Heritage hetherington lidar_pdf[1]

Earth Surface Processes and Landforms
Laser Surf. Process. fluvial geomorphology
Earth scanning in Landforms (in press) 1
Published online in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/esp.1375

Towards a protocol for laser scanning in fluvial
geomorphology
George Heritage* and David Hetherington
Built and Human Research Institute, School of Environment and Life Sciences, Peel Building, University of Salford, Manchester, M5 4WT, UK

*Correspondence to: Abstract
G. L. Heritage, Built and
Human Research Institute, Advances in spatial analytical software allow digital elevation models (DEMs) to be pro-
school of Environment and duced which accurately represent landform surface variability and offer an important
Life Sciences, Peel Building, opportunity to measure and monitor morphological change and sediment transfer across a
University of Salford, variety of spatial scales. Many of the techniques presently employed (aerial LIDAR, EDM
Manchester, M5 4WT, UK. theodolites, GPS, photogrammetry) suffer coverage or resolution limitations resulting in a
E-mail: trade-off between spatial coverage and morphologic detail captured. This issue is parti-
G.L.Heritage@salford.ac.uk cularly important when rates of spatial and temporal change are considered for fluvial
systems. This paper describes the field and processing techniques required for oblique laser
scanning to acquire 0·01 m resolution digital elevation data of an upland reach of the River
Wharfe in the UK. The study site is variable with rapidly changing morphology, diverse
vegetation and the presence of water, and these are evaluated with respect to laser data
accuracy. Scan location, frequency and distance are discussed with reference to survey accu-
racy and efficiency, and a field protocol is proposed. Scan data cloud merging was achieved
with a high degree of precision (sub-centimetre) and positional data are shown to be very
accurate for exposed surfaces. Vegetation and water decrease the accuracy, as the laser pulse
is often prevented from reaching the ground surface or is not returned. Copyright © 2006
Received 22 June 2005; John Wiley & Sons, Ltd.
Revised 30 January 2006; Keywords: terrestrial LIDAR; terrestrial laser scanning; oblique LIDAR; oblique laser
Accepted 9 February 2006 scanning; River Wharfe; geomorphological survey

Introduction
Sophisticated spatial analytical equipment and software now allow digital elevation models (DEMs) to be constructed
that accurately represent landform surface variability and offer an excellent opportunity to measure and monitor
morphological change across a variety of spatial scales (Brasington et al., 2000; Lane and Chandler, 2003; Fuller
et al., 2005). Coupled with this, the development of increasingly sophisticated surveying equipment (aerial LIDAR,
EDM theodolites, GPS, photogrammetry) has led to an increase in the amount of data collected during fieldwork. As
a result, new insights are being offered into fluvial dynamics utilizing three-dimensional DEMs of the riverine
environment (e.g. Lane et al., 1994; Milne and Sear, 1997; Heritage et al., 1998; Brasington et al., 2000; Fuller et al.,
2005). Many of these studies continue to suffer area or resolution limitations due to a trade-off between spatial
coverage and morphologic detail captured (Figure 1): techniques such as terrestrial photogrammetry produce dense
accurate morphometric data but aerial coverage is restricted; aerial photogrammetry offers increased spatial coverage
but reduced elevation accuracy; EDM theodolite surveys suffer from long collecting times resulting in reduced data
density if large areas are surveyed (Figure 1). Spatial coverage, data point density, data point accuracy and their
relationship with field survey and post-processing time are particularly important when viewed alongside the spatial
and temporal change that is occurring within fluvial systems where a general negative relationship exists between
scale of change and rapidity of change (see Knighton 1998).
Table I illustrates the high accuracy and resolution achieved by flume and field studies of relatively small
areas using ground-based photogrammetric techniques, with large areas being mapped only through construction of

Copyright © 2006 John Wiley & Sons, Ltd. Earth Surf. Process. Landforms (in press)
DOI: 10.1002/esp

2 G. Heritage and D. Hetherington

Figure 1. The spatial and temporal limits of conventional morphometric survey techniques.

