accuracy of laser scanners for measuring surfaces made of synthetic materials · 2017. 3. 8. ·...

PFG 2015 / 5, 0357–0372 ArticleStuttgart, October 2015

© 2015 E. Schweizerbart'sche Verlagsbuchhandlung, Stuttgart, Germany www.schweizerbart.deDOI: 10.1127/pfg/2015/0273 1432-8364/15/0273 $ 4.00

Accuracy of Laser Scanners for Measuring Surfacesmade of Synthetic Materials

GRZEGORZ LENDA, URSZULA MARMOL & GRZEGORZ MIREK, Cracow, Poland

Keywords: terrestrial lidar, measurement error, material tests

permeability, which may vary depending onthe structure of the synthetic material. Exist-ence of this phenomenon may be proved bythe following experiment. One can measurea distance to a synthetic element, for exam-ple 2 mm – 3 mm thick, consisting of white

1 Introduction

Synthetic materials (materials consisting ofsynthetic polymers) are a group of materi-als which, when observed using laser scan-ners, may cause additional measurement er-rors. These errors result from the optical

Summary: Synthetics are a group of materialswhich, when observed using laser scanners, maycause additional measurement errors. These errorsresult from optical permeability, which may varydepending on the structure of the synthetic materi-al. A laser beam which penetrates the material canbe subject to numerous physical phenomena such asabsorption, dispersion, full or partial reflection, re-fraction at the interface between two media, dif-fraction and interference that change the measureddistance. As a result, additional measurement er-rors occur, independent of the known errors, whenthe beam is reflected from the outer surface of thematerial. Their occurrence is influenced by lighttransmission, density and internal structure of thesynthetic material, its colour and thickness. Thedistance from the scanner, laser beam angle of inci-dence, temperature, scanner type (carrier wave-length), or even the distance between the syntheticmaterial and its background are important as well.This study is devoted to examining these depend-encies. Currently there are no studies distinguish-ing and examining synthetic materials as a separategroup of materials that may interfere with distancemeasurements due to the phenomena that occur intheir structure.

Zusammenfassung: Genauigkeit von Laser-Scan-nern bei der Oberflächenmessung von syntheti-schen Materialien. Synthetische Materialien kön-nen bei der Erfassung von Oberflächen mit Laser-scannern zu zusätzlichen Messfehlern führen.Diese Fehler ergeben sich aus der optischen Durch-lässigkeit, die in Abhängigkeit von der Struktur dessynthetischen Materials variieren kann. Ein Laser-strahl, der in das Material eindringt, kann mehre-ren physikalischen Phänomenen, z. B. Absorption,Streuung, vollständige oder teilweise Reflexion,Refraktion an Grenzflächen zwischen zwei Medien, Beugung oder Überlagerung, unterworfensein, welche die gemessene Distanz verändern. AlsErgebnis erhält man zusätzliche und von bisher be-kannten Fehlerursachen unabhängige Messfehler,wenn der Strahl an einer aus synthetischem Mate-rial bestehenden Oberfläche reflektiert wird. Diesewerden von der Lichtdurchlässigkeit, der Dichteund der inneren Struktur des synthetischen Mate-rials beeinflusst. Die Distanz vom Scanner, derAuftreffwinkel des Laserstrahls, die Temperatur,der Scannertyp (Trägerfrequenz des Lasers) oderauch die Distanz der gescannten Oberfläche vonObjekten, die sich aus Sicht des Sensors im Hinter-grund befinden, sind ebenfalls von Bedeutung.Dieser Beitrag widmet sich der Untersuchung die-ser Einflüsse. Bisher gibt es keine Studien, welchesynthetische Materialen als eigene Gruppe vonMa-terialen unterscheiden, welche in spezifischer Wei-se die Distanzmessung auf Grund ihrer innerenStruktur beeinflussen.

358 Photogrammetrie • Fernerkundung • Geoinformation 5/2015

midity, temperature), and the properties of thescanned surface.Considering the properties of the materi-

al, which this article focuses on, the researchwork carried out by BOEHLER et al. (2003) isworth mentioning. Their paper includes a se-ries of tests for checking the quantitative andqualitative accuracy of points measured us-ing nine laser scanners selected from differ-ent manufacturers. Among other things, therelation between surfaces of varying reflectiv-ity and the accuracy of the acquired informa-tion was tested. The following samples madeof various materials and in different colourswere used: white paint (90% and 80% reflec-tive), grey paint (40% reflective), black paint(8% reflective), metallic paint, various kindsof films. Measurements of different coloursurfaces yielded good accuracy results. Forthe white surfaces, a deviation of zero wasobtained. Also, the grey and black coloursbrought zero errors for most scanners. The sit-uation was much worse in the case of differ-ent types of films, where errors from a few upto tens of millimetres occurred. Contradictoryresults were shown in BUCKSCH et al. (2007),where the bright surfaces gave much greateraccuracy than the dark ones. The problem as-sociated with material reflectance has beenanalysed in CLARK & ROBSON (2004). The au-thors selected materials of standardized col-ours and textures, and scan them from vari-ous distances (4 m – 24 m) at the angles of20°, 40° and 60°. The experimental resultsrevealed systematic discrepancies in the re-corded distances, depending on the type ofthe scanned surface. Colours which exhibitlow reflectance, such as black or red, lead tolonger distance measurements than the refer-ence distance. The authors proposed a correc-tion factor, reducing these systematic errors.A similar problem was presented in PFEIFER etal. (2008), where the authors focused on thesignal energy returning to the receiver. It washypothesized that this energy can be deter-mined based on the observed range and inten-sity value. It was proven that it was possibleto reproduce the reflectance of a given surfacewith the accuracy at the level of 6%. Surfacereflectance is also dealt with in Zámečníkováet al. (2014b). The authors used the standardwhite colour (spectralon), which is the mate-

