In: Wagner W., Szekely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part7B
MONITORING OF BEACHES AND SAND DUNES USING DIGITAL AERIAL
PHOTOGRAPHY WITH DIRECT GEOREFERENCING
J. A. Gon9alves a,b *, L. Bastos a,b , Perez, B. a , Magalhaes, A. a
a University of Porto - Science Faculty, Rua Campo Alegre, 4169-007, Porto, Portugal
b CIIMAR - Centre for Marine and Environmental Research, Rua dos Bragas, 4050-123, Porto, Portugal
j agoncal@fc. up .pt
Commission VII
KEY WORDS: Photogrammetry, Change Detection, Matching, DEM/DTM, Georeferencing, Direct
ABSTRACT:
This paper presents an efficient methodology for coastal monitoring based on digital aerial photography acquired periodically, using
a Zeiss-Intergraph DMC camera, with a spatial resolution of 10 cm. Images are delivered with exterior orientation given by the
direct georeferencing system used by the company providing the images. An assessment of those exterior orientation parameters was
done using field check points. It could be concluded that the positional accuracy of 3D coordinates determined by photogrammetric
means using the exterior orientation elements has root mean square errors better than 30 cm, which can be accepted for the purposes
of this type of work. This level of accuracy was achieved for four different flights. It was decided not to do an aerial triangulation.
DSMs were obtained by stereo-matching (least-squares with region growing) using program DPCOR of the software package
BLUH. Sand areas are normally very bright and difficult for the matching. Since images were acquired with relatively small sun
elevation the automatic extraction of a DSM was very successful. Using check points the vertical accuracy of the DSMs was found
to be 30 cm or better. Orthoimages were created using the extracted DSM. The resulting datasets for the four different dates were
integrated in a GIS and the method proved to be a very useful tool for the detection of areas where significant volumetric changes
occurs and for the quantification of those beach changes.
1. INTRODUCTION
The western part of the Portuguese coast is subject to
significant degradation due to the action of the ocean (Ferreira,
et al., 1995). Coastal areas are also under a strong anthropic
pressure, which have as consequence the destruction of dunes
and other natural protections. Significant morphodynamic
changes occur in short periods of time which need to be
monitored. In order to detect and quantify those changes and
foresee the near term evolution of the coastal areas, both for
safety and economic reasons, very frequent surveys, are
required.
Field survey monitoring programmes are carried, using GPS, in
sensible coastal areas in Portugal, in order to detect and
quantify changes (Baptista et al., 2008, Gonpalves et al., 2009).
However, they are still time consuming and usually applied in
small areas if surface models are to be obtained.
Airborne surveying techniques are more cost effective and can
achieve a positional accuracy appropriate for coastal evolution
assessment. Elevation changes in beach landforms, due to
natural or antropic origin, can be of several meters, even in
short periods. Elevation errors of a few decimetres in digital
surface models (DSMs) are therefore still acceptable for the
detection of essential relief changes in coastal areas.
Laser scanning is nowadays a commonly used technique for
coastal monitoring. It has an important advantage in that the
final DSM/DTM is obtained essentially in an automatic manner
(Baltsavias, 1999). However, besides being still quite
expensive, laser scanning alone does not provide the image
component, which is also very important for the interpretation
and detection of the morphodynamic changes.
Aerial photography, especially with the recent digital aerial
cameras, can provide important information to the assessment
of changes while allowing a high degree of automation. That is
the case if direct georeferencing (DG) can be accurate enough
to avoid aerial triangulation and if the DSM generation can be
fully automated by stereo matching.
Images of sandy areas may be difficult to treat in terms of
standard photogrammetric procedures. For example, for the
purpose of aerial triangulation, in the absence of man made
features, well defined ground control points (GCPs) may be
difficult to obtain, especially in the case of high resolution
images. When large parts of the images are occupied by water
also tie points cannot be obtained. These are additional reasons
to use the exterior orientation data given by DG, if it is accurate
enough.
This paper describes the exploitation of aerial photography
acquired in the scope of a coastal monitoring program in the
Portuguese northwest coast through the referred methodology .
Part of the area does not have man made features to be used as
GCPs. There is usually a need for fast data production
immediately after image acquisition, so there is a strong interest
in relying on DG. One of the objectives of this study is to
determine the best accuracy achievable using image data
provided by a private company.
228