Content uploaded by Olivier Cohen
Author content
All content in this area was uploaded by Olivier Cohen on Jun 29, 2018
Content may be subject to copyright.
1
Mapping coastal dunes morphology and habitats evolution
using UAV and ultra-high spatial resolution photogrammetry
Olivier COHEN 1, Adrien CARTIER 2, Marie-Hélène RUZ 1
1 Laboratoire d’Océanologie et de Géosciences (UMR 8187)
2 Géodunes, Drone Littoral
LIFE+Flander Program
International workshop “Management of coastal dunes and sandy beaches”
Dunkirk, 12-14 june 2018
2
Introduction Method Results Conclusion 2
•Medium and long-term measurements
•Short-term measurements in the field
3
Introduction Method Results Conclusion 3
•Medium and long-term measurements
•Short-term measurements in the field
Elevation profile
4
Introduction Method Results Conclusion 4
•Medium and long-term measurements
•Short-term measurements in the field
Elevation profile
Digital Elevation Model
5
Introduction Method Results Conclusion 5
•Short-term telemetry using LiDAR
•Short-term telemetry by airborne photogrammetry
DEM calculated with LiDAR data
DSM calculated with photogrammetry data
Orthophotographie 2015
6
Introduction Method Results Conclusion 6
•Short-term telemetry by airborne photogrammetry: a complement to LiDAR surveys for smaller areas.
7
Introduction Method Results Conclusion
• Photogrammetry is a technique of restitution of the 3rd dimension (elevation) starting from photographs
of a site taken under various angles. It has been known since the middle of the 19th century, but could only
be implemented by a few providers (ie : French National Geographical Institute).
But until now, it was an expensive and complex technique to set up, which required overflights by
plane or helicopter.
The processing of the aerial photographs did not allow sufficiently precise results to be obtained to
highlight small variations in elevation.
•Today, thanks to the reduction in the costs of flying machines (UAVs) and the development of software
that is affordable and relatively easy to use, digital photogrammetry is within the reach of almost
everyone. It is rapidly successful in many disciplines: archaeology, architecture, geography, earth sciences
(geomorphology, geology).
7
Photogramm Zeiss (source : Université Laval)
8
Introduction Method Results Conclusion
• Photogrammetry uses the principle of stereoscopy: to perceive an object in relief, our eyes each see it
from a slightly different angle because they are a few centimeters apart (6 to 7 in general), it is the parallax.
The brain then reconstructs the impression of depth. So when you close one eye, you have trouble
estimating that depth of field.
8
9
Introduction Method Results Conclusion 9
Quadcopter UAV DJI Phantom IV Pro Powerful workstation
Optimized flight plan Ground control points
10
Introduction Method Results Conclusion 10
• A photorealistic rendering of the 3D model: example in Zuydcoote
11
Introduction Method Results Conclusion 11
• Ultra-high spatial resolution orthophotography (pixel = 1 cm)
12
Introduction Method Results Conclusion 12
• Ultra-high spatial resolution orthophotography (pixel = 1 cm)
13
Introduction Method Results Conclusion 13
• Ultra-high spatial resolution Digital Surface Model (pixel = 5 cm): many details can be distinguished.
Upper beach accumulations
Mid-beach bar
Fences alignment
Access to the beach
03 November 2017
Blowout
14
Introduction Method Results Conclusion 14
•Comparison of the DSM of 03/11/2017 and 15/05/2018: the mid-beach bar, the upper beach and the
dune slope have been eroded (red, orange), the blowouts are filling up (blue areas on the foredune).
•Calculation of sediment budget.
Evolution
Erosion = 6410 m3, -0.25m3/m2
Accumulation = 1368 m3, 0.22m3/m2
Sediment budget = -5041 m3
Surface = 3,2 ha
Budget/surface = -0,15 m3/m2
Evolution
15
Introduction Method Results Conclusion 15
• Comparison of topographic profiles: from November 2017 to May 2018, the dune foot remained at the
same position but the dune was scarped. The upper beach was lowered, the upper bar disapeared.