Plexiglas or polystyrene, using any type of re-flectorless range finder, then measure it againwithout changing the position of the elementbut with an identical element adhered to itsback surface. The observed distance will begreater due to laser beam penetration throughadded layers of material. On the other hand,covering the front surface with a sheet of pa-per will prevent the laser beam from penetrat-ing and result in a shorter measured distance.A laser beam penetrating the material may ini-tiate a number of physical phenomena, chang-ing the measured distance. As a result, addi-tional measurement errors occur, independentof the already familiar errors that occur whenthe beam is reflected from the outer surfaceof the material. Even cursory tests of select-ed materials demonstrate that in some casesthese errors can appear much bigger than theassumed accuracy of the scanner. Due to thegrowing use of synthetic elements in variousbranches of technology and construction, aswell as the high prevalence of scanning tech-niques for the measurements of all types ofstructures (BERENYI et al. 2010, HIREMAGALURet al. 2007, HOLST et al. 2012, JOHNSON& JOHN-SON 2012, MONSERRAT & CROSETTO 2008, SA-LEMI et al. 2008), a need arises to investigatethe influence of these materials on errors ofmeasured distances.In the literature on examining the accuracy

of laser scanners there have been no studiesthat address measurements to synthetic mate-rials, although some of them indicate poten-tial problems associated with measuring ma-terials having a certain degree of transparency(EREN et al. 2009, SALEMI et al. 2008, VOEGTLEet al. 2008). Generally, on the basis of sever-al laser scanning studies (ABBAS et al. 2013,EREN et al. 2009, GOTTWALD 2008, KAASALAI-NEN et al. 2011, KERSTEN et al. 2009, LEE et al.2010, LICHTI& JAMTSHO 2006, POLO et al. 2012,SOUDARISSANANE et al. 2007) and those givenbelow, it is known that the measurement ac-curacy is affected by several factors, includ-ing: the type of a scanner (pulse-based, phase-based), the precision of a scanning mechanism(laser footprint, accuracy and resolution of thehorizontal and vertical angular encoders, theeccentricity of the scanning system), scanninggeometry (beam angle of incidence, scanningdistance), external conditions (lighting, hu-

Grzegorz Lenda et al., Accuracy of Laser Scanners 359

dence on the resulting accuracy was analyzed.Considering the values of the mean squareerrors, the following relation was noted: forthe perpendicular direction of the beam inci-dence the errors are significantly larger thanfor smaller angles. A metal plate (aluminium,high reflectance) also appears in Zámečníkováet al. (2014b). The authors noticed that smallchanges in the scanning angle (0.1°) result in20% changes in the signal strength, and thusto changes in the recorded deviations. The re-sults presented in that paper confirmed that ahigher reflectance results in larger errors forshort distances, and for the distances exceed-ing 20 m their value decreases.In conclusion, it can be said that the research

hitherto conducted mainly analyzed the errorscaused by the reflection of a laser beam fromthe outer surface of a material. Synthetic ma-terials were neither tested nor distinguishedas a separate group of materials which mightdistort the measurements of a distance due totheir internal structure. Therefore, this workpresents error analysis for distance measure-ments of the elements of this type.The authors initially studied measurements

to synthetic materials in LENDA & MARMOL

(2010) by testing the distance measurementdeviations of reflectorless range finders fora few selected materials. It enabled the firstconclusions on the factors influencing the dis-tance measurement errors of synthetic materi-als, as well as to plan appropriate methodolo-gy for the implementation of research studies,taking into account a wider range of materialsfor laser scanners. The studies carried out inthis paper present a wide range of issues. Fromthe point of view of the users, this paper al-lows for a sufficient assessment of the impactof synthetic materials on the laser scanners’distance measurements.

2 Research Methodology

A laser beam in the structure of a syntheticmaterial may be subject to a series of phe-nomena, such as absorption, dispersion, fullor partial reflection, diffraction and interfer-ence, and refraction at the interface betweentwo media (FEYNMAN et al. 2001, FOWLES 1989,REES 1990, TRAGER 2007, WANDACHOWICZ

rial of almost perfect Lambertian reflectance,as well as cardboard boxes in varied colours.The materials were moved on a specially con-structed trolley in the range of 1.1 m to 29.7 mat every 7.5 cm. The authors found that theobtained deviations depend not only on therange, but also on the signal strength.Another study of the properties of the

scanned objects was performed by VOEGTLE etal. (2008). This study used materials common-ly applied on the façades of buildings, hav-ing different properties: colourful sheets andshades of gray, different kinds of wood, metalelements, plaster of various granularity, trans-parent films and materials with varying de-grees of moisture. The authors’ attention wasdrawn to the tests of partly transparent materi-als: errors ranged from 15 mm (for 35% trans-parency) to 34 mm (for 5% transparency).It is not possible to analyse properties of the

scanned material separately from the scan-ning geometry. Both the angle of the beamincidence and the distance have a significanteffect on the resulting accuracy. In SOUDARIS-SANANE et al. (2007) it was found that themeasurement noise for a fibreboard and ply-wood painted white increases with increas-ing angles. The effect of the scanning angleon the resulting accuracy was also analysedin SOUDARISSANANE et al. (2009) and SOUDARIS-SANANE et al. (2011). An increase in the meas-urement noise with the increasing scanningangle was identified. The authors proposed amodel which optimized the point position andcorrected the influence of the angle. Therewas a significant improvement in the stand-ard deviation, from the value of 3.25 mm to2.55 mm. The influence of the scanning an-gle on the final accuracy was also presentedin Zámečníková et al. (2014a). For the angles100° – 65° and 50° – 45° the measured dis-tances were longer than the reference dis-tance, while in the range of 65° – 50° theywere shorter. The authors did not explain thespecific character of this phenomenon. Hav-ing eliminated the systematic factor, using apolynomial model, the values of standard de-viations were significantly reduced.From the point of view of the research re-

