 |
| Titel |
Comparative analysis of landslide kinematics using ground-based optical images and terrestrial laser scanning |
| VerfasserIn |
Sabrina Rothmund, Jean-Philippe Malet, Manfred Joswig |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250051077
|
|
|
|
|
|
| Zusammenfassung |
| High-resolution multi-temporal digital surface models (DSMs) acquired by ground-based
optical images and terrestrial laser scanner (TLS) are used in combination to estimate the
three-dimensional surface deformation and displacements of the Super-Sauze landslide
(South French Alps) during a period of two months (June-July 2010). The analysis is part of a
field experiment with the objective to characterize the relationship of small fracture processes
with slope movement, fissure developments and pore water pressure build-up. Long-term
observations of the average displacement rate at the investigated area (about 10.000 m2) are
greater than 0.01 m.day-1.
During the monitoring period, 3D point clouds were acquired with an Optech ILRIS-3D TLS
system at different dates. On the same days, several optical images (2816 x 2112
pixels) were taken by a low-cost digital camera to test the validity of advanced
photogrammetric algorithms for precise data acquisitions. Additionally, DGPS
measurements were carried out in order to geo-reference and validate the point cloud
data.
The investigated area with an elevation difference of about 50 m is characterized by large
relief irregularities (flattened areas, blocks, small gullies). These irregularities lead to
shadowed zones and finally to gaps in the acquired point clouds. Nine laser scans from four
different viewpoints with an average point spacing between 10 to 30 mm were acquired in
order to minimize these undesired gaps. In a final processing step all TLS-point clouds were
merged and geo-referenced.
About 70 optical images were taken on points along a predefined line on the side part of the
landslide. To maximize the coverage, a point to point distance of 10 m and different viewing
directions were chosen. The distance between camera and relevant area ranges
from 5 to 60 m. The photogrammetric point cloud was then generated from all
images by applying a new feature-based multi-view stereo approach which does
not require any ground control point information. To determine the quality of the
photogrammetric DSMs, the point clouds were compared with the LiDAR-based
DSMs.
The surface deformation over a period of eight weeks was determined by comparison of the
photogrammetric DSM, the TLS-based DSMs as well as the 3D-displacement vectors
calculated from the DGPS coordinates. |
|
|
|
|
|
|