European Journal of Remote Sensing

ISSN: (Print) 2279-7254 (Online) Journal homepage: https://www.tandfonline.com/loi/tejr20

Adequacy of pseudo-direct georeferencing

of terrestrial laser scanning data for coastal
landscape surveying against indirect
georeferencing

Marion Jaud, Pauline Letortu, Emmanuel Augereau, Nicolas Le Dantec,

Mickaël Beauverger, Véronique Cuq, Christophe Prunier, Réjanne Le Bivic &
Christophe Delacourt

To cite this article: Marion Jaud, Pauline Letortu, Emmanuel Augereau, Nicolas Le Dantec,
Mickaël Beauverger, Véronique Cuq, Christophe Prunier, Réjanne Le Bivic & Christophe Delacourt
(2017) Adequacy of pseudo-direct georeferencing of terrestrial laser scanning data for coastal
landscape surveying against indirect georeferencing, European Journal of Remote Sensing, 50:1,
155-165, DOI: 10.1080/22797254.2017.1300047

To link to this article: https://doi.org/10.1080/22797254.2017.1300047

© 2017 The Author(s). Published by Informa Published online: 13 Mar 2017.

UK Limited, trading as Taylor & Francis
Group.

Submit your article to this journal Article views: 1102

View related articles View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tejr20
EUROPEAN JOURNAL OF REMOTE SENSING, 2017
VOL. 50, NO. 1, 155–165
https://doi.org/10.1080/22797254.2017.1300047

Adequacy of pseudo-direct georeferencing of terrestrial laser scanning data

for coastal landscape surveying against indirect georeferencing
Marion Jaud a, Pauline Letortub, Emmanuel Augereaua, Nicolas Le Danteca,c, Mickaël Beauvergera,
Véronique Cuqb, Christophe Pruniera, Réjanne Le Bivica and Christophe Delacourta
a
Laboratoire Domaines Océaniques – UMR 6538, Université de Bretagne Occidentale, IUEM, Technopôle Brest-Iroise, Plouzané, France;
b
LETG-Brest Géomer – UMR 6554, Université de Bretagne Occidentale, IUEM, Technopôle Brest-Iroise, Plouzané, France; cCEREMA -
Centre d’Etudes et d’expertise sur les Risques, l’Environnement, la Mobilité et l’Aménagement, DTecEMF, Plouzané, France

ABSTRACT ARTICLE HISTORY

The georeferencing process is crucial to the accuracy of terrestrial laser scanner (TLS) data, in Received 20 February 2017
particular in the context of diachronic studies relying on multi-temporal surveys. The use of Accepted 22 February 2017
Ground Control Points in the georeferencing process can however be complex when confronted
KEYWORDS
with the practical constraints of coastal surveying. A simple and quick alternative method called Terrestrial laser scanning;
“pseudo-direct georeferencing” is proposed in the present paper. This method involves internal coastal monitoring; direct
inclinometers to measure roll and pitch angles and a centimetric GPS to measure the position of georeferencing; indirect
the TLS center and the position of one backsight target. When assessing the transformational georeferencing;
uncertainty by using a set of independent ground validation points for both classical indirect and multi-temporal surveys;
proposed pseudo-direct methods, we respectively obtain root mean square errors of 4.4 cm for geomorphological evolution
the indirect method and 3.8 cm for the pseudo-direct method.

Introduction Among other topographic survey instruments, ter-

restrial laser scanners (TLSs) are convenient since
Coastal landscapes are shaped by a set of forcing factors,
they allow acquisitions at high temporal and spatial
highly variable both in time and location, making the
resolution with rather limited constraints: the equip-
littoral zone a very dynamic environment. A good
ment is very portable and the scan positions are
understanding of this hydro-geomorphological system
defined by the user. TLS data, combined with data
requires considering a range of time scales from geolo-
from other sources (high resolution satellite, aerial or
gical to historical, seasonal and events scale (especially
UAV imagery, photogrammetry, Differential GPS
to account for extreme events) as well as different
[DGPS] survey, etc.), is thus a key asset for diachro-
spatial scales from regional (kilometers) to local and
nic surveys, which are essential for coastal
fine (centimeters). Coastal morphodynamics therefore
monitoring.
needs high spatial resolution and high precision mon-
The TLS technique has been widely used in the
itoring, but also monitoring at a high frequency (e.g.
study of coastal mass transfers to
before and after storm events, and ideally at each low
● analyze rock falls including those affecting
tide during the storm) with extensive spatial coverage.
Furthermore, working with datasets georeferenced in coastal cliffs (Abellán et al., 2014; Kuhn &
absolute coordinates and high absolute accuracy in 3D Prüfer, 2014; Letortu et al., 2015; Quinn,
positioning are necessary to combine multisource data, Rosser, Murphy, & Lawrence, 2010; Rosser,
to carry out diachronic analyses in order to study evolu- Brain, Petley, Lim, & Norman, 2013; Rosser,
tion trends or even to easily share results with coastal Petley, Lim, Dunning, & Allison, 2005);
● quantify changes on beach sediment budget, dune
stakeholders.
The most basic method for coastal landscape survey- and tidal marsh (Leroux, 2013; Lim, Dunning,
ing is the GPS or tacheometer pole measurement, a Burke, King, & King, 2015; Nield, Wiggs, &
point-wise method associated with low spatial resolution Squirrell, 2011; Pietro, O’Neal, & Puleo, 2008;
and therefore strong uncertainties. Structure-from- Schubert, Gallien, Majd, & Sanders, 2015).
Motion photogrammetric surveys from unmanned aerial Paffenholz (2012) considers three approaches for
vehicles (UAVs) are an emerging alternative to provide georeferencing 3D point clouds, the raw output gen-
high resolution topography (Delacourt et al., 2009; erated by a TLS:
Harwin & Lucieer, 2012; Mancini et al., 2013). (1) indirect georeferencing via ground control
points (GCPs), using targets, either fixed