sults presented in our study, experiments witha metal plate were interesting. In VOEGTLE etal. (2008) the effect of the beam angle of inci-


was favourable from the point of view of therange and accuracy of the reflectorless meas-urements. Plexiglas had a specified degree oflight transmission. This parameter was un-known for the other materials. The tests as-sociated with different thicknesses of the syn-thetic materials were performed for two mate-rials: with high density, i.e. for Plexiglas, andwith low density, i.e. for Styrofoam.Temperature also influences the results of

the reflectorless distance measurements. Allmaterials were tested at two temperatures: atabout 15 °C and at about 40 °C. The tempera-ture of 15 °C (cold) was ambient temperature,and the temperature of 40 °C (warm) wasreached by heating the materials with a porta-ble heater. The temperature measurement wascarried out using a Testo 830-T2 pyrometer.The results of the reflectorless measure-

ments are also affected by the incidence angleof the beam on the target. To assess the sig-nificance of this factor, basic tests were per-formed at the most favourable, normal inci-dence of the laser beam on the target (90°). Forcomparison, measurements were also carriedout for the beam incidence at an angle of 45°.Materials with partial transparency placed atan angle increase the distance which the beamcan travel within the structure of the materi-al. Potentially, this could affect the increase in

2000). Depending on the type of the material,the influence of each of these properties var-ies with respect to the errors of the measureddistance. On the basis of the tests carried outin LENDA & MARMOL (2010), it was observedthat the magnitude of these phenomena is af-fected mainly by factors such as the degree oflight transmission, type of the synthetic ma-terial (chemical composition), density, thick-ness and the temperature of the material. Thecolour of the material is of some importanceas well, which, however, was not included inthe previously conducted research. Accordingto these observations, a group of 16 materialswas selected for testing. The analysis of 13 ofthese (Tab. 1), due to the similar properties ofthe materials and obtained results, turned outto be sufficient.Plexiglas was the material with the big-

gest changes in measured distances during therange finder tests. Therefore, for this particu-lar synthetic material, the most detailed testswere planned, taking into account its colour,thickness and light transmission. Other highdensity materials included: HIPS, Polysty-rene and Polyolefin (rubber). Materials withlow density were represented by foam syn-thetic materials: Styrofoam and Polyurethanefoam. All tested materials, except for the col-oured Plexiglas, were white or gray, which

Tab. 1: The materials tested.

No. Material type Colour Thickness(mm)

Transparency(%)

1 Plexiglas (Polymethyl methacrylate) red 5 30

2 Plexiglas green 5 30

3 Plexiglas blue 5 30

4 Plexiglas white 5 30


6 Plexiglas black 5 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 -


and then the distance between them was com-pared, taking into account the thickness of thecardboard. The compared planes were createdfrom two different point clouds, so they werenot exactly parallel. The determination of thedistance between them was carried out as fol-lows: The regression planes have been fittedin the reference cloud and the test cloud. Thedense point grid with a resolution of 1 pointper 3 mm² has been interpolated on the testplane. Then the distances from all points ofthe sample of the reference plane have beendetermined, along the normal to that plane.Then the average distance has been calculatedand taken as the distance between the planes.Such a method of determining the distanceallowed eliminating inaccuracies associatedwith a slight rotation of both planes. The dif-ference was the distance measurement devia-tion of the synthetic material with reference tothe white cardboard, and, consequently, to thereflective tape. The deviation of the distancemeasurement of synthetic materials observedin this way (ΔD) is not an overall error of dis-tance measurement. It is, however, an addi-tional contribution, which should be added tothe standard errors of a laser scanner, assum-ing reflection of a beam from the outer sur-face of the material. In addition, the precision(mp) of fitting the regression plane into a setof points representing the synthetic material,expressed by the RMSerror of the distancesof the points from the regression plane was de-termined.The tests were performed using two laser

scanners with similar accuracy, but with adifferent carrier wavelength: Leica C10 andRiegl VZ400. Basic data of these scannerswith reference to the subject in question werepresented in Tab. 2. The carrier wavelength isimportant especially when measuring targetsof varying colours since spectral curves maydiffer even in the case of very similar materi-als (Toś 2014). The colour of the material mayinfluence the absorption of beams with differ-ent wavelengths. It can also affect the meas-urement accuracy of synthetic materials in avaried manner, depending on the size of theinhomogeneity in their structure.In the production process it is not always

possible to obtain a uniform chemical compo-sition and repetitive structure of synthetic ma-

the error values. On the other hand, due to thepartial reflection of the beam from the surfaceof the material placed at an angle, the pene-tration of the interior of the material can belimited, which will reduce the value of the er-rors. For these reasons, it is difficult to find ananalogy with other studies at varying anglesof the beam incidence, contained in the exist-ing literature.Since our previous studies of reflectorless

range finders presented the effect of a distanceon the obtained errors associated with greaterpenetration of the material by the laser beam onshorter distances, tests of the scanners werecarried out on the bases of three lengths: 5 m,15 m and 50 m.The procedure of determining the distance

measurement deviations for each of the mate-rials requires an explanation. We built a tar-get of 20 cm × 30 cm for each material. Thosetargets were split into two halves of 20 cm ×15 cm each (Fig. 1). For the cardboard, whichserved as a reference, the distance deviationin relation to the original Leica reflective tapetarget had been predetermined – it equals0.6 mm – taking into account the differencein thickness of the cardboard and tape target(LEICA 2015). It was decided not to stick thetape target directly on the tested synthetic ma-terials, due to a possibility of it being trans-mitted by the laser beam, which would en-able the penetration into the structure of thesynthetic material, distorting the results. Thetested synthetic materials adhered directly tothe opaque background. The planar target wasscanned at a high resolution (2 mm density).Points near the boundaries of the two halvesof the targets were removed from the cloud ofpoints in order to obtain homogeneous sets ofpoints. Based on the filtered data, a regressionplane was determined for each of the fields,

Fig. 1: Experimental setup.