16
Introduction Method Results Conclusion 16
• Semi-automatic classification of 3D model points for mapping:
Areas of bare sand
Vegetation zones
17
Introduction Method Results Conclusion 17
Advantages Disadvantages
An affordable cost
Increase of productivity in the field :
oa technique that is quick and easy to set
up
olarger areas measured faster than by GNSS
A high density of measurements
(several
hundred per m2 depending on flight
height,
images resolution and available
computing
power) much higher than that of
a
topographic LiDAR (1 to 2 points/m2).
A high level of accuracy (error margin ± 3 to
5
cm)
A 3D study of the shoreline
3D models showing invisible elements
with
GNSS surveys (e.g. sand/vegetation
boundary,
vegetation type, wrach line, etc.)
Useful 3D models for
scientific
communication
Sensitivity to weather conditions (wind, rain)
Flight time and limited range (battery life:
20
to 25 minutes)
Long processing times, high computing
power
required (workstation)
Need ground control points to obtain a
very
precise geo-referencing of the model
Vegetation (trees) must sometimes
be
removed from models
Administrative regulations sometimes
binding
for UAV overflights
18
Thank you for your attention.
Olivier COHEN 1, Adrien CARTIER 2, Marie-Hélène RUZ 1
1 Laboratoire d’Océanologie et de Géosciences (UMR 8187)
2 Géodunes, Drone Littoral
LIFE+Flander Program
International workshop “Management of coastal dunes and sandy beaches”
Dunkirk, 12-14 june 2018
19
Coastal dunes evolution can be analysed over medium time scale (few decades) using high resolution aerial photographs which are the best
source of information about mesoscale changes. In the field measurements carried out with surveying instruments (e.g. total electronic stations
or high precision GNSS) complement these observations in the shorter term and in 3D. However their achievement is time-consuming and
difficult to carry out over large areas. It is therefore necessary to measure topographic profiles or Digital Elevation Models on restricted size
zones (a few thousand square meters in general).
In recent years, Light Detection And Ranging (LiDAR) systems deployed on aircraft have proven to give accurate horizontal and vertical
measurements and high spatial resolution 3D data (1 or 2 measurements per m2) on broader areas (several km2 to tens of km2). This is the case
along the Hauts de France coastline where LiDAR surveys have been carried out every 3 to 4 years since 2008. However, LiDAR surveys are
usually not sufficiently frequent to accurately characterize short-term changes, such as storm-related ones.
Unmanned Aerial Vehicles (UAV) photogrammetry is a convenient technique to complement LiDAR surveys. It is experiencing growing success in
the field of Earth sciences. It provides data at ultra-high spatial density and high precision (several hundred measurements per m2, margin of
error from 1 to 5 cm in plane and altitude). Photogrammetric images processing allows to compute Digital Surface Models to precisely describe
beach and dunes morphology. For example on these DSM, it is possible to clearly distinguish embryo-dunes on the upper beach that cannot be
mapped with LiDAR data. Comparison of several DSM helps to determine evolutions in the short term, e.g. migration and development of
embryo-dunes, erosion or accretion of the foredune.
Photogrammetric processing also provides ultra-high resolution orthophotos (1 cm2/ pixel) on which fine details (wrack line) can be identified.
Vegetation cover can be mapped using automatic classification of the aerial photographs. Comparison of several classification maps provides
information on the habitats evolution.
This communication presents the UAV photogrammetric survey carried out by members of the Oceanology and Geosciences Laboratory (UMR
8187) in the East of Dunkirk.
Keywords: coastal survey, UAV, photogrammetry, geomatics, dunes morphodynamics, vegetation cover evolution, habitats
LIFE+Flander Program
International workshop “Management of coastal dunes and sandy beaches”
Dunkirk, 12-14 june 2018