Table I. Conventional field survey techniques: extent and accuracy

Environment; location Method Survey area (m2) Overall accuracy point spacing Reference

Glacier-fed, braided; Levelling, oblique 700 <1 m (x,y), <0·01 m (z) / 0·1 m−2 Goff and Ashmore (1994)
Sunwapta River, Canada photography
Flume data Photogrammetry 1 0·002 m/11 100 m−2 Chandler et al. (2000)
Waimakariri River, Photogrammetry 3·3 × 106 0·17 m/ 0·25 m−2 Westaway et al. (2003)
New Zealand
Gravel bar; Kingwater Stream, Photogrammetry 10–100 ± 0·026–0·057 m/4–12 m−2 Heritage et al. (1998)
Cumbria, UK
Braided; River Coquet, Aerial LIDAR 2 000 000 ± 0·15–0·3 m/ 0·065 m−2 Charlton et al. (2003)
Northumberland, UK
Wandering; River Nent, Theodolite EDM 8 000 ± 0·05 m (x,y,z) / 0·075– 0·275 m−2 Chappell et al. (2003)
Cumbria, UK

a mosaic of separate surveys requiring increased field and post-processing time. Theodolite and GPS surveys have
covered larger areas (Chappell et al., 2003; Fuller et al., 2003) achieving average data point resolution of around 1 m−2;
however, such surveys require several field days and can suffer from operator bias and/or DEM interpolation problems
depending on the survey technique adopted. Still larger areas have been the focus of topologic investigation using
laser-based airborne LIDAR (light detection and ranging) approaches (Charlton et al., 2003) and aerial photographic
techniques (Westaway et al., 2003). In both cases data resolution limitations and surface elevation estimations led to a
failure to pick up grain- and microtopographic-scale morphometric data.
Hardware and software developments in the area of oblique field-based laser scanning (LIDAR) now offer a
significant improvement in the speed, accuracy, resolution and aerial coverage of topographic data acquisition. Initial
studies using the new approach have concentrated in the urban environment (Maas and Vosselman, 1999; Yamada et
al., 2003) capturing detail on building facades. More recently, use of the equipment has been extended to the natural
environment. Bitelli et al. (2004) reported on landslide dynamics through repeat slope survey, Pyysalo and Hyyppa
(2001) have quantified vegetation metrics, Hetherington et al. (2005) quantified barform change across a glacial
outwash plain and Nagihara et al. (2004) have demonstrated the utility of the new technique in acquiring a high-
resolution topographic data set of a barchan dune. Details on survey logistics and resultant dataset character (Table II)
indicate that the new approach is potentially both accurate (sub-centimetre to centimetre scale elevation error) and
efficient in terms of field data collection and post-processing.
It is clear that the new approach opens up major new opportunities to investigate morphology and processes in the
fluvial system. Information may be acquired at scales ranging from gravel- (m2 × 10−2) through to reach-scale (m2 × 102)

DOI: 10.1002/esp

Laser scanning in fluvial geomorphology 3

Table II. Summary application data for laser scanning in the natural environment

Scan number/ Overall accuracy /
Environment; study type survey duration Survey area (m2) point spacing Reference

Barchan dune; measurement 11 scans/0·5 days 400 ± 6 mm / 100 m−2 Nagihara et al. (2004)
Glacial outwash plain; daily 4 scans/daily survey, 4 000 ± 2 mm / 2500 m−2 Hetherington et al. (2005)
morphological change measurement 2 h duration
Felled forest; roughness estimation 80 m2 / day 240 ± 4 mm / 50 000 m−2 Schmid et al. (2004)
and volume balance measurement
Valley side; landslide measurement 3 scans 4 000 ± 10–25 mm / 1000 m−2 Bitelli et al. (2004)
Glacial lake/valley; ice cliff evolution 14 scans/1 day 36 000 ± 3–5 mm /1000–1 400 m−2 Conforti et al. (2005)
Cliff face; stratigraphic modelling Not specified 37 500 ± 5 mm / >10 000 m−2 Bellian et al. (2005)
Pine forest; tree canopy structure 3 scans/0·5 day 60 ± 5–15 mm / up to 100 000 m−2 Danson et al. (2006)
measurement

Figure 2. The River Wharfe at Deepdale, North Yorkshire, UK. This figure is available in colour online at www.interscience.wiley.com/
journal/espl

areas (Figure 1) and operator bias and interpolation error are significantly reduced. Despite the apparent ease of use
and the ability to densify morphometric data collection, none of the previously published studies (Table II) has
provided a full evaluation of survey errors.
This paper critically assesses the accuracy of the new technique in representing rapidly changing morphology,
diverse vegetation and the presence of water. Scan location, frequency and distance impact on survey accuracy and
efficiency, as does data cloud meshing and extreme value handling during post-processing. A protocol for survey error
minimization is proposed based on the experiences gained through the acquisition of 0·01 m resolution digital eleva-
tion data of a topographically diverse upland reach of the River Wharfe (Figure 2). Suggestions are also made
concerning the value of the raw and processed data sets in relation to roughness characterization, channel change and
fluvial processes.

