Mobile mapping system performance – an initial investigation
into the effect of vehicle speed on laser scan lines
C.Cahalane, T.McCarthy and C.McElhinney
National Centre for Geocomputation, NUI Maynooth, Kildare, Ireland
Mobile mapping systems (MMS) are becoming an increasingly popular method for collecting
high quality near-3D information in terrestrial environments. One of the primary reasons for
this is technological advances in laser scanning. When a modern laser scanner is mounted on
a moving platform and combined with a GPS and navigation system, mobile mapping
systems can produce millions of geo-referenced points per minute which can then be used to
create accurate near-3D models. The development of processing algorithms for these point
clouds has been the focus of the research community to date. However, given an arbitrary
known static object positioned at a specific distance away from a moving mobile mapping
system the resolution and accuracy of the resulting point cloud which will describe the object
is unknown. It is this resolution and accuracy which is the underlying limit of these point
cloud processing algorithms. We are in the process of developing a method for determining
the quantitative resolution and accuracy of point clouds collected by a mobile mapping
system with respect to known objects. In this paper, we will demonstrate our initial
investigation into the effect that vehicle speed has on laser scan lines. Speed has an impact on
the physical distance between sequential laser scan lines and will also influence the angle of
individual scan lines. We have developed a system to calculate this information about laser
scan lines based on the position and orientation of the laser scanner on the vehicle and the
vehicles speed. We will verify our equations and analysis by comparing our simulated data to
the point cloud data collected by our XP-1 mobile mapping system.
The focus of the research community to date has largely been on developing automated or
semi automated algorithms for processing the large point clouds captured by modern
terrestrial or mobile mapping systems (Becker and Haala, 2009, Hammoudi et al., 2009, Pu
and Vosselman, 2007). However, other than accuracy tests on specific systems (Barber et al.,
2008, Haala et al., 2008) little research exists assessing the performance of generic mobile
mapping systems. Further research in this area is important as one of the underlying problems
facing research groups working with extraction algorithms is how many scan lines (or
profiles) will hit an object at a certain range, and how many points can they expect to return
from each profile. For example, work by Kukko et al., (2009) and Lehtomaki et al., (2010)
require a minimum number of profiles on post objects for them to be detected. Circular
objects need a minimum number of points on each profile to recognise a circular shape. Each
algorithm performs differently, and from Kaartinen et al., (2005) we can see that the point
density directly impacts on the accuracy of the resulting extracted model. Mobile mapping
systems (MMS) are new to the market, and to date there has been no concerted effort to
assess their combined capabilities. This paper will focus entirely on laser based systems.
One of the fundamentals for a laser based mobile mapping system is the location and
orientation of the scanner on the vehicle. Although there have been tests investigating the
best scanner configuration to minimise occlusions (Yoo et al., 2009), there does not appear to
have been research carried out to find the optimal location for a single scanner (i.e. rear, side,
front) that will provide the highest point density. The solution to date has been to increase the
number of scanners. Our system is a single scanner system, so we hope to provide a definitive
view of the capabilities of such a system which we anticipate will then be of use to systems
with more hardware. Scanner orientation is also of importance. Scan lines cannot be
perpendicular to the direction of travel or they will miss objects whose sides are also
perpendicular to it. A horizontal rotation of the scanner solves this problem, and a vertical
rotation deals with structures above the vehicle which would otherwise be missed, such as
overhead road signs, bridge faces etc. We hope to be able to define what the optimum
When safe to do so, mobile mapping systems are capable of operating at highway speeds.
However, point density decreases as vehicle velocity increases and this necessitates multiple
passes to ensure a dense point cloud (multiple passes are also employed to ensure all sides of
an object are captured) that will meet project specifications. To ensure a high point density,
projects have been carried out at low speed (Goulette et al., 2006, Graefe, 2007), which in a
commercial situation would impact on the productivity of a MMS. It is our hope that when
completed our work will allow us define the maximum speed for specific scanner
configurations that will provide a required point density, and also define the minimum
number of passes required. This should help to minimise survey time and also the file size of
To date there has been some interesting work in this area. Kukko et al., (2007) and Hesse and
Kutterer (2007) have detailed profile spacing at various mirror speeds and vehicle velocities.
We hope to improve on this by providing a generic formula which will work for any mirror
speed, vehicle velocity and importantly, will incorporate scanner orientation into the system.
Hoffman and Brenner (2009) have included in their work on theoretic point density some
interesting results on the effect change in vehicle direction and velocity has on scan lines.
