Digital Image Correlation glossary

Definitions of the key terms used in digital image correlation (DIC) and simulation validation, as used throughout our technical articles.

Glossary

Digital Image Correlation (DIC)

An optical, non-contact, non-destructive measurement method that tracks the displacements and deformations of a solid by comparing camera images to a reference image. It works regardless of material or shape, as long as the region of interest can be filmed.

Speckle pattern

A random texture applied to a part’s surface (often spray paint producing millimetre-sized spots) that lets the algorithm recognise each zone from one image to the next. Without a suitable speckle pattern, correlation is impossible.

Subset

A small window of an image (a subset of pixels) that the correlation algorithm tracks from one image to the next to measure its displacement. Correlation amounts to tracking all the subsets across the observed surface.

Camera calibration

Estimation of a camera’s intrinsic and extrinsic parameters to determine its projection matrix (mapping 3D points to the 2D image). An essential step for any multi-camera metrological measurement.

Self-calibration

A calibration method that uses the tested part itself (via a dense description of its geometry) as the calibration object, allowing a stereovision system to be calibrated from a single image pair and recalibrated during the test.

Local DIC

A correlation approach that processes each subset independently to derive a displacement, following the subset principle.

Global DIC

A correlation approach that solves the displacement field on a finite element mesh covering the whole region of interest, ensuring field continuity and direct comparison with simulation.

Mechanical regularization

A technique that imposes mechanical consistency on the measured displacement field, useful when the mesh is too fine relative to the speckle; it ties elements together by imposing the ideal characteristic length.

Model updating (FEMU)

Automatic adjustment of a finite element model’s parameters to minimise the gap between simulation and measurements, via the FEMU method (Finite Element Model Updating), weighting the identification by measurement uncertainty.

Correlation residual

The residual difference between the deformed image and the reference image registered by the measured field. A low residual indicates a reliable measurement; its analysis notably allows crack detection.

Digital twin

A virtual representation of a physical system, fed by test data, that allows the simulated behaviour to be continuously validated and refined by comparison with reality.

VVUQ

Verification, Validation and Uncertainty Quantification: a systematic approach to assessing the quality and reliability of numerical simulations, by verifying the models and algorithms and quantifying uncertainties.