Field Site
The study site at Deepdale on the River Wharfe, North Yorkshire, England (OS grid ref: SD890799), comprises a
150 m straight reach flowing over limestone that may be divided into an upstream gravel-cobble bed grading into
cobbles and boulders before flowing over smooth bedrock downstream (Figure 2). In-channel sediment D50 ranges
from 0·06 m to 0·16 m from the gravel to the boulder bed, excluding large boulders and bedrock outcrops. The
channel width is fairly uniform at around 12 ± 1 m and has a natural boundary along its length apart from a short
stretch of masonry wall protecting the right bank in the bedrock section. Gauged flow data for the river at the study
site are not readily available. However, a rated section of the bedrock sub-reach indicates that ‘bankful flow’ occurs at
around 25 m3 s−1. Surface flow in the channel may cease at the study site during drier periods in the summer due to

DOI: 10.1002/esp


water entering sinkholes upstream. This leaves only a few small pools of water in the bedrock section exposing much
of the bed for accurate laser survey. No submerged aquatic vegetation exists at the site; upstream the banks are lined
by mature alder (Alnus glutinosa), grading into patches of burdock (Arctium minus), common stinging nettle (Urtica
dioica) and grass species on the banks and across the confined right-bank floodplain.

Methodology
The channel and the surrounding confined floodplain were surveyed using a Riegl LMS Z210 terrestrial laser scanner
over a period of 2 days in August 2003. The technical specifications of the scanner may be found in Hetherington
et al. (2005).
A series of 21 high resolution scans were conducted from locations in the centre of the channel, upstream and along
the right and left bank using the field laser attached to a toughbook computer. Survey control was facilitated by the
RiScan-Pro survey software package, capable of visualizing point cloud data in the field. Scans were generally
restricted to 180° in front of the scanner and were collected with substantial overlap up- and downstream and across
the channel, ensuring that the surface of the study reach was recorded from several directions. This increases the point
resolution across the surface and reduces the possibility of unscanned ‘shadowed’ areas due to large roughness
elements shielding objects in their lee from the laser scanner pulses. At each scan location the instrument was mounted
as high as possible above the surface being scanned, allowing greater laser pulse penetration of topographic lows,
further reducing the problem of large object shadowing. Individual scans contained around 1 million points which
were stored in files approximately 10 Mb in size, recording angular information, range data colour and reflectivity.
Individual scan times, when conducted efficiently, took less than half an hour including setup.
Data from individual scans are recorded in the internal scanner co-ordinate system which varies with each setup.
These were merged to a single project co-ordinate system using 60 independent EDM theodolite points on prominent
features such as corners of walls, gates and road signs, angular rocks and fence post tops. To increase accuracy the
independent field EDM theodolite points were measured with a prism reflector without the pogo attachment. The
locations of these tie points were manually identified through visual inspection of each scan using the RiScan-Pro
software and subsequently used in the merging process only where they could be fixed with a precision of ±0·01 m.
Scan cloud point alignment was achieved using between 4 and 13 tiepoints depending on scanner location. Mean
tiepoint alignment error was 0·017 m, with a mean standard deviation of 0·026 m and a range between 0·0017 and
0·055 m.
The resultant meshed set of laser scans was manually clipped within the RiScan-Pro software to remove unwanted
information such as distant points, overhanging tree canopy and any spurious aerial data points returned from aerosols
or water droplets. Further refinement of the individual scan data points was then conducted using the POLYWORKS
post-processing software package. Deviations in individual data point locations between scan clouds were minimized
to improve overall point cloud alignment. Finally, data alignment to the project co-ordinate system was checked using
119 independent EDM theodolite points.