3 Results of Test Measurements

The results obtained from the conducted testsare shown in Tab. 3 (Leica C10 scanner) andTab. 4 (Riegl VZ400 scanner). Due to high-ly ambiguous results (scattered data) of themeasurements performed at the angle of 45°at the distances of 15 m and 50 m, they arenot presented here. The diagrams presentedin Figs. 2 and 3 were drawn based on theseTables. The measurement results of differentsamples of the same materials are included in

terials. For this reason, comparative tests wereperformed for several materials (Plexiglaswhite, 5 mm thickness and 30% transparen-cy (hereinafter briefly stated with x mm/Y%),Polystyrene 5 mm, and Styrofoam 10 mm),taking into account the measurement of threesamples of the same production lot, as wellas the measurement to three samples of un-known origin. The tests were performed usingthe Leica C10 scanner, at the baseline of 5 m,the distance between the laser scanner and thetest material.

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

Tab. 3: Distance measurement deviations of the Leica C-10 scanner (ΔD = distance measurementdeviations relative to the comparative model, mp = precision of fitting the regression plane for asynthetic material, tr. = transparency).


Tab. 5. During the tests with the VZ400 scan-ner, it was noticed that synthetic materialswere not directly adhering to the background,but they stood out approximately 2 cm – 3 cmbehind it. This resulted in a significant deteri-oration in the distance measurement deviationfor some materials. The measurements wererepeated and, for comparison, Tab. 6 showsthe results for the background adhering andnot adhering to the synthetic material.The precision of fitting the regression plane

for the reference cardboard was similar forboth scanners and it was ±1.2 mm for 90°beam incidence angles, and ±0.8 mm for the45° incidence angles, with the scatter for thesevalues reaching a maximum of ±0.2 mm.

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

Tab. 5: Distance measurement deviations ofthe samples for selected materials: A1–A3 –sample material from the same production lot,B1–B3 – sample material from unknown pro-duction lots, Leica C-10 scanner, distance 5 m.

Plexiglaswhite, 5 mm,

30%transparency

(mm)

Polystyrene,5 mm

(mm)

Styrofoam10 mm

(mm)A 1 8.5 4.8 3.3A 2 8.1 4.6 3.9A 3 8.8 5.1 4.2B 1 8.5 4.8 3.3B 2 6.9 4.7 5.2B 3 11.8 6.4 2.8


Fig. 2: Distance measurement deviations for the materials 1–6, ΔD – distance measurement de-viations relative to the comparative model.


Fig. 3: Distance measurement deviations for the materials 7–13, ΔD – distance measurementdeviations relative to the comparative model.


4 Analysis of the Results

The assessment of the results was divided intoseveral categories, resulting from the prop-erties of the material itself, the conditions inwhich it was found, and the tested instrument.The analyses included the following catego-ries: type of material (light transmission, den-sity and internal structure, colour, thickness),distance, angle of incidence, temperature, typeof scanner (carrier wavelength), and distanceof the synthetic material from the background.Some phenomena were especially noticeablefor the shortest measurement distance (5 m), atlow temperature of the material (15 °C), whichis explained later in the text. Therefore, dur-ing the comparisons, we will frequently referto the results obtained under such conditionsas the “basic” conditions.

4.1 Light Transmission

This is the primary factor affecting the dis-tance measurement of synthetic materials. Ifthe laser beam is not able to penetrate into thestructure of the synthetic material, there areno changes in the measured distance in com-parison with the reference cardboard. Such asituation can be observed only for one of thetested materials, i.e. opaque rubber (Polyole-fin). Regardless of the conditions, the distancemeasurement error for this material did notexceed the value of 2 mm, usually remainingat the level of about 1 mm, unlike other mate-rials, where a certain degree of transparencyallowed for the occurrence of additional phe-nomena affecting the distance measurement.The changes in the distance observed for themranged from about 2 mm to 15 mm. The influ-ence of transmission was clearly visible whencomparing the two samples of Plexiglas white,with a transmission of 30% and 70%. Meandeviations for the material with lower trans-mission were approximately 7 mm (C10) and8 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 asignificant light transmission. Other materialswere measured with mean deviations within

Tab. 6: Comparison of the distance measure-ment deviations of the VZ-400 scanner forvarious locations of the synthetic materialbackground. Test 1: The background adheringto the rear surface of the synthetic material.Test 2: The background placed approximately2 cm – 3 cm behind the synthetic material.Comparison for the 5 m distance (tr. – trans-parency).


dex will be subject to changes. Comparing themeasurements results performed with the C10scanner to various samples of materials fromthe same production lot (Tab. 5), a stable levelof deviations for Plexiglas white 5 mm / 30%– (8.1 mm – 8.8 mm) and Polystyrene 5 mm /30% – (4.6 mm – 5.1 mm) is noticeable. Forthe foamed material, 10 mm Styrofoam, therelative differences are larger – (3.3 mm –4.2 mm). When comparing samples of thesematerials originating from different produc-tion lots, possibly from different manufactur-ers, the results look somewhat different. Dif-ferences for 5 mm Polystyrene increased to4.7 mm – 6.4 mm. For other materials, the rel-ative differences were nearly doubled: 10 mmStyrofoam: 2.8 mm – 5.2 mm, Plexiglas white5 mm/30%: 6.9 mm – 11.8 mm. Therefore,the type of a material and its internal structurehave a significant impact on the level of themeasurement deviations. At the same time,this influence is difficult to determine accu-rately due to variations in the production.

