In-
the error for Styrofoam increases significant- vestigating laser scanner accuracy. – 19th CIPA
International Symposium, Antalya, Turkey.
ly with its thickness, and therefore covering
Bucksch, a., lindenBergh, r. & van r ee, J., 2007:
such balls with opaque lacquer should be con- Error budget of terrestrial laserscanning: Influ-
sidered. Some of the tests yielded inconclu- ence of the intensity remission of the scan qual-
sive or questionable results (incidence angle), ity. – III International Scientific Congress Geo­
which may have resulted from the methodol- Siberia, Novosibirsk, Russia.
ogy, dependent upon the occurrence of addi- clark, J. & roBson, s., 2004: Accuracy of mea-
tional phenomena, which the authors had not surements made with a CYRAX 2500 laser
anticipated. The factors that interfere with the scanner against surfaces of known color. – Sur-
results of the research studies may include for vey Review 37 (294): 626–638.
example: additional lighting, reflections of the EREN, G., AUBRETON, O., MERIAUDEAU, F., SECADES,
beam from the surrounding elements, type S., FOFI, D., NASKALI, A.T., TRUCHETET, F. & ER-
CIL, A., 2009: A 3D Scanner for Transparent
of the synthetic material background and the
Glass. Image Analysis and Processing. – Inter-
manner of its adherence to the sample. The national Conference on Image Analysis and Pro-
deviations resulting from a complete passage cessing (ICIAP) 2009, Lecture Notes in Com-
of light through the synthetic material and the puter Science (LNCS) 5716: 519–527.
reflection of a beam from the background lo- Feynman, r.P., leighton, r.B. & sands, m., 2001:
cated at a certain distance from the material The Feynman Lectures on Physics. – PWN,
may reach significant values. In general, due Warszawa, Poland.
to a large number of factors affecting the dis- FOWLES, G.R., 1989: Introduction to Modern Op-
tance measurements and their mutual con- tics. – Dover Publications, Mineola, NY, USA.
figurations, it is difficult to tabulate distance GOTTWALD, R., 2008: Field Procedures for Testing
Terrestrial Laser Scanners (TLS) – A Contribu-
measurement deviations for specific synthetic
tion to a Future ISO Standard. – Integrating
materials. The situation is made worse by an Generations, FIG Working Week 2008: 1–14,
average repeatability of their production. This Stockholm, Sweden.
study draws attention to the problem, allowing hiremagalur, J., yen, k., a kin, k., Bui, t., lasky,
estimating the scale of the existing dependen- t. & r avani, B., 2007: Creating standards and
cies. specifications for the use of laser scanning in
caltrans projects. – AHMCT (Advanced High-
way Maintenance & Construction Technology
Acknowledgements Research Center) Research Report, University
of California at Davis, California Department of
Transportation, Davis, CA, USA.
This work has been made within scien-
HOLST, C., ZEIMETZ, P., NOTHNAGEL, A., SCHAUERTE,
tific research program no. 11.11.150.005, W. & kuhlmann, H., 2012: Estimation of Focal
11.11.150.949. Length Variations of a 100­m Radio Telescope’s
Main Reflector by Laser Scanner Measure-
ments. – Journal of Surveying Engineering 138
References (3): 126–135.
Johnson, W. & Johnson, a., 2012: Operational Con-
a BBas, m.a., setan, h., m aJid, Z., chong, a.k., siderations for Terrestrial Laser Scanner Use in
idris, k.m. & asPuri, a., 2013: Calibration and Highway Construction Applications. – Journal
Accuracy Assessment of Leica ScanStation C10 of Surveying Engineering 138 (4): 214–222.
Terrestrial Laser Scanner. Developments in k aasalainen, s., Jaakkola, a., k aasalainen, m.,
Multidimensional Spatial Data Models. – Lec- k rooks, a. & kukko, a., 2011: Analysis of Inci-
ture Notes in Geoinformation and Cartography dence Angle and Distance Effects on Terrestrial
10: 33–47. Laser Scanner Intensity: Search for Correction
Berenyi, a., lovas, t. & Barsi, a., 2010: Terres- Methods. – Remote Sensing 3 (10): 2207–2221.
trial laser scanning – civil engineering applica- k ersten, t.P., mechelke, k., lindstaedt, m. &
tions. – International Archives of the Photo- STERNBERG, H., 2009: Methods for Geometric
grammetry, Remote Sensing and Spatial Infor- Accuracy Investigations of Terrestrial Laser
372 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

Scanning Systems. – PFG – Photogrammetrie, Influencing factor on the quality of terrestrial

Fernerkundung, Geoinformation 2009 (4): 301– laser scanning points. – ISPRS Journal of Photo-
315. grammetry and Remote Sensing 66 (4): 389–399.
lee, i.s., lee, J.o., Park, h.J. & Bae, k.h., 2010: toś, c., 2014: The possibilities of using the terres-
Investigations into the influence of object char- tial scanning data for classification of rocks in
acteristics on the quality of terrestrial laser scan- limestone mine ‘Czatkowice’. – Reports on Ge-
ner data. – KSCE Journal of Civil Engineering odesy and Geoinformatics 97: 80–90.