CONTACT Marion Jaud [email protected] Laboratoire Domaines Océaniques - UMR 6538, Université de Bretagne Occidentale, IUEM,
Technopôle Brest-Iroise, Rue Dumont D’Urville, Plouzané F-29280, France
© 2017 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
156 M. JAUD ET AL.

positioned or temporally placed before each surveys (Alho et al., 2011; Barber, Mills, & Smith-Voysey,
survey, that are measured with an independent 2008; Kukko, Kaartinen, Hyyppä, & Chen, 2012), which
method (GPS, theodolite) (Lague, Brodu, & are highly dependent on the positioning and inclination
Leroux, 2013; Alba et al., 2007; Letortu et al., angle measurements. In recent years, significant progress
2015); has been made in sensor technology, improving data
(2) direct georeferencing, using the internal or reliability and reducing acquisition time. Many studies
external sensors attached to the TLS in order to address the comparison of georeferencing techniques for
directly provide the required registration para- TLS data. However, these studies mainly focus on archi-
meters (Lichti & Gordon, 2004; Paffenholz, tectural surveying (Alba, Giussani, Roncoroni, & Scaioni,
2012; Reshetyuk, 2009; Scaioni, 2005); 2007), civil surveying (Scaioni, 2005), landslide surveying
(3) cloud-matching techniques using overlapping (Giussani & Scaioni, 2004; Kasperski et al., 2010), for
parts of point clouds (Besl & McKay, 1992; which purposes operational constraints and require-
Olsen, Johnstone, Driscoll, Ashford, & Kuester, ments are quite different from those for coastal monitor-
2009; Olsen, Johnstone, Kuester, Driscoll, & ing. This paper reports on an improvement to the
Ashford, 2011; Schürch, Densmore, Rosser, surveying protocol, precisely a reduction in survey dura-
Lim, & McArdell, 2011). Registration uncer- tion without affecting georeferencing accuracy. After a
tainty may be assessed comparing fixed points brief theoretical description of TLS data registration, this
in datasets as buildings, roads, etc. This data- paper presents a field survey where both methods –
driven approach is only be mentioned here with- classical (indirect-georeferencing) and alternative
out further details as it relies on prior (direct or (pseudo-direct georeferencing) – are tested. The results
indirect) georeferencing of an initial point cloud of both methods are then compared and discussed.
in order to obtain absolute georeferencing for
the subsequent surveys.
When defining an optimized protocol for TLS Background on TLS data registration
surveying applicable to coastal environment monitor- Coordinate systems definition and
ing, the following practical constraints are relevant: transformational uncertainty
● foreshore accessibility is time-limited because of
tides, which restricts the number of scans that During data acquisition, the position of an object
can be performed and thus imposes compromis- which has reflected the laser beam is defined rela-
ing between survey area for coverage purpose tively to the TLS position knowing the slant range ρ
and scan overlap for co-registration purpose; (radial distance), the azimuthal angle α and the incli-
● the TLS tripod may have to be placed on nation angle β (Figure 1). The transformation from
undrained sands, in the absence of a better spherical coordinates to Cartesian coordinates is
option, and thus be subject to settling, in given by Equation (1):
which case assessing the position and inclination 0 1 0 1
x ρ cos β cos α
of the TLS device becomes crucial; @ y A ¼ @ ρ cos β sin α A
● adequate positioning of the reflective targets
(1)
z ρ sin β
may not be possible in certain areas (partial
masking of satellite constellation at the cliff
foot, submerged zone, masking by rocks, The position of the measured points is thus
stranded seaweeds, etc.). expressed in a local coordinate system associated with
● setting up a fixed network of GCPs or perma- the sensor, generally referred in literature and further
nent geodetic marks for georeferencing purpose mentioned as Intrinsic Coordinate System (ICS). The
is challenging because the TLS is generally ICS is defined by the position of the TLS reference point
located on the foreshore, which is subject to and the orientation of the TLS axes (Figure 1).
significant topographic changes (seasonal and To exploit TLS datasets for (1) multi-temporal or
even event-driven variations in beach level), in multi-sensor comparisons, (2) integration with other
addition to tampering risks due to the high geospatial data and (3) diffusion to stakeholders, they
frequentation of coastal sites. need to be registered in a common reference system.
Such constraints call for a time-efficient protocol for Georeferencing consists in transforming the regis-
TLS data acquisition in coastal environments. In par- tration of the point cloud from the ICS into an
ticular, setting up the targets used as GCPs and mea- absolute or GLobal Coordinate Systems (GLCS), gen-
suring their position is a time-consuming step, erally associated to a geodetic datum. This transfor-
especially with respect to the scanning speed of cur- mation is associated with the matrix operator MIGL
rent TLS technology. (Figure 1), a unique combination of a translation and
The quality of the georeferencing is a key issue not a rotation. Thus, a set of six transformation para-
only for TLS surveys but also for Mobile Laser Scanning meters is required for this registration.
 EUROPEAN JOURNAL OF REMOTE SENSING 157

Figure 1. Definition of the coordinates systems.