4.3 Colour

The colour of the synthetic material signifi-cantly influences the measurement deviations.A sample Plexiglas was tested under basicconditions using the same parameters regard-ing thickness and permeability. Then, the de-viations resulting from the changing colourof the material were within 4.2 mm – 9.1 mm(C10) and 7.1 mm – 10.5 mm (VZ400). Forboth scanners, the differences between the ex-treme values of the deviations had a factor of1.5 to 2. As only a few selected colours weretested, a discrepancy may be even larger un-der real conditions. The colour of the materialstrongly influenced the absorption of beamswith different wavelengths. Due to the factthat two scanners emitting laser radiation ofdifferent carrier wavelengths, 532 nm (C 10)and 1550 nm (VZ400), were selected for thisexperiment, the differences in the values ofthe deviations for the same colours were clear-ly noticeable. Under basic conditions, for theC10 scanner, the smallest deviations were ob-served for the following colours: blue: 4.2 mm,black: 4.4 mm, and red: 5.3 mm. It is interest-ing that, at the same time, for those colours

the range of 1 mm – 3 mm (C10) and 1 mm –5 mm (VZ400). For the C10 scanner, Plexiglaswith 70% transmission caused a slight dete-rioration in the precision of fitting the regres-sion plane relative to other materials (up to ca.±2 mm). For the VZ400 scanner, the fittingprecision for this material decreased in somecases to ca. ±3 mm.

4.2 Density and Internal Structure

The studies covered two groups of materialswith large differences in density and inter-nal structure. The first one included materialswith a continuous structure (Plexiglas, HIPS,Polystyrene, Polyolefin), and the second oneincluded foamed materials (Styrofoam, Poly-urethane Foam). In the foamed materials thebeam was subject to changes, alternately infragments of the material and in air cham-bers. Foamed materials were measured withsmaller deviations in relation to the thicknessof the material than their continuous counter-parts. This could be assessed by comparingthe measurement deviations of Polystyreneand Polystyrene Foam (Styrofoam) (5 mm) inbasic conditions: Polystyrene: 4.8 mm (C10),6.3 mm (VZ400) and Styrofoam: 2.1 mm(C10), 2.9 mm (VZ400). The foamed materi-al achieved deviations of approximately halfthat size. It is interesting that for some mate-rials, especially for Plexiglas, the measure-ment deviations were larger than the thicknessof the material itself. Examples under basicconditions were Plexiglas white 5 mm / 30%– 8.5 mm (C10), 9.4 mm (VZ400), Plexiglaswhite 5 mm / 70% – 13.7 mm (C10), 15.2 mm(VZ400). During the tests, the materials ad-hered directly to the opaque background, sothat the beam, having passed through the ma-terial, did not have a possibility of incidenceon other targets. The values of the deviations,therefore, must have resulted from additionalphenomena associated with the beam propa-gation in the structure of the material. Theymight be related, among others, to the repeatedrefraction or the reflection of a beam. In addi-tion, the refractive index depends on the inter-nal structure of materials. If it is not homoge-neous, and the inhomogeneity has dimensionsgreater than the wavelength, the refractive in-


be recognized with other synthetic materialsmeasured with the VZ400 scanner, e.g. HIPSand Polystyrene. The dependencies relatedto the thickness of the synthetic material aretherefore significant, nonlinear, and for betterassessment it would be advisable to carry outmore extensive tests.

4.5 Distance

Dependencies resulting from the variable dis-tance of the target are easiest to evaluate fromthe diagrams in Figs. 2 and 3. For most materi-als, the deviations are reduced with increasingdistance. The differences between the resultsfor the distances of 5 m and 15 m do not exhib-it this phenomenon clearly, i.e. sometimes thedeviations even increase. However, the differ-ences between the results at the distances of5 m and 50 m exhibit a certain tendency andreach the values of up to 3.5 mm. The increas-ing tendency of the deviations was observedfor two foamed materials: Styrofoam (10 mm)and Polyurethane Foam for the VZ400 scan-ner. Deviations decreasing with distance arecontrary to current experiments for reflector-less instruments. Moreover, the equipmentmanufacturers themselves inform about in-creasing deviations with increasing distance.However, this phenomenon can be explainedby a reduced ability of radiation to penetratethe material which, at the larger distance, hasa lower energy density at the surface of thesynthetic material. In this way, the scale of thephenomena described above is limited. Dueto the fact that the largest deviations were ob-served at short distances, the 5-meter baselinewas set as the “basic” condition. With the in-creasing distance there was no significant in-crease in the precision of fitting the regressionplane.

4.6 Incidence Angle

Studies related to the varying incidence anglesof the beam yielded inconclusive results. Thetests were carried out at the angle of 45° with-out the materials being heated. The results forthe distances of 15 m and 50 m did not demon-strate any systematic order, the measurement

the lowest precision of fitting a regressionplane were obtained: blue: ±2.8 mm, black:±3.0 mm, red: ±3.9 mm. Larger deviationswere observed for white (8.5 mm) and green(9.1 mm), with better precision of fitting theplane: white ±1.2 mm, green ±1.4 mm. TheVZ400 scanner demonstrated the smallest de-viations for red (7.1 mm) and green (7.5 mm),and the largest for the blue (10.2 mm) andblack (10.5 mm) colours. Therefore, for green,blue and black, the dependencies observedfor the C10 scanner were inverted. In basicconditions, the precision of fitting the regres-sion plane were at a low level for all colours(±1.0 mm – ±1.6 mm). For the test configu-rations other than the basic ones, the devia-tions for red and blue increased. The colour ofthe synthetic material in combination with itslight transmittance is therefore a very impor-tant factor determining the value of the dis-tance measurement deviations.