In section 2 we will look at mobile mapping systems in general and the platform we have
developed at StratAG, followed in section 3 by the theory and processes behind our work. In
section 4 we will present the results of our test data, and finally in section 5, our conclusions.
2 Mobile Mapping & Experimental Platform
MMS enable high density spatial data to be collected along route networks and in urban
environments. These data can then be utilised in a number of ways, such as route safety
audits, road authorities GIS, infrastructure surveys and change detection for national mapping
agencies. Combining high accuracy GPS/INS, LiDAR and imaging sensors onboard a
moving platform enable surveys to be carried out rapidly with significant cost savings. Land
based MMS compliment existing ground based survey and aerial surveying activities in a
number of ways. Large scale detail such as road sign detail or detailed infrastructure
condition can be recorded. Additionally, extensive ground control is not required and these
systems can capture features that are sometimes obscured from aerial platforms.
The multi-disciplinary research group StratAG, established to research advanced
geotechnologies at NUI Maynooth have recently completed design and development of a
multipurpose, state of the art, land based experimental platform (XP-1) Mobile Mapping
System as shown in Figure 1. The primary components of the XP-1 are an IXEA LandINS
GPS/INS, a Riegl VQ-250 300KHz laser scanner and an imaging system consisting of 6
progressive-scan cameras. Additional imaging sensors include a FLIR thermal (un-cooled)
SC-660 camera and an innovative 5-CCD multispectral camera capable of sensing across
blue, green, red and two infrared bandwidths.
Figure 1. The XP1 Mobile Mapping System.
3 Laser scan line theory
This paper details our initial investigation into the effects of vehicle speed on scan lines for
certain scanner mirror speeds, positions and orientations. For these initial steps, a number of
necessary assumptions were made. The first assumption was that the vehicle was operating
on a plane surface, so road cross fall or surface deformities could be ignored. The second was
that the vehicle maintains a constant speed between each laser profile. The final assumption
was that we would not take into account occlusions, traffic or obstructions. Once these
assumptions had been made, the factors impacting upon scan lines were identified.
3.1 Profile spacing
Before exploring the effect of scanner rotations on profile spacing, it is important to define
what we mean by a vertical and horizontal rotation of the laser scanner. On the XP1, the
scanner is located at the rear of the vehicle. Figure 2, displays the our laser
scanner/GNSS/INS mount, and the 45° horizontal and 45° vertical rotations of the scanner. It
can be seen in Figures 3 and 4 the effect that horizontal and vertical rotations of the scanner
have on profile spacing. The distance travelled, d, is no longer the profile spacing. The profile
spacing is now a function of the scanner angle, mirror speed and vehicle velocity.
Figure 2. XP1 Laser scanner rotations: (a) top view (b) side view.
Figure 3. Profile effect due to vertical scanner rotation: (a) overview (b) detail view
Figure 4. Profile effect due to horizontal scanner rotation: (a) overview (b) detail view
Once the variables affecting the profile spacing were identified (see Table 1), it was possible
to create a formula to calculate profile spacing for a generic system.
Table 1. Profile spacing variables.
Speed in m/s
Mirror rotation frequency
Distance travelled between mirror rotations
Laser scanner horizontal-rotation angle
Laser scanner vertical-rotation angle
Profile distance on ground surface
Profile distance on vertical structures
To calculate profile spacing, the distance covered in one mirror rotation for any speed is
required (Equation 1).
d = v/f (1)
From this, the effect of the scanner rotation is calculated by trigonometry (Equation 2) for a
right angled triangle, as shown in Figures 3 and 4.
Ph = cosθh.d or Pv = cosθv.d (2)
3.2 Motion effect
In this paper, the effect of motion on profiles refers to their angular change due to the speed
of a vehicle. A stationary vehicle exhibits profiles whose angle is entirely dependent on the
scanner angle, however, when the MMS is in motion this will change. In our system, it takes
0.01s for each mirror rotation. During this time the vehicle will have moved, altering the
angle of the scan line, as shown in Figure 5. We have designed a system to calculate this
effect for a generic system, based on scanner height, vertical scanner rotation, vehicle speed
and mirror speed. The variables involved in this calculation can be found in Table 2.
Figure 5. Motion effect: (a) overview (b) detail view
Table 2. Motion effect variables.