The error in the final TLS dataset is due to various from scanner space, with coordinates (x, y, z) in the
factors affecting the measurement and the processing. A ICS, to the corresponding point P, with coordinates
complete description of sources of uncertainty in TLS (X, Y, Z) in the GLCS is given by Equation (2):
data can be found in the work by Paffenholz (2012) and 2 3 2 3 2 3
Cuartero, Armesto, Rodríguez, and Arias (2010). The X X0 x
4Y 5 ¼ 4 Y0 5 þ R4 y 5 (2)
sole focus of this study is on transformational uncer-
tainty, i.e. the error induced by the applied georeferen- Z GLCS Z0 GLCS z ICS
cing method. Random and systematic errors of the TLS,
with
environmentally induced errors or object-related errors
are not addressed in this paper. R¼
2 3
cosψ cosθ sinψ cosφ þ cosψ sinθ sinφ sinψ sinφ þ cosψ sinθ cosφ
6 7
4 sinψ cosθ cosψ cosφ þ sinψ sinθ sinφ cosψ sinφ þ sinψ sinθ cosφ 5
Indirect georeferencing sinθ cosθ sinφ cosθ cosφ

This method is widely adopted for high precision mea- (3)

surement applications. It is based on registration of each where φ, θ and ψ are the roll, pitch and yaw angles of
point cloud using a set of well-spread reflective targets, the TLS in the GLCS and
serving as GCPs. Their coordinates have to be known (X0,Y0,Z0) are coordinates in GLCS of the TLS
both in ICS and in GLCS. Their position in GLCS is center position.
measured by differential GPS or tacheometer, while their In the literature (Alba & Scaioni, 2007; Lichti &
position in ICS is obtained with the TLS by semiauto- Gordon, 2004; Scaioni, 2005), “direct georeferencing”
matic detection and re-scanning at very fine spatial generally refers to a method analogous to total station
resolution. survey, meaning that the TLS is leveled and centered
During the post-processing, the six parameters of over a known point by an optical plummet. Setting
rotation and translation allowing projecting the 3D up the TLS over a known location reduces the
point cloud from ICS to GLCS are computed from the unknown transformation parameters to three angles.
coordinates of the GCPs in both coordinate systems. A If leveling the TLS, it is hypothesized that φ = θ = 0.
minimum of three GCPs has to be measured for each The orientation in the horizontal plane (i.e. azimuthal
scan, but in practice a higher number of GCPs is gen- angle ψ) is carried out by a pointing system (tele-
erally used since redundancy improves the reliability of scopic sight) or by measuring a target.
registration (Alba & Scaioni, 2007; Reshetyuk, 2009). With the development of mobile scanning, the con-
A least squares algorithm is applied to compute cept of direct georeferencing is changing. Commercially
the transformation matrix MIGL providing the best fit available TLS are now indeed generally equipped with
between GCPs position in both coordinate systems. internal positioning sensors, compasses, inclinometers
and gyroscopes. Also, direct georeferencing now
implies real-time georeferencing.
Direct georeferencing
In this paper, “direct georeferencing” refers only to
The MIGL matrix can be decomposed into a rotation approaches using internal sensors (or external sensors
and a translation. The transformation of a point p attached to the TLS) directly providing, in real-time,
158 M. JAUD ET AL.