Results
The final data set of some 17 million points covers a channel and overbank area of roughly 150 m × 15 m surveyed at
an average 1 cm resolution (Figure 3). Such a data set is able to capture detail of the riverbed gravel, bedrock surface
and vegetation character and, as far as the authors are aware, represents the most detailed reach-scale topographic
survey of a river channel produced to date. The co-ordinate data files may be exported as ASCII format for use with
other software. However, files in this format are large, some 750 Mb in this case.
The accuracy of the laser scan survey points was evaluated through the collection of 257 independent surface point
co-ordinates using an EDM theodolite (Table III). These data were obtained at the same time as the laser scan survey
and taken in a downstream direction to avoid disruption of the post-scan surface. Direct comparison of the EDM
theodolite data and laser survey points was achieved through the extraction from the laser database of all laser points
within a 0·02 m radius of the theodolite point. The radius accounts for potential errors in the location of the EDM
theodolite reflector in space.
The overall accuracy of the laser data was assessed against EDM theodolite data collected from exposed cobble
surfaces that would have been unaffected by any survey ‘shadowing’ problems during data collection. Residual errors
between the theodolite and laser elevations were computed (Table III), revealing a mean error of only 0·004 m.
Examination of the error distribution shows a generally good agreement between the laser data surface and the

DOI: 10.1002/esp


Figure 3. Final laser scan model of the River Wharfe at Deepdale, North Yorkshire, UK.

Table III. Laser scan discrepancy statistics defining the agreement between scanner data and EDM theodolite data across differing
surfaces on the River Wharfe study reach at Deepdale

Broad-leaved
Alignment Bedrock Rock gaps vegetation Grass Water Water edge

Mean 0·0038 − 0·0065 0·2547 0·0749 0·0686 − 0·2555 − 0·2359
Varience 0·0279 0·0211 0·0171 0·0168 0·0077 0·0195 0·0066
Standard deviation 0·1673 0·1454 0·1308 0·1296 0·0881 0·1398 0·0815
Point number 157 93 44 33 19 47 23

independent theodolite points: 55 per cent of the data are accurate to ±0·02 m. However, approximately 35 per cent
displayed positive and negative error residuals between 0·02 m and 0·1 m, and 10 per cent between 0·1 m and 0·2 m.
The errors were further analysed with respect to surface colour (Figure 4a) and intensity of laser pulse reflection
(Figure 4a). In all cases no obvious trend could be detected in the plots. A single co-ordinated scan was chosen to
determine the effect of scan distance on residual discrepancies (Figure 4b); again, no trend was apparent in the data.
The terrain across the study reach was variable including bedrock, boulders, cobbles, gravel, vegetation and water.
The form of the laser scan return data was assessed across each of these surface types (Table III, Figure 5). Across
smooth bedrock surfaces visible from many scan locations, laser scan accuracy was as good as for the exposed cobble
surfaces used as scan tie points, with a mean error of 0·0065 m (Figure 5a). The data are slightly negatively skewed
indicating that the laser-based surface is below the theodolite survey points; however, the error is minimal. Larger
errors of up to ±0·3 m are occasionally recorded.
Laser data captured in the gaps between cobbles and boulders were less accurate (Figure 5b) displaying a consist-
ently positive discrepancy relative to the theodolite points with mean errors of 0·025 m. This error is likely to have
been caused by the local shadowing effect of adjacent clasts preventing penetration of the pulsed laser light into the
interstitial spaces, leading to an overestimate of the true surface. Such errors are greatest where the surrounding
material is large and the gaps narrow and where the angle of incidence of the laser pulse is low. The range of errors is
also high (0·05 to 0·45 m) reflecting the heterogeneous nature of the cobble and boulder bed. Vegetation also leads to
apparent discrepancy in the determination of the study surface; the effects have been investigated separately for
broad-leaved vegetation and grasses. The average discrepancy in identifying the ground surface in areas covered in
broad-leaved vegetation (principally burdock, A. minus) is apparently quite low (0·07 m) with the laser survey
overestimating the true height due to the laser pulse being reflected off the leaf surface before reaching the ground

DOI: 10.1002/esp


Figure 4. Effect of (a) colour ( , red; , green; , blue; , return intensity) and (b) distance on laser scan accuracy.

Figure 5. Effect of different terrain types on laser scan DEM discrepancy for the River Wharfe at Deepdale: (a) bedrock, (b) rock
gaps, (c) broad-leaved vegetation, (d) grass, (e) water, (f ) water’s edge.