Laser scanner vertical-rotation angle
Scanner height above ground
Distance travelled between mirror rotations
Motion amended angle
For this initial investigation, the effect of motion on profiles can be viewed as part of a group
of parallelograms and triangles (Figure 5). Using trigonometric ratios, the hypotenuse ‘Y’ of
the triangle formed with the vehicle is found (Equation 3). Accurate measurement of the
vehicle height (h) is important.
Y = h / (sin (90° – θv) (3)
In this calculation, because using the distance travelled during one full mirror rotation would
mean that we would now be on a new scan line we only deal with the bottom half of the
circular scan, hence the use of d/2. This is followed by use of the cosine rule to calculate the
short diagonal length ‘X’ in the parallelogram (Equation 4). This value is one of the variables
for the last triangle, and by using the sine rule θm is found (Equation 5).
X = √ ((d/2) 2 + (Y)2 – 2(d/2)(Y)cos (90°- θv)) (4)
θm = sin-1 ((sin (90°- θv) * d/2)/x) (5)
We will demonstrate the capabilities of this method using two datasets. One, a theoretical test
dataset verified with CAD models and the second an actual dataset from a MMS. As our test
dataset, we applied our formula to a selection of velocities, varying mirror rotation speeds
and varying vertical and horizontal rotations of the laser scanner. This theoretical test dataset
was compared against CAD models and in each case it agreed with our results. To improve
on this and provide a real world test of our system, we chose to verify our equations and
hypotheses by comparing our simulated data to a point cloud captured by our XP1 mobile
mapping system. Since relative accuracy and not absolute accuracy is what is important to
this study, we ignored the GPS conditions and quality of the combined navigation solution.
By analysing the navigation files, varying areas velocity from 5m/s to 16m/s were identified.
By plotting this navigation data on the survey data and selecting straight sections of road
only, errors due to course deviation could be kept to a minimum. The reference data exhibited
a point spacing that one would expect from a laser scanner according to its specifications.
This made accurate measurements difficult, so we interpolated a 2d linear fit (shown in
Figure 5) as it was the most suitable approximation of the scan lines that we were
investigating and our matches our assumption that the road is a flat plane.
Figure 6. Scan line measurements: (a) linear interpolation (b) subsequent measurements
After analysing the data, and plotting the predicted versus measured values for ground
profiles(Figure 7), it becomes clear that our system is performing satisfactorily. Measured
profile spacings ranged from 0.031m at 4.1m/s to 0.12m at 16.34 m/s. Overall the predicted
figures matched the measured values, with a minimum percentage error of 0%, an average
percentage error of 2.68% and a maximum of 6%. All measurements were within 5mm of
their predicted value. Outliers can be attributed to errors in measurement of the reference data
or possibly from the linear interpolation process of the scan lines. This is something that we
hope to improve on in the near future.
Initial inspection of the results from the motion effect tests identified an issue (see Figure 8).
However, it soon became apparent that there was a problem with the measured rather than
our predicted values. Minus values for angular changes had been recorded and these should
not have been present unless the vehicle velocity was negative – i.e. reversing. This suggests
the error was made when selecting the navigation data for comparison. We believe that a
more rigorous approach to ensuring there was no course deviation between scan lines will
lead to better results in the future. It also became apparent that a θm value of over 90° will
cause our system to fail – due to the change of trigonometric quadrant. However, this is
unlikely to happen in practice as the mirror speed would have to be extremely low and the
vehicle speed extremely high, but is worth noting.
6.9339 6.9339 6.9339 6.9339 6.9339 6.9339
y = - 0.48355*x + 1.0735e+006
Figure 7. Ground profile spacing due to vehicle speed and 45° horizontal scanner rotation.
Figure 8. Angular change of scan lines due to 45° vertical scanner rotation and vehicle
This study has proven by validating with real world data that our system is capable of
predicting profile spacing for differing vehicle speeds, mirror frequencies and scanner
rotations. Although initial tests were promising, the results from real world data of our
motion effect tests demonstrates that a more rigorous approach to selecting areas of 0° course
deviation needs to be taken we believe that we have identified the cause of error and will be
able to rectify this in the near future. Further to this, future work will involve eliminating the
error sources in the profile spacing process and attempting to incorporate point spacing into
the profile spacing system.
The authors would like to acknowledge the support received from the Irish Research Council
for Science, Engineering and Technology(IRCSET) and the Enterprise Partner, Pavement
Management Systems Ltd.
6 8 10 12 14 16
Profile Spacing (m)
6 8 10 12 14 16
Angular Change (dec. deg.)