the six rotation and translation parameters. As using performed with the RiScan Pro® software (Riegl®,
a scope mounted on the TLS or measuring a tie point Austria).
implies post-processing registration of the point
cloud, these methods are not considered in this
paper as direct georeferencing methods, but as Indirect georeferencing
“pseudo-direct georeferencing methods”. Indirect georeferencing is the most traditional
method to register stationary TLS data (Bitelli,
Dubbini, & Zanutta, 2004; Earlie, Young, Masselink,
Field test survey for comparison of & Russell, 2015; Jaud et al., 2011). Creating steady
georeferencing methods known locations for placing targets can be challen-
ging in the coastal environment. Therefore, the con-
Experimental setup figuration of the targets network changes from one
A test survey was carried out in July 2015 to compare session to the next.
the different georeferencing methods herein consid- Typically, 8–15 reflective targets (depending on
ered. The selected test area is the beach of Porsmilin the complexity of the area) are spread throughout
(Figure 2), located near Brest (France). Various surveys the scanned scene. For the present survey, 14 reflec-
(topography, bathymetry and hydrodynamics) are reg- tive targets (depicted by red crosses in Figure 3) were
ularly performed to monitor this beach, which is one of used as GCPs and taken into account for the indirect
the sites of the National Observation System DYNALIT. georeferencing process. In parallel, 14 distinct reflec-
In the present study, the TLS is a Riegl® VZ-400. For tive targets (depicted by blue dots in Figure 3) served
this survey configuration, TLS acquisition involves a as ground validation points (GVPs). These GVPs
360° horizontal and 100° (from 30° to 130°) vertical were used as calibration points to assess the georefer-
scan with an angular resolution of 0.04° in both direc- encing uncertainties.
tions, providing a dense 3D point cloud distributed over Depending on the size and the topography of the
the scanned area. With the aforementioned parameters, scanning scene, placing GCPs, measuring their posi-
a full scan is completed in 9 min and generates more tion in GLCS and gathering them up may be a time-
than 22.5 million points. Taking advantage of existence consuming task. Moreover, the fine-scan process is
of a geodetic marker on the site of Porsmilin to set up one of the longest steps of a TLS acquisition.
our GPS reference station and use Real-Time Kinematic Nevertheless, increasing the number of GCPs
(RTK) GPS positioning, centimetric positioning accu- improves the global redundancy of the observations,
racy can be achieved. At sites without geodetic points, thus reducing the impact of target positioning errors.
post-processing kinematics may be used for GPS data in
order to achieve the same positioning accuracy. For this
test survey, 28 targets were distributed around the TLS Pseudo-direct georeferencing
standpoint. These targets are reflective cylinders 10 cm “Pseudo-direct georeferencing” qualifies an alterna-
in diameter and 10 cm in height. tive referencing method, for the most part based on
The Riegl® VZ-400 TLS can automatically identify the methods proposed by Reshetyuk (2009) and
reflective targets in the point cloud. The results are Mårtensson, Reshetyuk, and Jivall (2012). These
then manually checked before programming the TLS methods are themselves largely based on a geore-
to perform a fine scan of the targets to precisely ferencing approach analogous to procedure used
measure their centroid. for total station survey (Alba & Scaioni, 2007;
The survey is controlled via a PC. Point cloud Lichti & Gordon, 2004; Scaioni, 2005). The TLS is
capture and the following georeferencing steps are centered over a known point and precisely leveled.

Figure 2. (a) Location map of the study area. (b) The test area: The beach of Porsmilin (France).
 EUROPEAN JOURNAL OF REMOTE SENSING 159

Figure 3. Configuration of the targets for indirect georeferencing.

As integrated compasses are generally not accurate We use the dedicated adapter on our TLS to
enough to measure the azimuth orientation, it is mount an external GPS antenna, a Topcon RTK
then performed by aligning the scanner head along GNSS receiver, overtop, as shown in Figure 4. The
a known direction via the measurement of a tie TLS position is thus measured by differential RTK
point. This approach is called “backsighting”. GPS, with centimeter accuracy. The offsets between
The aforementioned challenge in creating per- the GPS antenna center and the scanner center
manent known positions in a coastal setting applies (Figure 4(b)) are extracted from the technical data-
to the scanner station as well. The position of the sheets: the vertical offset is 34.99 cm and the hori-
TLS has therefore to be measured at each survey. zontal offset along X-axis is 1.85 cm.
As was also mentioned by Scaioni (2005), the qual- In this configuration, the TLS tripod mounting
ity of the parts supplied by TLS vendors for direct platform has been coarsely leveled (±1°). Here, it is
georeferencing is generally inadequate for high not necessary to measure the height of the TLS nor
accuracy applications. The roll and pitch para- that of the GPS antenna, since the position of the
meters are measured by internal inclination sensors scanner center is directly obtained from the measured
with a precision of ±0.01°, which can induce errors GPS antenna position. Before the rotation of the
of 1.7 cm at a range of 100 m. The yaw angle, scanner, RTK GPS measurements of the scanner
corresponding to the heading parameter, is mea- position are collected during 5 min.
sured by an integrated compass with a typical Roll and pitch angles are measured by the internal
accuracy of 1°, which can generate errors up to inclinometers. The azimuth (yaw angle) is computed
1.75 m at a range of 100 m. Several repetitive using backsighting techniques. For this purpose, the
tests have shown that the internal sensors are in only reflective target used as GCP, called backsight
reality less precise than the typical accuracy values target, is measured by RTK GPS during 2 min. To
specified by the manufacturer. The internal posi- avoid operator-induced measurement uncertainty,
tion sensor of our instrument (Riegl VZ-400) is an the GPS rover pole can be held on position on a
autonomous GPS L1 receiver. Tests show an uncer- bipod (Figure 5). The sole target is placed about
tainty around 2.5 m with eight satellites in view, 55 m away from the TLS standpoint (Figure 6),
which is insufficient. Since these results do not within unobstructed view of the TLS and the satel-
meet our accuracy requirements, we propose an lites. Since the standard deviation of the scanner
alternative method with a straightforward and azimuth is inversely proportional to the distance to
time-efficient protocol. the backsight target (Reshetyuk, 2009), choosing a
160 M. JAUD ET AL.

Figure 4. (a) Riegl® VZ-400 TLS equipped with a Nikon D200 camera and a Topcon GNSS antenna. (b) Dimensioned front and
top views of the coupled TLS, camera and GPS antenna.

coordinates of the backsight target have to be

known a priori. Moreover, using only one target
yields a significant gain in survey time. It makes
also the logistic of the survey easier, because there
are fewer targets to carry to the different standpoints
distributed over the scanning site, fine scan and mea-
sure with RTK GPS.