4.4 Thickness

The effect of the thickness of the synthetic ma-terial was assessed based on two samples ofmaterials (Plexiglas 30%, 5 mm and 10 mm,and Styrofoam, 5 mmand 10 mm).Under basicconditions, the measurement deviations usingthe C10 scanner for Plexiglas ranged from8.5 mm to 11.6 mm, and for the VZ400 scan-ner from 9.4 mm to 13.3 mm. For Styrofoam,the deviations for the C10 scanner ranged from2.1 mm to 3.3 mm, and for the VZ400 scannerfrom 2.9 to 5.2 mm. The deviations were dif-ferent for each sample and scanner, and werewithin the limits of 36% – 79%, at 100%change in thickness of the material. Greaterdifferences were observed for the foamed ma-terial of low density, i.e. for Styrofoam 57%– 79%, while for Plexiglas they amounted to36% – 41%. As Styrofoam in practice had athickness greater than 5 mm or 10 mm, thiscould translate into relatively large measure-ment deviations. In view of the fact that colourhas also a strong influence on the deviations ofthe measurements on Plexiglas the deviationscould reach values higher than the thicknessof the material itself. This happened in morethan half of the cases for the C10 scanner andall of the cases for the VZ400. This could also


deviations of fitting the regression plane fromabout ±1.2 mm to ±0.7 mm on an average.Similar observations were made in VOEGT-LE et al. (2008), where it was also found thatwith a decreasing incidence angle, the devia-tion decreased. However, in SOUDARISSANANEet al. (2007), for some materials an oppositetendency was observed. It was also confirmedby the previous research of the authors associ-ated with the tests of reflectorless range find-ers (LENDA&MARMOL 2010). Similar relationswere observed for the instrument VZ400,where the deviations decreased by more than4 mm. However, for some materials (Plexiglasgreen and blue, 10 mm Styrofoam), a slight,about 1 mm increase in the deviations oc-curred. The precision of fitting the regressionplane were also subject to a decrease an av-erage of about ±1.2 mm to ±1.0 mm. In ourexperiments, at the angle of incidence of 45°,the materials Plexiglas red and black were im-measurable for the Leica scanner. Based onthe obtained results and the results of other re-searchers, it was apparent that the angles ofincidence affect the measured distance. Thetests exhibited a decrease in the deviationsat more acute angles of the beam incidence –which may be the result of reduced penetra-tion by laser beam due to partial reflectionfrom the surface of the sample at incidenceangle of 45°.

deviations just increased or decreased in arandom manner. This was associated with sig-nificant decrease of precision of fitting the re-gression planes with up to several millimetres.The ambiguous test results at greater dis-

tances could have resulted from the variablereaction of fragments of the materials to theillumination at an angle. At a greater distancewhere the density of the radiation energy atthe sample surface is smaller, the non-homo-geneity of the material could have determinedthe phenomenon of partial reflection fromthe surface of the inclined sample, changingthe degree of the material penetration by thebeam. This is particularly visible for the bluePlexiglas measured with the C10 scanner at adistance of 15 m (Fig. 4). The left figure pre-sents a sample scanned at the angle of 90°,and the right one at the angle of 45°. Signifi-cant losses in the point cloud (front view, rightside) together with highly scattered individualpoints (top view, right side) are visible in thelatter case.For this reason, only the results for the

distance of 5 m are presented, because theyseemed not to be subject to blunders. With thedecreasing angle of the beam incidence (fromthe standard 90° to 45°), for the C10 scanner,a decreasing measurement deviation of up toabout 3 mm in almost all the cases was ob-served. This was related to the decrease in the

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.


transmitted through the material, can be re-flected by the background, causing significantchanges to the obtained results, of up to sev-eral centimetres. For some synthetic materi-als it does not really matter (HIPS, Polysty-rene, Styrofoam, Polyolefin), for others thesechanges may be small (Polyurethane foam).For some of the materials (Plexiglas, espe-cially with small thickness) it may, however,result in very significant discrepancies. Theobserved increase in the deviations for dif-ferent colours of 5 mm Plexiglas, under ba-sic conditions, ranged from 71% (white) to asmuch as 345% (red), reaching the values ofup to 32 mm. It was combined with a slightincrease in the plane fitting precision. The de-pendencies associated with the colours werequite different from the ones observed for thebackground adhering to the synthetic materi-al. The loss of accuracy related to a completepassage of light through the synthetic mate-rial may thus result in obtaining significant,unpredictable values. Further studies, takinginto account various distances between thesynthetic material and the background, mayfoster interesting results.

5 Summary

The paper analyses a number of factors affect-ing the distance measurement of synthetic ma-terials. The most important of these is opticalpermeability, which induces other phenomenaoccurring within the material structure. Someof them may have a significant impact on theobtained results, e.g. degree of permeability,type of material, colour in combination with acarrier wavelength of the laser radiation, andthickness. The effects of other factors seemto be less significant, but noticeable and sys-tematic such as distance and temperature ofthe material. The angle of the beam incidenceis also relevant. Synthetic elements may there-fore exhibit shifts of several, sometimes overa dozen millimetres in the point cloud relativeto other objects. Especially when they not ad-here to the other materials, shifts in some cases(Plexiglas) may increase to several centimetres.In some cases, measurement deviation of thesynthetic material can affect the accuracy ofthe entire scanned structure. This may happen

