PE&RS July 2016 Public - page 556

12:50, one hour and 22 minutes before the rapid increase. The
dual axis inclinometer seems less sensitive and measured a
much lower level change very close to the collapse. After the
initial small increase of the pore water pressure for about first
18 minutes measured by all five piezometers, two of them
increased and three decreased in a small amount and approxi-
mately leveled until a significant increase of all five in an av-
eraged period of 47 minutes from about 13:53 to 14:40 (Plate
5b), Three piezometer measurements continued to increase
after the collapse, which did not happen to other contact
sensors and the imaging sensors. The osmometer showed the
pressure measured in soil due to saturation that is a similar
curve as the piezometers, but with the significant increase at
about 13:21, 32 minutes earlier.
The imaging sensors detected changes on the large surface
of the slope, while the contact sensors measured changes in
different depths at various points. The initial comparative
analysis showed some degree of cohesiveness among the
observations. During the period (12:40 ~ 14:27) of the calcu-
lated sliding blocks area (Figure 2) the surface soil was soaked
at different levels and built sliding blocks, which expanded
the areas as the precipitation continued. Gradually the energy
was accumulated and the soil saturation level measured by
the osmometer started to rise rapidly at about 13:21 when the
first local collapse, Event 1 in Figure 3, was detected by the
camera. As both the saturation level and pore water pres-
sure continued to increase (Plate 5b), the water can penetrate
to the low layer and lead to the development of positive in-
terstitial pressures. Thus three more local collapse events and
the final failure occurred as shown in Figure 3.
In this study, an image-based approach for measuring high-
speed surface changes during a simulated landslide using a
scaled-down slope model was designed and tested. The results
and analysis demonstrated a non-contact photogrammetric
methodology for monitoring surface deformation in terms of
event detection and parameter estimation during the entire ex-
periment and during the short period of intense slope failure.
1. The developed system demonstrated an improved
capability of monitoring relatively slow landslide
deformation events and studying rapid surface changes
during the final slope failure by a combination of low-
and high-frequency imaging and associated feature
detection and parameter estimation methods. The
achieved position accuracy is 3.8 mm (one sigma) and
the speed accuracy is 0.11 m/s.
2. The proposed surface change descriptors, includ-
ing percentage of sliding block areas, percentage of
changed features and average speeds, can be used to
describe the characteristics of overall and local defor-
mations and to analyze behavior of the landslide slope
related to energy build-up or release, block sliding, lo-
cal collapses, and slope failure. The developed method
successfully detected four pre-failure local collapse
events and the final slope failure with a peak average
speed of 0.8 m/s.
3. The local deformation behavior and spatio-temporal
speed change patterns derived in different sections of
the slope model allowed a detailed systematic analysis
of sectional deformation changes as well as the interac-
tions between the materials moved from one section to
the other. Especially the developed imaging and change
detection capability for the last five seconds of the
slope failure process is valuable for a slow play back
and quantitative study of the failure mechanism that
would be otherwise impossible.
4. The initial analysis demonstrated a collaborative
capability between the imaging sensor based surface
change detection and the contact sensor based point-
wise change detection. The addition of the image based
technology provided a correspondence between the sur-
face deformation events and the development of posi-
tive interstitial pressures in the subsurface. This time
period of 66 minutes before the failure is confirmed by
significant signals from both imaging and contact sen-
sors and is important for landslide early warning.
Our future research will include the further in-depth col-
laborative analysis using both photogrammetric processing
results and observations of in situ contact sensors installed on
the surface and in the subsurface of the slope. To have an insight
of the subsurface activities, we will add some colored layers of
materials on both sides of the slope. Additional images through
the glass walls will allow the measurements and analysis of
deformation and motion of the mass both on the surface and be-
tween layers. Considering the deployment of the system and the
environment limit in the field, for instance imaging under a par-
tially covered canopy or at night, additional lighting or infrared
cameras can be introduced. A field implementation scheme in
western China will be designed and realized in the near future.
We appreciate the constructive comments of three reviewers. This
study was supported by the National Key Basic Research Pro-
gram of China (2012CB719903, 2012CB957701, 2013CB733203,
2013CB733204), the National Science Foundation of China
(41171327, 41201425, 41201379), Specialized Research Fund for
the Doctoral Program of Higher Education (20120072120057), the
China Special Fund for Surveying, Mapping and Geoinformation
Research in the Public Interest (201412017), and the Fundamen-
tal Research Funds for the Central Universities.
Abellán, A., T. Oppikofer, M. Jaboyedoff, N.J. Rosser, M. Lim, and M.J.
Lato, 2014. Terrestrial laser scanning of rock slope instabilities,
Earth Surface Processes and Landforms
, 39(1):80–97. doi:
Arbanas, Ž., T.F. Fathani, Z. Shoaei, B.G. Chae, and P. Tommasi, 2014.
Introduction: Monitoring, prediction and warning of landslides,
Landslide Science for a Safer Geoenvironment
(K. Sassa, P.
Canuti, and Y. Yin, edtors), Springer International Publishing,
pp: 539–544. doi: 10.1007/978-3-319-05050-8_83.
Barazzetti, L., M. Scaioni, and F. Remondino, 2010. Orientation and
3D modelling from markerless terrestrial images: Combining
accuracy with automation,
The Photogrammetric Record
25(132):356–381. doi: 10.1111/j.1477-9730.2010.00599.x.
Brönnimann C., M. Stahli, P. Schneider, L. Seward, and S.M.
Springman, 2013. Bedrock exfiltration as a triggering mechanism
for shallow landslides,
Water Resources Research
, 49: 5155–
5167. doi: 10.1002/wrcr.20386.
Cai, Y., and H. Zhu, 2004. A meshless local natural neighbour
interpolation method for stress analysis of solids,
Analysis with Boundary Elements
, 28(6): 607–613. doi:
Cardenal, J., E. Mata, J.L. Perez-Garcia, J. Delgado, M.A. Hernandez,
A. Gonzalez, and J.R. Díaz de Terán, 2008. Close range digital
photogrammetry techniques applied to landslides monitoring,
International Archives of the Photogrammetry, Remote Sensing
and Spatial Information Sciences
, 37:235–240.
Castagnetti, C., E. Bertacchini, A. Corsini, and A. Capra, 2013. Multi-
sensors integrated system for landslide monitoring: Critical
issues in system setup and data management,
European Journal
of Remote Sensing
, 46:104–124. doi: 10.5721/EuJRS20134607.
Cina, A., and M. Piras, 2014. Performance of low-cost GNSS receiver
for landslides monitoring: Test and results,
Hazardsand Risk
, 1–18. doi: 10.1080/19475705.2014.889046.
July 2016
447...,546,547,548,549,550,551,552,553,554,555 557,558,559,560,561,562,563,564,565,566,...582
Powered by FlippingBook