Results and discussion

Table 1 shows the comparisons between the TLS center
positions deduced from GPS measurements and com-
puted by indirect georeferencing. The root mean square
(RMS) error between these positions is 0.9 cm. Given
our accuracy expectation on the order of a few centi-
meters, this result fulfills the requirements.
Then, the absolute accuracy of the results is
assessed using the set of GVPs. Because these GVPs
are reflective targets which are not taken into account
in the registration process, 14 GVP are available to
assess indirect georeferencing, while 27 GVPs (by
adding 13 control points from the GCPs used in the
case of indirect georeferencing) are available for
Figure 5. Backsight target accurately measured by RTK GPS,
with GPS pole held by an adjustable bipod.
pseudo-direct georeferencing. The error budget com-
puted from the GVPs can arise from georeferencing
method errors, as well as from intrinsic TLS measure-
ment errors, and GPS errors when measuring the
sufficiently large distance is important to reduce the GVPs position. Results are presented in Table 2.
associated error. However, automatic identification of These RMS errors are both of the order of magnitude
the target centroid can be less accurate if the back- of the RTK GPS error. On average, the RMS error for
sight target is too far from the TLS. the indirect georeferencing approach is 4.4 cm,
The other 27 reflective targets (Figure 6) are used whereas the average RMS error for the pseudo-direct
as GVPs to assess the registration accuracy. georeferencing approach is 3.8 cm (Table 2). The
This method is quite convenient since neither the uncertainty of both indirect and pseudo-direct geor-
coordinates of the TLS station position nor the eferencing methods appears to be very similar, with
 EUROPEAN JOURNAL OF REMOTE SENSING 161

Figure 6. Configuration of the targets for pseudo-direct georeferencing.

Table 1. Comparison of the position of the TLS center com-

puted with the different georeferencing methods. CloudCompare© (Telecom ParisTech - EDF,
TLS position computed by indirect TLS position measured by France). The mean cloud-to-cloud distance between
georeferencing RTK GPS the clouds registered by both georeferencing meth-
XTLS = 132039.928 m XTLS = 132039.927 m
YTLS = 6833812.888 m YTLS = 6833812.879 m
ods, pseudo-direct and indirect, is 2.9 cm with a
ZTLS = 7.002 m ZTLS = 6.989 m standard deviation of 4.2 cm.
The different error assessments (comparative and
absolute) are quite consistent, with the cloud-to-
cloud distance and the absolute accuracy of the mea-
Table 2. Georeferencing uncertainty computed for the surements of the same order. The former is slightly
ground validation points (GVPs). lower than the latter meaning that the difference in
Indirect georeferencing RMS Pseudo-direct georeferencing RMS
error computed on 14 GVP error computed on 27 GVP
output between both georeferencing methods is smal-
XRMS = 4.4 cm XRMS = 3.6 cm ler than the intrinsic error of either method.
YRMS = 2.5 cm YRMS = 2.5 cm Within the overall error budget, the present study
ZRMS = 6.3 cm ZRMS = 5.3 cm
Total RMS = 4.7 cm Total RMS = 4.0 cm is addressing the georeferencing error. The uncer-
tainty of registration is mainly due to the uncertainty
in target position measurements. The indirect geor-
errors of a few centimeters. Horizontal positioning eferencing method is less affected by target measure-
error is lower than vertical error, which is consistent ment errors than the pseudo-direct georeferencing
with the fact that RTK GPS error is lower horizon- method due to GCPs redundancy in the indirect
tally than vertically. georeferencing method. The configuration of the
As the registration error can be anisotropic and GCPs network by the survey operator and the quality
may not be spatially uniform (Lague et al., 2013), of the automatic target detection within the point
point-wise assessment of the error may be insuffi- cloud, performed by the data acquisition software,
cient. The point clouds registered by pseudo-direct can also affect the georeferencing uncertainty. The
georeferencing and indirect georeferencing were main drawback of this method is the time-consuming
compared relative to each other. Both point clouds procedure consisting in laying out the targets, taking
being affected by the same intrinsic TLS measure- a GPS measurement of their position and performing
ment errors, only the differences between both geor- a fine-scan with the TLS.
eferencing methods are thus taken into account. The On the opposite, the pseudo-direct georeferencing
“cloud-to-cloud” distance was computed in approach is very quick since the same steps are required
162 M. JAUD ET AL.