4.7 Temperature of the Material

The influence of the temperature is easiest toassess from the diagrams contained in Figs.2 and 3. A decrease in the deviations for thenon-foamed materials (except rubber) relat-ed to the temperature growth is quite clear,as well as its lack for the foamed materials.Non-foamed synthetic materials, when heat-ed, generate deviations that are smaller byabout 1 mm – 3 mm (C10) and by about 1 mm– 4 mm (VZ400) than cold ones. Continu-ous materials, having no air bubbles in theirstructures, exhibited susceptibility to heat.This phenomenon could be explained by thethermal expansion of materials and the relatedlower density of synthetic materials, reducingthe refractive index value. However, since thechange in the volume of the synthetic materi-als when heated was negligible, other factorsrelated to temperature, unnoticed by the au-thors, must have affected the reduction of themeasurement deviations. Due to the fact thatlarger deviations were observed for cold tar-gets, low temperature was adopted as another“basic” conditions.

4.8 Scanner Type

In this category, the carrier wavelength of thelaser, different for the two scanners (532 nm(C10) and 1550 nm (VZ400)) should be distin-guished. The light of various wavelengths isabsorbed in a different manner by the coloursand materials of a different size and orderingof the molecules. Taking into account all 13tested materials, almost all of the tests demon-strated slightly higher deviation values for theVZ400 scanner. This dependency could havebeen expected when comparing the precisionparameters of both instruments.

4.9 Distance between SyntheticMaterial and the Background

The location of the background of the testedmaterial is of great importance to the meas-ured distance of some synthetic materials.If it is close, just a few centimetres from thesynthetic material, a laser beam, after being


mation Sciences, Commission V Symposium 38(5): 80–85, Newcastle, United Kingdom.

Boehler, W., Bordas, v.m. & marBs, a., 2003: In-vestigating laser scanner accuracy. – 19th CIPAInternational Symposium, Antalya, Turkey.

Bucksch, a., lindenBergh, r. & van ree, J., 2007:Error budget of terrestrial laserscanning: Influ-ence of the intensity remission of the scan qual-ity. – III International Scientific Congress GeoSiberia, Novosibirsk, Russia.

clark, J. & roBson, s., 2004: Accuracy of mea-surements made with a CYRAX 2500 laserscanner against surfaces of known color. – Sur-vey Review 37 (294): 626–638.

EREN, G., AUBRETON, O., MERIAUDEAU, F., SECADES,S., FOFI, D., NASKALI, A.T., TRUCHETET, F. & ER-CIL, A., 2009: A 3D Scanner for TransparentGlass. Image Analysis and Processing. – Inter-national Conference on Image Analysis and Pro-cessing (ICIAP) 2009, Lecture Notes in Com-puter Science (LNCS) 5716: 519–527.

Feynman, r.P., leighton, r.B. & sands, m., 2001:The Feynman Lectures on Physics. – PWN,Warszawa, Poland.

FOWLES, G.R., 1989: Introduction to Modern Op-tics. – Dover Publications, Mineola, NY, USA.

GOTTWALD, R., 2008: Field Procedures for TestingTerrestrial Laser Scanners (TLS) – A Contribu-tion to a Future ISO Standard. – IntegratingGenerations, FIG Working Week 2008: 1–14,Stockholm, Sweden.

hiremagalur, J., yen, k., akin, k., Bui, t., lasky,t. & ravani, B., 2007: Creating standards andspecifications for the use of laser scanning incaltrans projects. – AHMCT (Advanced High-way Maintenance & Construction TechnologyResearch Center) Research Report, Universityof California at Davis, California Department ofTransportation, Davis, CA, USA.

HOLST, C., ZEIMETZ, P., NOTHNAGEL, A., SCHAUERTE,W. & kuhlmann, H., 2012: Estimation of FocalLength Variations of a 100m Radio Telescope’sMain Reflector by Laser Scanner Measure-ments. – Journal of Surveying Engineering 138(3): 126–135.

Johnson, W. & Johnson, a., 2012: Operational Con-siderations for Terrestrial Laser Scanner Use inHighway Construction Applications. – Journalof Surveying Engineering 138 (4): 214–222.

kaasalainen, s., Jaakkola, a., kaasalainen, m.,krooks, a. & kukko, a., 2011: Analysis of Inci-dence Angle and Distance Effects on TerrestrialLaser Scanner Intensity: Search for CorrectionMethods. – Remote Sensing 3 (10): 2207–2221.

kersten, t.P., mechelke, k., lindstaedt, m. &STERNBERG, H., 2009: Methods for GeometricAccuracy Investigations of Terrestrial Laser

when spatial markers in the form of Styrofoamballs are used to connect neighbouring scans.As it was demonstrated by our experiments,the error for Styrofoam increases significant-ly with its thickness, and therefore coveringsuch balls with opaque lacquer should be con-sidered. Some of the tests yielded inconclu-sive or questionable results (incidence angle),which may have resulted from the methodol-ogy, dependent upon the occurrence of addi-tional phenomena, which the authors had notanticipated. The factors that interfere with theresults of the research studies may include forexample: additional lighting, reflections of thebeam from the surrounding elements, typeof the synthetic material background and themanner of its adherence to the sample. Thedeviations resulting from a complete passageof light through the synthetic material and thereflection of a beam from the background lo-cated at a certain distance from the materialmay reach significant values. In general, dueto a large number of factors affecting the dis-tance measurements and their mutual con-figurations, it is difficult to tabulate distancemeasurement deviations for specific syntheticmaterials. The situation is made worse by anaverage repeatability of their production. Thisstudy draws attention to the problem, allowingestimating the scale of the existing dependen-cies.

Acknowledgements

This work has been made within scien-tific research program no. 11.11.150.005,11.11.150.949.

