EARTH SURFACE PROCESSES AND LANDFORMS
Earth Surf. Process. Landforms 34, 1831–1838 (2009)
Copyright © 2009 John Wiley & Sons, Ltd.
Published online in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/esp.1879
Letters to ESEX
Application of boat-based laser scanning for
river survey
P. Alho,
1
* A. Kukko,
2
H. Hyyppä,
3
H. Kaartinen,
2
J. Hyyppä
2
and A. Jaakkola
2
1
University of Turku, Department of Geography, FI-20014 University of Turku, Finland
2
Finnish Geodetic Institute, Department of Remote Sensing and Photogrammetry, Geodeetinrinne 2, P.O. Box 15, FI-02431
Masala, Finland
3
Helsinki University of Technology, Research Institute of Modelling and Measuring for the Built Environment, P.O. Box 1200,
FI-02015 TKK, Finland
Received 12 December 2008; Revised 22 June 2009; Accepted 29 June 2009
*Correspondence to: P. Alho, University of Turku, Department of Geography, FI-20014 University of Turku, Finland. E-mail: petteri.alho@utu.fi
Abstract: The boat-based, mobile mapping system (BoMMS) with a laser scanner allows the derivation of detailed riverine topo-
graphical data for fluvial applications. Combined with data acquisition from static terrestrial LiDAR (light detection and range) or
mobile terrestrial LiDAR on the ground, boat-based laser scanning enables a totally new field mapping approach for fluvial studies.
The BoMMS approach is an extremely rapid methodology for surveying riverine topography, taking only 85 min to survey a reach
approximately 6 km in length. The BoMMS approach also allowed an effective survey angle for deep river banks, which is dif-
ficult to achieve with aerial or static terrestrial LiDAR. Further, this paper demonstrates the three-dimensional mapping of a point-
bar and its detailed morphology. Compared with the BoMMS surface, approximately, 80% and 96% of the terrestrial LiDAR points
showed a height deviation of less than 2 cm and 5 cm, respectively, with an overall standard deviation of ± 2·7 cm. This level
of accuracy and rapidity of data capture enables the mapping of post-flood deposition directly after a flood event without an
extensive time lag. Additionally, the improved object characterisation may allow for better 3D mapping of the point bar and other
riverrine features. However, the shadow effect of the BoMMS survey in point bar mapping should be removed by additional
LiDAR data to acquire entire riverine topography. The approach demonstrated allowed a large reach to be surveyed compared
with static terrestrial LiDAR and increased the spatial limit of survey towards aerial LiDAR, but it maintains the same or even
better temporal resolution as static terrestrial LiDAR. Copyright © 2009 John Wiley & Sons, Ltd.
KEYWORDS: laser scanning; LiDAR; river morphology; TLS; mobile mapping
Introduction and background
In fluvial studies, different survey and modelling approaches
have been used to study the interaction of landscape and flow
processes, including response thresholds, feedback elements
and other such complexities, requiring high-quality topo-
graphical and bathymetrical data at different scales. Currently,
synoptic aerial photogrammetry, tachymetry and GPS surveys
are widely used in fluvial geomorphology (Brasington et al.,
2000; Fuller et al., 2003, Lane et al., 2003), while more
sophisticated survey methods such as close-range photogram-
metry (Lane et al., 1994; Butler et al., 1998; Smith et al., 2009)
and terrestrial LiDAR (Heritage and Hetherington, 2007) are
less common.
Fluvial processes have been efficiently simulated using
hydraulic models, which require digital terrain models or
digital elevation models (DTMs, DEMs) as input data. The
effects of DTM characteristics in fluvial studies and particu-
larly in flood inundation modelling have been investigated in
a number of studies (Néelz et al., 2006; Sanders, 2007). More
recently, remote sensing techniques have also been utilised to
acquire high-resolution digital terrain models using aerial
LiDAR, or in reconstructing past inundations with either air-
borne synthetic aperture radar (SAR) or airborne photogram-
metry (Néelz et al., 2006). In a number of studies, it has been
reported that DTM accuracy is crucial for fluvial geomorpho-
logical mapping and hydraulic modelling (Cobby et al., 2001;
Bates, 2004; Alho et al., 2009). Therefore, a number of
researchers have sought to develop better approaches for
acquiring topographic data for fluvial studies (Alsdorf et al.,
2000; Cobby et al., 2001; Bates et al., 2003, Hudson and
Colditz, 2003; Fonstad and Marcus 2005; Heritage and
Hetherington, 2007).
Promising results have been obtained by various authors
using satellite remote sensing data (Bates, 2004) or highly
accurate aerial LiDAR DTMs (Cobby et al., 2001; French,
2003; Mason et al., 2007), rather than the more traditional
ground or national survey maps. Nevertheless, ground surveys
could be applied using terrestrial LiDAR, based on distance
measurements and the precise orientation of these measure-
ments between a sensor at a known position and a reflecting
object (the position to be defined). This technique has been
shown to be useful in producing high-quality 3-D models of
forested environments (Hyyppä et al., 2001; Yu et al., 2006;