for only one target. To shorten surveying time is critical more accurate leveling does not appear necessary at this
in the coastal environment where survey plans tend to stage. Indeed, a 1° error in platform tilt induces 6 mm of
require accommodating for the tide schedule, as well as uncertainty in TLS center positioning. Such uncertainty is
under certain climates where favorable weather win- acceptable since it is lower than RTK GPS accuracy. The
dows can be quite brief. Naturally, this method is highly measurement of the platform tilt does not impact the
dependent on the accuracy of the position measure- rotation matrix R (Equation (3)) since the roll and pitch
ment for the TLS center and the backsight target. The angles are measured by the internal inclinometers.
cloud-to-cloud mean distance between the unbiased In theory, 3D positioning of one backsight target
and biased point clouds is 1.8 cm, with a standard yields not only the azimuth angle ψ but also the
deviation of 3.3 cm. The distribution of the error is pitch angle θ. Considering both the RTK GPS
very uneven: the points close to the TLS (where the accuracy and the uncertainty on TLS automatic
point cloud is denser) are nearly not affected (less than target centroid detection and comparing with the
1 cm of error up to 10 m from the TLS), while the internal inclinometers accuracy, it appears prefer-
positioning error for the farthest points (up to 400 m) able to use the pitch value from the inclinometer
reach up to 6 m. Depending on the satellite configura- when available. Nevertheless, the pseudo-direct
tion, the observation time of the backsight target may be georeferencing method may still be used with TLS
increased in order to improve the accuracy of the posi- which are not equipped with internal sensors, pro-
tion measurement of the backsight target. For example, vided that the TLS platform is precisely leveled (i.e.
introducing an error of 5 cm in the northing position of in this configuration, φ = θ = 0). Then, RTK GPS
the backsight target (situated 55 m away from the TLS) measurements of the TLS center and of one back-
induces an error of 0.05° in the estimation of yaw angle, sight target are sufficient for direct georeferencing
and thus a bias in the point cloud georeferencing. of the point cloud. It is also possible to carry out
Figure 7(a) shows the effects of (1) positioning errors the Survey without internal inclinometer or precise
of the single backsight target and of (2) the distance leveling of the platform, but by using two backsight
between TLS and the target on the estimation of heading targets in order to obtain the roll, pitch and yaw
angle. It appears that the farther is the backsight target, angles.
the better is the estimation of heading angle. But the The pseudo-direct georeferencing method is pro-
farther the TLS, the less accurate is the automatic estima- posed in this study for a single TLS station survey. It
tion of the target center in the TLS point cloud. can be reproduced for each TLS station, with DGPS
Considering that the TLS precision is 3 mm at 100 m measurement of the TLS position and measurement
of range (RIEGL VZ-400), and the spatial resolution of of the backsight target position. Nevertheless, to opti-
the fine-scan of the target given by Figure 7(b), the mize fieldwork when several TLS stations are needed,
impact of the relative positioning of the target centroid the same backsight target can be used for several
in the point cloud is negligible compared to GPS posi- scans. The backsight target has so to be placed so as
tioning errors. Nevertheless, placing the backsight target to be seen from the different TLS stations. Figure 8
far from the TLS, it is recommended to use a bigger provides some guidance to optimize the fieldwork for
target. multi-station survey in linear and nonlinear context.
In this study, the TLS tripod mounting platform is The position of the TLS (XTLS, YTLS, ZTLS) has to be
coarsely leveled, about ±1°. Due to the small vertical offset measured at each station; however, it may be not
(34.9 cm) between the GPS antenna and the TLS center, necessary to create a backsight target for each station.

Figure 7. (a) Impact of the positioning error of the backsight target (from 1 to 5 cm) and of the distance between the TLS and
the backsight target on the estimation of the heading angle. (b) Impact of the distance between the TLS and the backsight
target on the resolution of the fine-scan of the backsight target.
 EUROPEAN JOURNAL OF REMOTE SENSING 163

Figure 8. Guidance to optimize fieldwork for multi-station survey with pseudo-direct georeferencing in linear (a) and nonlinear
(b) contexts. BT is the position of the backsight target. At each station, the position of the TLS (XTLS, YTLS, ZTLS) has to be
measured.

Conclusion Disclosure statement

High-resolution and high-accuracy data from TLS No potential conflict of interest was reported by the
surveys are now commonplace in coastal monitoring. authors.
A well-defined acquisition procedure can signifi-
cantly shorten surveying time, which can be critical
again in the coastal environment. The choice of the Funding
georeferencing method depends on (1) the accuracy This work was supported by the “Laboratoire d’Excellence”
required and (2) the constraints imposed by the study LabexMER [ANR-10-LABX-19] and co-funded by a grant
area. from the French government under the program
This paper assesses the performance of a field “Investissements d’Avenir”.
protocol aimed at reducing survey time without
increasing georeferencing errors. ORCID
Considering the suboptimal quality of the internal
sensors supplied by vendors, it is not possible to carry Marion Jaud http://orcid.org/0000-0002-7629-9710
out real-time direct georeferencing. Consequently, a
pseudo-direct georeferencing method is proposed. The
protocol involves internal inclinometers to measure roll References
and pitch angles and a RTK GPS to measure the position Abellán, A., Oppikofer, T., Jaboyedoff, M., Rosser, N.J.,
of the TLS center and the position of one backsight target. Lim, M., & Lato, M.J. (2014). Terrestrial laser scanning
This method is much quicker than classical indirect geor- of rock slope instabilities. Earth Surface Processes and
eferencing. The cloud-to-cloud mean distance resulted in Landforms, 39(1), 80–97. doi:10.1002/esp.v39.1
Alba, M., Giussani, A., Roncoroni, F., & Scaioni, M. (2007).
2.9 cm, while the intrinsic error is lower for the pseudo-
Review and comparison of techniques for terrestrial 3D-
direct georeferencing method. Assessing the accuracy of View georeferencing. In Proceedings of the 5th
each method with GVPs, we get an average RMS error of International Symposium on Mobile Mapping
4.4 cm for classical indirect georeferencing, and of 3.8 cm Technology (ISPRS). Padua, Italy.
for pseudo-direct georeferencing. Accordingly, despite Alba, M., & Scaioni, M. (2007). Comparison of techniques
the low redundancy of the pseudo-direct georeferencing for terrestrial laser scanning data georeferencing applied
to 3-D modelling of cultural heritage. The International
method (only two points are positioned through the RTK Archives of the Photogrammetry, Remote Sensing and
technique, the TLS center and the backsight target), this Spatial Information Sciences, 36, XXXVI-5/W47.
method proved to be more precise than the indirect Alho, P., Vaaja, M., Kukko, A., Kasvi, E., Kurkela, M.,
georeferencing approach. Hyyppä, J., . . . Kaartinen, H. (2011). Mobile laser scan-
ning in fluvial geomorphology: Mapping and change
detection of point bars. Zeitschrift Für Geomorphologie,
55, 31–50. doi:10.1127/0372-8854/2011/0055S2-0044
Acknowledgments Barber, D., Mills, J., & Smith-Voysey, S. (2008). Geometric
validation of a ground-based mobile laser scanning sys-
This work was supported by the “Laboratoire d’Excellence” tem. ISPRS Journal of Photogrammetry and Remote
LabexMER (ANR-10-LABX-19) and co-funded by a grant Sensing, 63, 128–141. doi:10.1016/j.isprsjprs.2007.07.005
from the French government under the program Besl, P.J., & McKay, N.D. (1992). A method for registration
“Investissements d’Avenir”. of 3-D shapes. IEEE Transactions on Pattern Analysis
164 M. JAUD ET AL.