DOI: 10.1002/esp


(Table III). Conventional EDM theodolite survey ignores the vegetation. The discrepancy appears bimodal (Figure 5c)
with the positive peak of 0·2 m reflecting the average leaf height and returns where the laser pulse penetrated to the
ground surface. Occasional underestimates of surface height are more difficult to explain and require further study.
Grasses show a different response, with more variable results around a mean discrepancy of 0·07 m (Figure 5d). It is
suggested that this is due to the lower leaf density and more diffuse leaf structure of the grasses allowing some of the
laser pulses to reach the ground, thus returning a co-ordinate value equivalent to that recorded by theodolite survey.
The general spread of data points across this vegetation type probably results from pulses penetrating the vegetation to
varying degrees before being reflected off grass stems or the ground surface. The effect of water on scan data accuracy
is apparent from Figure 5e. Often the laser pulse is refracted off the water surface and no return is recorded. However,
where the water is clear, calm and shallow and the angle of incidence of the laser pulse is high, some penetration was
measured. In the case of the Wharfe pools the water was clear and calm with an average depth of around 0·1 m; the
error in elevation returns from the laser survey was of the order of 0·23 m. Further work is required to determine
whether a correction factor may be applied to these data to provide more accurate elevation data in this environment.
A similar effect is noticeable from water edge data (Figure 5f ) that also have mean errors of 0·25 m, comparable to
results for still areas of shallow pools. The inaccuracy here raises questions over the use of laser data in defining water
boundaries and in deriving accurate water surface slope information. The authors’ experience with field scanning on
the Arolla and Ferpecle glacial outwash plain suggests data are also returned when the water is turbid and where the
water surface is disrupted due to high bed roughness, although the accuracy is highly variable.

Discussion
The use of oblique laser scanners to generate detailed accurate DEMs of landforms represents a major improvement
on previous survey methods, both in aerial coverage and DEM accuracy (Figure 1). Data may now be rapidly acquired
and processed at a density sufficient to represent the surface at the grain scale estimated at between 4000 and
10 000 points m−2 by Lane et al. (1994). The issue of point distribution and potential operator bias (Lane et al., 2003)
is also rendered obsolete as a dense cloud of meshed data points ensures that a surface is sampled many times. Data
point quality may still potentially prove to be an issue for some studies aimed at the grain scale as the range error on
current instruments may lead to unacceptable inaccuracies in the DEM surface. Studies at the sub-bar scale, identified
as a significantly under-researched area by Charlton et al. (2003), would suffer much less from such small errors and
the rapidity with which the laser data may be acquired enables change at this scale to be monitored in detail.
Despite the obvious advantages of terrestrial laser scanning, great care needs to be exercised during data collection
and processing in order to ensure that data accuracy is maximized. In the field, survey positions should be chosen that
maximize scan angle across a surface, thus reducing topographically induced data shadows (see Figure 5b) and
considerable scan overlap should occur to capture more fully the three-dimensional structure of the scanned surface.
Care must be taken to select clearly recognizable tiepoint locations if manual data cloud meshing is attempted,
ensuring that at least four are visible from each scan position to facilitate accurate meshing. A surveyed network of
retro-reflectors placed across the study area may also be utilized, allowing automated tiepoint selection from each scan
based on the reflectivity values returned for each data point (see Hetherington et al., 2005).
The ability of the laser scanner to detect solid exposed surfaces such as bedrock (Figure 5a) has been demonstrated,
generally achieving centimetre-scale accuracy comparable with other studies (Table II) and exceeding accuracies
previously reported for large-scale topographic surveys using conventional techniques (Table I). Caution must be
exercised, however, as a small proportion (c. 10 per cent) of the data may be in error by more than 10 cm. It is
suggested that multiple scanning from the same scan position followed by scan cloud averaging would minimize these
errors, although this has not been demonstrated in this study which meshed a series of single scans.
The effect of vegetation obscuring the ground surface has been demonstrated (Figure 5c and d); similar effects have
also been reported in airborne LIDAR (Charlton et al., 2003) and photogrammetric studies. The data from terrestrial
laser scanning must be treated with caution when trying to model the ground surface. However, this study has noted
the ability of some terrestrial laser scanner pulses to penetrate the vegetation and return a signal from the ground
surface, as evidenced by the near-zero errors when compared with the EDM theodolite data (Figure 5d). An algorithm
to select the basal data from the scan returns would thus remove the vegetation effect.
The presence of water also poses significant problems for the interpretation of the laser data (Figure 5e) with partial
penetration occurring in clear shallow water and high laser pulse incidence angles and beam refraction off the water
surface at reduced incidence angles. Irish and Lillycrop (1999) note that the use of multispectral laser data may
improve accuracy in clear water environments. Atmospheric moisture may also be important: the authors have noted
occasional spurious data points above the survey landform surface following laser scanner operation in rain and fog.