Barber, D. Mills, J. and Smith-Voysey, S., 2008. Geometric validation of a ground-based mobile laser
scanning system, ISPRS Journal of Photogrammetry and Remote Sensing, 63, (1), 128-141.
Becker, S. & Haala, N., 2009. Grammar Supported Facade Reconstruction from Mobile LIDAR
Mapping. In International Archives of Photogrammetry, Remote Sensing and Spatial Information
Sciences, Vol.XXXVIII, Part3/W4 (pp. 65-70). Paris, France.
Goulette, F., Nashashibi, F., Abuhadrous, I., Ammoun, S. and Laurgeau, C., 2006. An integrated on-
board laser range sensing system for on-the-way city and road modelling. In: ISPRS Commission I
Symposium, “From Sensors to Imagery”. Paris, 3-5.7.2006. Internationa Archives of
Photogrammetry, Remote Sensing and Spatial Information Sciences 36 (1B).
Gräfe, G., 2007. Kinematic surveying with static accuracy. In: Grün, A., and H. Kahmen, (Eds.) 8th
Conference on Optical 3-D Measurement Techniques, July 9- 12, 2007 - ETH Zurich, Switzerland.
Haala, N., Peter, M., Kremer, J., Hunter, G., Stuttgart, U. and Cloud, P., 2008. Mobile lidar mapping
for 3d point cloud collection in urban areas a performance test. In: ISPRS Congress, ISPRS, Beijing,
China, pp. 1119–1124.
Hammoudi, K., Dornaika, F. and Paparoditis, N., 2009. Extracting building footprints from 3D point
clouds using terrestrial laser scanning at street level. In International Archives of Photogrammetry,
Remote Sensing and Spatial Information Sciences, Vol.XXXVIII, Part3/W4 (pp. 65-70). Paris,
Hesse, C. and Kutterer, H., 2007. A mobile mapping system using kinematic terrestrial laser scanning
(KTLS) for image acquisition. In: Grün, A., and H. Kahmen, (Eds.) 8th Conference on Optical 3-D
Measurement Techniques, July 9- 12, 2007 - ETH Zurich, Switzerland.
Kaartinen, H., Hyyppä, J., Gülch, E., Vosselman, G., Hyyppä, H., Matikainen, L., Hofmann, A.D.,
Mäder, U., Persson, Å., Söderman, U., Elmqvist, M., Ruiz, A., Dragoja, M., Flamanc, D., Maillet, G.,
Kersten, T., Carl, J., Hau, R.,Wild, E., Frederiksen, L., Holmgaard, J. and Vester, K., 2005. Accuracy
of 3D city models: EuroSDR com- parison. International Archives of Photogrammetry, Remote Sens-
ing and Spatial Information Sciences 36 (Part 3/W19), 227–232.
Kukko, A., Jaakola, A., Lehtomäki, M. and Kaartinen, H., 2009. Mobile mapping system and
computing methods for modelling of road environment. 2009 Joint Urban Remote Sensing Event, 20-
22, May Shanhgai, China.
Kukko, A., Andrei, C.-O., Salminen, V.-M., Kaartinen, H., Chen, Y., Rönnholm, P., Hyyppä, H.,
Hyyppä, J., Chen, R., Haggrén, H., Kosonen, I. and Čapek, K., 2007. Road Environment Mapping
System of the Finnish Geodetic Institute – FGI Roamer. In Proceedings of the ISPRS Workshop
„Laser Scanning 2007 and SilviLaser 2007‟, Espoo, Finland, 12-14 Sept. 2007; Rönnholm, P.;
Hyyppä, H.; Hyyppä, J. Eds.; pp. 241-247.
Lehtomäki, M., Jaakola, A., Hyyppä, J., Kukko, A. and Kaartinen, H., 2010. Detection of Vertical
Pole-Like Objects in a Road Environment Using Vehicle-Based Laser Scanning Data. Remote
Sensing, 2(3), 641-664.
Pu, S., and Vosselman, G., 2007. Extracting windows from terrestrial laser scanning. In International
Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. XXXVI, Part
3/W52 (pp. 320-325). Espoo, Finland.
Yoo, H.J., Goulette, F., Senpauroca, J., Lepère and G., 2009. Comparative analysis based on
simulation for the design of laser terrestrial mobile mapping systems, 6th International Symposium on
Mobile Mapping Technology, 2009, Sao Paolo.