Lecture 4: Digital Image correlation Flashcards
(8 cards)
Landslide
The term ‘landslide’ describes
a variety of processes that
result in the downward and
outward movement of slope-
forming materials, including
rock, soil, artificial fill or a
combination of these.
Landslide causes and triggers
can be natural or/and human-
related
What does control the occurrence and motion of landslides?
Threshold phenomenon
Factor of safety:
Shear strength resists motion; shear stress drives motion.
Forces Acting on a Slope
▪ Gravity stress split into normal stress (perpendicular) and shear stress (parallel).
▪ Pore-water pressure weakens the slope by reducing effective normal stress.
▪ Failure occurs along a potential failure plane when shear stress > shear strength.
External factors affecting slope stability:
▪ Loading (buildings, fill)
▪ Rainfall/infiltration (raises pore pressure)
▪ Weathering, seismicity
▪ Vegetation removal, reservoir filling
▪ These can increase shear stress or reduce shear strength, pushing the system past threshold.
Susceptibility Factors per Region
Interpretation of a comparative bar chart showing key contributing factors across regions.
E.g. in Europe, slope angle and rainfall may dominate, whereas in Asia, seismicity may play a larger role.
Useful for understanding regional controls on landslide hazard.
Threshold phenomenon
A condition that must be
exceeded for state to change
within the system
Digital Image Correlation (DIC)
Digital Image Correlation (DIC)
➢ Method combining registration and tracking techniques to measure 2D or 3D changes in images and derive displacements and strain changes
➢ = specific branch of Photogrammetry and Computer Vision
➢ Based on the comparison of a reference image and a target image
➢ Not just image comparison → can track deforming subsets of images over time
Examples:
➢ Engineering, mechanical testing
→ Usually called Digital Image Correlation (DIC)
➢ Ground deformation and surface displacements (earthquakes, faults, landslides, etc.)
→ Usually called DIC or (sub-)pixel offset tracking
➢ Fluid movements (water, air, pollutants, lava, etc.)
→ Usually called Particle Image Velocimetry (PIV)
Main principle of 2D-DIC (Digital Image Correlation)
Two core approaches to compare images:
1. Using the spatial domain → e.g. Normalized Cross-Correlation (NCC)
≈ Template matching
2. Using the frequency domain → Phase correlation using FFT
≈ Faster, more robust to noise
2D-DIC with a pinhole camera
Distortions related to the camera
and optics
Different types of distortion:
➢ Radial (K): barrel or pincushion effect
➢ Tangential (P): caused by lens alignment issues
➢ Skew (B): non-orthogonality between sensor axes
In photogrammetry, the geometry of
image acquisition characterized by
these distortion parameters is called:
“interior orientation”
Photographs must be orthorectified before DIC analysis:
➢ Lens correction / camera calibration (removes distortion)
➢ Topographic correction (removes terrain-induced distortion)
2D-DIC with satellite imagery
➢ Preferable use of orthorectified
products
➢ Sometimes correction needed for
distortion related to image
acquisition
▪ e.g., jitter effect = short-term periodical variation of the sensor’s position
Photography
Photography = Art of capturing light with a camera, usually via a digital sensor or film, to create an image.
3 main settings:
▪ Shutter speed = length of time during which the camera shutter is open
▪ Aperture = opening or diaphragm in which the light enters the camera to reach the sensor.
▪ ISO = Level of sensitivity (to the light) of the sensor