and Machine Intelligence, 14(2), 239–256. doi:10.1109/ Lichti, D.D., & Gordon, S.J. (2004). Error propagation in
34.121791 directly georeferenced terrestrial laser scanner point
Bitelli, G., Dubbini, M., & Zanutta, A. (2004). Terrestrial clouds for cultural heritage recording. In The proceedings
laser scanning and digital photogrammetry techniques to of FIG Working Week. International Federation of
monitor landslide bodies. In Proceedings of the XXth Surveyors, Athens, Greece.
ISPRS Congress Geo-Imagery Bridging Continents. Lim, M., Dunning, S.A., Burke, M., King, H., & King, N.
International Society of Photogrammetry and Remote (2015). Quantification and implications of change in
Sensing (ISPRS), Istanbul, Turkey. organic carbon bearing coastal dune cliffs: A multiscale
Cuartero, A., Armesto, J., Rodríguez, P.G., & Arias, P. analysis from the Northumberland coast, UK. Remote
(2010). Error analysis of terrestrial laser scanning data Sensing of Environment, 163, 1–12. doi:10.1016/j.
by means of spherical statistics and 3D graphs. Sensors, rse.2015.01.034
10, 10128–10145. doi:10.3390/s101110128 Mancini, F., Dubbini, M., Gattelli, M., Stecchi, F., Fabbri,
Delacourt, C., Allemand, P., Jaud, M., Grandjean, P., S., & Gabbianelli, G. (2013). Using unmanned aerial
Deschamps, A., Ammann, J., . . . Suanez, S. (2009). vehicles (UAV) for high-resolution reconstruction of
DRELIO: An unmanned helicopter for imaging coastal topography: The structure from motion approach on
areas [Special issue]. Journal of Coastal Research, 56, coastal environments. Remote Sensing, 5, 6880–6898.
1489–1493. doi:10.2307/25738037 doi:10.3390/rs5126880
Earlie, C.S., Young, A.P., Masselink, G., & Russell, P.E. Mårtensson, S.G., Reshetyuk, Y., & Jivall, L. (2012).
(2015). Coastal cliff ground motions and response to Measurement uncertainty in network RTK GNSS-based
extreme storm waves. Geophysical Research Letters, 42 positioning of a terrestrial laser scanner. Journal of
(3), 847–854. doi:10.1002/2014GL062534 Applied Geodesy, 6, 25–32. doi:10.1515/jag-2011-0013
Giussani, A., & Scaioni, M. (2004). Application of TLS to Nield, J.M., Wiggs, G.F., & Squirrell, R.S. (2011). Aeolian
support landslides study: Survey planning, operational sand strip mobility and protodune development on a
issues and data processing. International Archives of drying beach: Examining surface moisture and surface
Photogrammetry, Remote Sensing and Spatial roughness patterns measured by terrestrial laser scan-
Information Sciences, 36, 318–323. ning. Earth Surface Processes and Landforms, 36(4), 513–
Harwin, S., & Lucieer, A. (2012). Assessing the accuracy of 522. doi:10.1002/esp.2071
georeferenced point clouds produced via multi-view Olsen, M., Johnstone, E., Driscoll, N., Ashford, S., &
stereopsis from unmanned aerial vehicle (UAV) ima- Kuester, F. (2009). Terrestrial laser scanning of extended
gery. Remote Sensing, 4, 1573–1599. doi:10.3390/ cliff sections in dynamic environments: Parameter ana-
rs4061573 lysis. Journal of Surveying Engineering, 135(4), 161–169.
Jaud, M., Delacourt, C., Allemand, P., Deschamps, A., doi:10.1061/(ASCE)0733-9453(2009)135:4(161)
Cancouët, R., Ammann, J., & Cuq, V. (2011, Olsen, M., Johnstone, E., Kuester, F., Driscoll, N., &
December). Comparison of some very high resolution Ashford, S. (2011). New automated point-cloud align-
remote sensing techniques for the monitoring of a sandy ment for ground-based light detection and ranging data
beach. Poster session presented at the AGU Fall Meeting, of long coastal sections. Journal of Surveying
San Francisco, CA. Engineering, 137(1), 14–25. doi:10.1061/(ASCE)
Kasperski, J., Delacourt, C., Allemand, P., Potherat, P., SU.1943-5428.0000030
Jaud, M., & Varrel, E. (2010). Application of a terrestrial Paffenholz, J.-A. (2012). Direct geo-referencing of 3D point