References

aBBas, m.a., setan, h., maJid, Z., chong, a.k.,idris, k.m. & asPuri, a., 2013: Calibration andAccuracy Assessment of Leica ScanStation C10Terrestrial Laser Scanner. Developments inMultidimensional Spatial Data Models. – Lec-ture Notes in Geoinformation and Cartography10: 33–47.

Berenyi, a., lovas, t. & Barsi, a., 2010: Terres-trial laser scanning – civil engineering applica-tions. – International Archives of the Photo-grammetry, Remote Sensing and Spatial Infor-


Influencing factor on the quality of terrestriallaser scanning points. – ISPRS Journal of Photo-grammetry and Remote Sensing 66 (4): 389–399.

toś, c., 2014: The possibilities of using the terres-tial scanning data for classification of rocks inlimestone mine ‘Czatkowice’. – Reports on Ge-odesy and Geoinformatics 97: 80–90.

trager, F., 2007: Springer Handbook of Lasers andOptics. – Springer.

voegtle, t., schWaB, i. & landes, t., 2008: Influ-ence of different materials on the measurementsof terrestrial laser scanner. – The InternationalArchives of the Photogrammetry, Remote Sens-ing and Spatial Information Sciences 37 (B5):1061–1066.

WANDACHOWICZ, K., 2000: Wyznaczanie rozkładuluminancji we wnętrzach z uwzględnieniemkierunkowo-rozproszonych charakterystyk od-biciowych materiałów (Determination of the lu-minance distribution in the interiors taking intoaccount the characteristics of the directional-diffuse reflectance materials). – doctoral disser-tation, Poznań, Poland.

Zámečníková, m., neuner, h. & PegritZ, s., 2014a:Influence of the Incidence Angle on the Reflec-torless Distance Measurement in Close Range.– INGEO 2014 – 6th International Conferenceon Engineering Surveying: 257–262, Prague,Czech Republic.

Zámečníková, m., Wieser, a., WoschitZ, h. &ressl, c., 2014b: Influence of surface reflectivi-ty on reflectorless electronic distance measure-ment and terrestrial laser scanning. – Journal ofApplied Geodesy 8 (4): 311–325.

Addresses of the Authors

GRZEGORZ LENDA, Ph.D. & URSZULA MARMOL,Ph.D., AGH University of Science and Technology,Faculty of Mining Surveying and EnvironmentalEngineering, Cracow, Mickiewicza 30, Poland,Tel.: +48-12-617-23-14, -26, Fax: +48-12-617-50-76,e-mail: {grzenda}{entice}@agh.edu.pl

GRZEGORZ MIREK, Ph.D., Cracow University ofTechnology, Faculty of Environmental Engineer-ing, Cracow, Warszawska 24, Poland, Tel.: +4812628-21-87, e-mail: [email protected]

Manuskript eingereicht: Februar 2015Angenommen: Juli 2015

Scanning Systems. – PFG – Photogrammetrie,Fernerkundung, Geoinformation 2009 (4): 301–315.

lee, i.s., lee, J.o., Park, h.J. & Bae, k.h., 2010:Investigations into the influence of object char-acteristics on the quality of terrestrial laser scan-ner data. – KSCE Journal of Civil Engineering14 (6): 905–913.

LEICA, 2015: http://www.allenprecision.com/survey/robotic-total-stations/accessories/prisms/60x60mmreflectivetape20pack/(19.7.2015).

lenda, g. &marmol, u., 2010: Accuracy of reflec-torless range finders for the measurements ofobjects made form synthetic materials (polish).– Pomiary Automatyka Kontrola 11: 1280–1286.

lichti, d.d. & Jamtsho, s., 2006: Angular resolu-tion of terrestrial laser scanners. – The Photo-grammetric Record 21 (114): 141–160.

monserrat, o. & crosetto, m., 2008: Deformationmeasurement using terrestrial laser scanningdata and least squares 3D surface matching. –ISPRS Journal of Phtotogrammetry & RemoteSensing 63: 142–154.

PFeiFer, n., höFle, B., Briese, c., rutZinger, m. &HARING, A. 2008: Analysis of the backscatteredenergy in terrestrial laser scanning data. – Inter-national Archives of the Photogrammetry andRemote Sensing and Spatial Information Sci-ence (IAPRS) 37 (B5): 1045–1051.

Polo, m.e., Felicísimo, a.m., villanueva, a.g. &martíneZ-del-PoZo, J.a., 2012: Estimating theUncertainty of Terrestrial Laser Scanner Meas-urements. – IEEE Transactions on Geoscienceand Remote Sensing 50 (11): 4804–4808.

REES, W., 1990: Physical principles of remote sens-ing. – Cambridge University Press, UK.

salemi, g., achilli, v. & Boatt, g., 2008: 3D vir-tual modelling of a gothic stained-glass panel. –The International Archives of the Photogram-metry, Remote Sensing and Spatial InformationSciences 37 (B5): 297–302.

soudarissanane, s., van ree, J., Bucksch, a. &lindenBergh, r., 2007: Error budget of terres-trial laser scanning: influence of the incidenceangle on the scan quality. – 3DNordOst, Berlin.

SOUDARISSANANE, S., LINDENBERGH, R., MENENTI, M.& teunissen, P.J.g., 2009: Incidence angle influ-ence on the quality of terrestrial laser scanningpoints. – ISPRS International Archives of thePhotogrammetry, Remote Sensing and SpatialInformation Sciences 38-3/W8: 183–188.

SOUDARISSANANE, S., LINDENBERGH, R., MENENTI,m., teunissen, P.J.g., 2011: Scanning geometry:

accuracy of laser scanners for measuring surfaces made of synthetic materials · 2017. 3. 8. ·...

Documents