How DIC works · semper

Enough theory to use semper and read the Manual — no prior DIC experience needed.

Digital image correlation watches a speckle pattern move between photos and turns that motion into full-field maps of how the surface stretched.

Five steps, start to finish:

A specimen is painted with a random speckle pattern — a spray of dots.
A reference photo is taken, then more photos while it deforms.
The image is divided into small windows called subsets (a subset is the patch of pattern tracked at each point).
For each subset the engine finds where it moved in the deformed image, to about 1/100th of a pixel. It uses ICGN to home in and ZNSSD (a match-quality score that ignores lighting changes; lower is better) to judge the fit.
Displacements (U, V in pixels) are differentiated into strains (Exx, Eyy, Exy in millistrain).

Field	Meaning
U, V	Horizontal / vertical displacement (px)
Exx, Eyy, Exy	Normal and shear strain (mε)

semper runs this pipeline on the phone. There is no scale calibration — convert pixels to physical units yourself. For experiment design, see the iDICs Good Practices Guide.