laser scanner (TLS) to the study of the Séchilienne land- clouds with 3D positioning sensors (Doctoral dissertation,
slide (Isère, France). Remote Sensing, 2, 2785–2802. Deutsche Geodätische Kommision (DGK) in the
doi:10.3390/rs122785 Bayrischen Akademie der Wissenschaften, München,
Kuhn, D., & Prüfer, S. (2014). Coastal cliff monitoring and Germany). Retrieved from https://www.researchgate.net/
analysis of mass wasting processes with the application Pietro, L.S., O’Neal, M.A., & Puleo, J.A. (2008). Developing
of terrestrial laser scanning: A case study of Rügen, terrestrial-LIDAR-based digital elevation models for
Germany. Geomorphology, 213, 153–165. doi:10.1016/j. monitoring beach nourishment performance. Journal of
geomorph.2014.01.005 Coastal Research, 24(6), 1555–1564. doi:10.2112/07-
Kukko, A., Kaartinen, H., Hyyppä, J., & Chen, Y. (2012). 0904.1
Multiplatform mobile laser scanning: Usability and perfor- Quinn, J.D., Rosser, N.J., Murphy, W., & Lawrence, J.A.
mance. Sensors, 12, 11712–11733. doi:10.3390/s120911712 (2010). Identifying the behavioural characteristics of clay
Lague, D., Brodu, N., & Leroux, J. (2013). Accurate 3D cliffs using intensive monitoring and geotechnical
comparison of complex topography with terrestrial laser numerical modelling. Geomorphology, 120(3), 107–122.
scanner: Application to the Rangitikei canyon (N-Z). doi:10.1016/j.geomorph.2010.03.004
ISPRS Journal of Photogrammetry and Remote Sensing, Reshetyuk, Y. (2009). Self-calibration and direct georeferen-
82, 10–26. doi:10.1016/j.isprsjprs.2013.04.009 cing with GPS in terrestrial laser scanning (Doctoral
Leroux, J. (2013). Chenaux tidaux et dynamique des prés- dissertation, Royal Institute of Technology (KTH),
salés en régime méga-tidal: Approche multi-temporelle du Stockholm, Sweden). Retrieved from http://citeseerx.ist.
siècle à l’événement de marée. (Doctoral dissertation, psu.edu/
Université Rennes 1 - Université Européenne de RIEGL VZ-400. 3D Terrestrial Laser Scanner RIEGL VZ-
Bretagne, Rennes, France). Retrieved from https://tel. 400 /RIEGL VZ-1000 (version 03/09 CE – Rev. 10/11/
archives-ouvertes.fr/tel-01005360/ 2014).
Letortu, P., Costa, S., Maquaire, O., Delacourt, C., Augereau, Rosser, N.J., Brain, M.J., Petley, D.N., Lim, M., & Norman,
E., Davidson, R., . . . Nabucet, J. (2015). Retreat rates, mod- E.C. (2013). Coastline retreat via progressive failure of
alities and agents responsible for erosion along the coastal rocky coastal cliffs. Geology, 41(8), 939–942. doi:10.1130/
chalk cliffs of Upper Normandy: The contribution of ter- G34371.1
restrial laser scanning. Geomorphology, 245, 3–14. Rosser, N.J., Petley, D.N., Lim, M., Dunning, S.A., &
doi:10.1016/j.geomorph.2015.05.007 Allison, R.J. (2005). Terrestrial laser scanning for
 EUROPEAN JOURNAL OF REMOTE SENSING 165

monitoring the process of hard rock coastal cliff erosion. Schubert, J.E., Gallien, T.W., Majd, M.S., & Sanders, B.F.
Quarterly Journal of Engineering Geology and (2015). Terrestrial laser scanning of anthropogenic beach
Hydrogeology, 38(4), 363–375. doi:10.1144/1470-9236/ berm erosion and overtopping. Journal of Coastal
05-008 Research, 31(1), 47–60. doi:10.2112/JCOASTRES-D-14-
Scaioni, M. (2005). Direct georeferencing of TLS in survey- 00037.1
ing of complex sites. In The Proceedings of the ISPRS Schürch, P., Densmore, A.L., Rosser, N.J., Lim, M., &
Working Group V/4 Workshop 3D-ARCH “Virtual McArdell, B.W. (2011). Detection of surface change in
Reconstruction and Vizualization of Complex complex topography using terrestrial laser scanning:
Architectures”. International Society of Application to the Illgraben debris-flow channel. Earth
Photogrammetry and Remote Sensing (ISPRS), Mestre- Surface Processes and Landforms, 36(14), 1847–1859.
Venice, Italy. doi:10.1002/esp.2206