DOI: 10.1002/esp


Conclusions
This paper reports on a new, rapid, detailed topographic survey technique for the natural environment, which has
potentially wide application across a variety of disciplines. However, the accuracy of the data obtained is dependent
on field operation of the equipment, terrain character and instrument specifications. Optimal results will be gained by
observing the following field protocol:

(a) minimize the scan distance to ensure greater scan point density;
(b) tilt the scanner towards the river channel to maximize the amount of data collected locally;
(c) select scan locations to minimize scan shadow effects caused by large obstructions;
(d) where possible, optimize the scan angle by setting the instrument well above the scanned surface;
(e) collect independent tiepoint/error check data to minimize systematic bias introduced during scan cloud merging;
(f ) use manually selected tiepoints for more accurate scan merging due to the ability to select their location in the
scan data with high precision;
(g) where centimetre-scale accuracy is not required utilize a reflector system to exploit reflector auto-detection algo-
rithms in the software and reduce post-processing time;
(h) ensure that some reflectors/tiepoints are placed at the edges of the scanned area to minimize propagation of
meshing errors;
(i) ensure a good variation in x, y and z dimensions when selecting tiepoints/reflector locations; this improves scan
merging accuracy and reduces the possibility of ‘chance’ scan merging due to similar distances and elevations
between tiepoints or reflectors;
(j) repeat scans from the same location to densify the data collected and potentially reduce extreme errors;
(k) avoid low angle scans across water surfaces.

Whilst laser scanning has impacted significantly on areas of engineering, it remains under-exploited as a technique for
studying the character and dynamics of the natural environment. The relatively small errors reported in this paper are
likely to be improved upon as the protocol for surveying the natural environment is improved and instruments
improve in terms of their technical specifications opening up research opportunities in the fields of facies analysis,
grain-size characterization, surface roughness evaluation and channel hydraulics. Aerial laser data have advanced
vegetation characterization and this may be complemented by terrestrial data impacting on floodplain characterization
and biotope definition. System change may be efficiently and accurately measured, advancing morphologic dynamics
and sediment budgeting studies and potentially improving CFD modelling through more detailed model meshes.

References
Bellian JA, Kerans C, Jennette DC. 2005. Digital outdrop models: applications of terrestrial scanning LiDar technology in stratigraphic
modeling. Journal of Sedimentary Research 75(2): 166–167.
Bitelli G, Dubbini M, Zanutta A. 2004. Terrestrial laser scanning and digital photogrammetry techniques to monitor landslide bodies. In
Proceedings of the XXth ISPRS Congress, Istambul: 6.
Brasington J, Rumsby BT, McVey RA. 2000. Monitoring and modelling morphological change in a braided gravel-bed river using high-
resolution GPS-based survey. Earth Surface Processes and Landforms 25: 973–990.
Chandler JH, Lane SN, Ashmore P. 2000. Measuring river-bed and flume morphology and paramterising fed roughness with a Kodak
DCS460 digital camera. In International Archives of Photogrammetry and Remote Sensing, XXXIIIB4, Beek KJ, Molenaar M (eds).
GITC 6v, XIXth Congress: Amsterdam; 250–257.
Chappell A, Heritage GL, Fuller I, Large ARG, Milan D. 2003. Spatial and temporal river channel change: geostatistical analysis of ground-
survey elevation data. Earth Surface Processes and Landforms 28: 349–370.
Charlton ME, Large ARG, Fuller IC. 2003. Application of airborne LiDAR in river environments: the River Coquet, Northumberland, UK.
Earth Surface Processes & Landforms 28: 299–306.
Conforti C, Deline P, Mortara G, Tamburini A. 2005. Terrestrial Scanning Lidar Technology applied to study the evolution of the ice-contact
image lake (Mont Blanc, Italy). In Proceedings of the 9th Alpine Glaciological Meeting, February, Milan, Italy.
Danson FM, Hetherington D, Koetz B, Morsdorf F, Allgöwer B. 2006. Three-dimensional forest canopy structure from terrestrial laser
scanning. In Proceedings of 3D Remote Sensing in Forestry, Vienna (in press).
Fuller IC, Large ARG, Charlton ME, Heritage GL, Milan DJ. 2003. Reach-scale sediment transfers: an evaluation of two morphological
Budgeting approaches. Earth Surface Processes and Landforms 28: 889–904.
Fuller IC, Large ARG, Heritage GL, Milan DJ, Charlton ME. 2005. Derivation of reach-scale sediment transfers in the River Coquet,
Northumberland, UK. In Blum M, Marriott S, Leclair S (eds). Fluvial Sedimentology VII. Special Publication 35. International Associ-
ation of Sedimentologists: 61–74.