14 (6): 905–913. trager, F., 2007: Springer Handbook of Lasers and
LEICA, 2015: http://www.allenprecision.com/ Optics. – Springer.
survey/robotic-total-stations/accessories/ voegtle, t., schWaB, i. & landes, t., 2008: Influ-
prisms/60x60mm­reflective­tape­20­pack/ ence of different materials on the measurements
(19.7.2015). of terrestrial laser scanner. – The International
lenda, g. & m armol, u., 2010: Accuracy of reflec- Archives of the Photogrammetry, Remote Sens-
torless range finders for the measurements of ing and Spatial Information Sciences 37 (B5):
objects made form synthetic materials (polish). 1061–1066.
– Pomiary Automatyka Kontrola 11: 1280–1286. WANDACHOWICZ, K., 2000: Wyznaczanie rozkładu
lichti, d.d. & Jamtsho, s., 2006: Angular resolu- luminancji we wnętrzach z uwzględnieniem
tion of terrestrial laser scanners. – The Photo- kierunkowo-rozproszonych charakterystyk od-
grammetric Record 21 (114): 141–160. biciowych materiałów (Determination of the lu-
monserrat, o. & crosetto, m., 2008: Deformation minance distribution in the interiors taking into
measurement using terrestrial laser scanning account the characteristics of the directional-
data and least squares 3D surface matching. – ­diffuse reflectance materials). – doctoral disser-
ISPRS Journal of Phtotogrammetry & Remote tation, Poznań, Poland.
Sensing 63: 142–154. Zámečníková, m., neuner, h. & PegritZ, s., 2014a:
PFeiFer, n., höFle, B., Briese, c., rutZinger, m. & Influence of the Incidence Angle on the Reflec-
H ARING, A. 2008: Analysis of the backscattered torless Distance Measurement in Close Range.
energy in terrestrial laser scanning data. – Inter- – INGEO 2014 – 6th International Conference
national Archives of the Photogrammetry and on Engineering Surveying: 257–262, Prague,
Remote Sensing and Spatial Information Sci- Czech Republic.
ence (IAPRS) 37 (B5): 1045–1051. Zámečníková, m., Wieser, a., WoschitZ, h. &
Polo, m.e., Felicísimo, a.m., villanueva, a.g. & r essl, c., 2014b: Influence of surface reflectivi-
m artíneZ-del-PoZo, J.a., 2012: Estimating the ty on reflectorless electronic distance measure-
Uncertainty of Terrestrial Laser Scanner Meas- ment and terrestrial laser scanning. – Journal of
urements. – IEEE Transactions on Geoscience Applied Geodesy 8 (4): 311–325.
and Remote Sensing 50 (11): 4804–4808.
R EES, W., 1990: Physical principles of remote sens-
ing. – Cambridge University Press, UK.
salemi, g., achilli, v. & Boatt, g., 2008: 3D vir- Addresses of the Authors
tual modelling of a gothic stained-glass panel. –
GRZEGORZ LENDA, Ph.D. & URSZULA M ARMOL,
The International Archives of the Photogram-
Ph.D., AGH University of Science and Technology,
metry, Remote Sensing and Spatial Information
Faculty of Mining Surveying and Environmental
Sciences 37 (B5): 297–302.
Engineering, Cracow, Mickiewicza 30, Poland,
soudarissanane, s., van r ee, J., Bucksch, a. &
Tel.: +48-12-617-23-14, -26, Fax: +48-12-617-50-76,
lindenBergh, r., 2007: Error budget of terres-
e-mail: {grzenda}{entice}@agh.edu.pl
trial laser scanning: influence of the incidence
angle on the scan quality. – 3D­NordOst, Berlin. GRZEGORZ MIREK, Ph.D., Cracow University of
SOUDARISSANANE, S., LINDENBERGH, R., MENENTI, M. Technology, Faculty of Environmental Engineer-
& teunissen, P.J.g., 2009: Incidence angle influ- ing, Cracow, Warszawska 24, Poland, Tel.: +48­12­
ence on the quality of terrestrial laser scanning 628-21-87, e-mail: gmirek@pk.edu.pl
points. – ISPRS International Archives of the
Photogrammetry, Remote Sensing and Spatial
Information Sciences 38-3/W8: 183–188.
SOUDARISSANANE, S., LINDENBERGH, R., MENENTI, Manuskript eingereicht: Februar 2015
m., teunissen, P.J.g., 2011: Scanning geometry: Angenommen: Juli 2015