DOI: 10.1002/esp


Goff JR, Ashmore PE. 1994. Gravel transport and morphological change in braided Sunwapta River, Alberta, Canada. Earth Surface
Processes and Landforms 19: 195–212.
Heritage GL, Fuller IC, Charlton ME, Brewer PA, Passmore DP. 1998. CDW Photogrammetry of low relief ﬂuvial features: Accuracy and
implications for reach scale sediment budgeting. Earth Surface Processes and Landforms 23: 1219–1233.
Hetherington D, Heritage GL, Milan DJ. 2005. Reach scale sub-bar dynamics elucidated through oblique lidar survey. In International
Association of Hydrological Scientists Red Book Publication. IAHS Publication 291: 278–284.
Irish JL, Lillycrop WJ. 1999. ‘Scanning laser mapping of the coastal zone: the SHOALS System’. ISPRS Journal of Photogrammetry &
Remote Sensing 54(2–3): 123–129.
Knighton D. 1998. Fluvial Forms and Processes. Arnold: London.
Lane SN, Chandler JH. 2003. The generation of high quality topographic data for hydrology and geomorphology: new data sources, new
applications and new problems. Earth Surface Processes and Landforms 28.
Lane SN, Chandler JH, Richards KS. 1994. Developments in monitoring and terrain modelling small-scale river-bed topography. Earth
Surface Processes and Landforms 19: 349–368.
Lane SN, Westaway RM, Hicks DM. 2003. Estimation of erosion and deposition volumes in a large gravel-bed, braided river using synoptic
remote sensing. Earth Surface Processes and Landforms 28: 249–271.
Maas HG, Vosselman G. 1999. Two algorithms for extracting building models from raw laser altimetry data. ISPRS Journal of Photogrammetry
and Remote Sensing 54(2–3): 153–163.
Milne JA, Sear DA. 1997. Surface modelling of river channel topography using GIS. International Journal of Geographic Information
Science 11(5): 499–519.
Nagihara S, Mulligan KR, Xiong W. 2004. Use of a three-dimensional laser scanner to digitally capture the topography of sand dunes in
high spatial resolution. Earth Surface Processes and Landforms 29: 391–398.
Pyysalo U, Hyyppa H. 2001. Reconstructing tree crowns from laser scanner data for feature extraction. In Automated Interpretation of High
Spatial Resolution Digital Imagery for Forestry Institute of Photogrammetry and Remote Sensing. Helsinki University of Technology.
Schmid T, Schack-Kirchner H, Hildebrand E. 2004. A case study of terrestrial laser-scanning in erosion research: calculation of roughness
indices and volume balance at a logged forest site. International Archives of Photogrammetry, Remote Sensing and Spatial Information
Sciences 36(8/ W2): 114 –118.
Westaway RM, Lane SN, Hicks DM. 2003. Remote survey of large-scale braided rivers using digital photogrammetry and image analysis.
International Journal of Remote Sensing 24: 795–816.
Yamada O, Takase Y, Shimoda I, Nakagawa T. 2003. Signiﬁcance of digital reconstruction of historic buildings using 3d laser scanner case
study: prasat suor prat n1 tower, Angkor, Cambodia. International Archives of Photogrammetry, Remote Sensing and Spatial Information
Sciences Xxxiv(5/w12): 342–346.

DOI: 10.1002/esp

Heritage hetherington lidar_pdf[1]

Recomendados

Recomendados

Más contenido relacionado

La actualidad más candente

La actualidad más candente (20)

Destacado

Destacado (6)

Similar a Heritage hetherington lidar_pdf[1]

Similar a Heritage hetherington lidar_pdf[1] (20)

Más de llica

Más de llica (20)

Último

Último (20)

Heritage hetherington lidar_pdf[1]