Adaptive Digital Image Correlation for imaging deformations in biological materials
Computes deformation fields out of image natural texture by performing digital image correlation between test and reference images.
The ADIC_deformation algorithm is a texture correlation (TC) algorithm, which is optimized and expanded to robustly extract deformation information from neutron radiographies while ignoring "shadow regions", where only poor correlation statistics are available. The inner TC core performs a zero-order search (rigid subset) with a zero-normalized cross-correlation calculated at integer positions and refined with bicubic interpolation. The center position and size of the search window are adaptively adjusted. Error control ensure correlation, uniqueness and continuity of deformation estimates.
The ADIC_strain algorithm computes strain fields by fitting discontinuous deformation fields to continuous cubic smoothing splines and differentiating along the spatial coordinates Strains are defined as: eps_xx = dux/dx; eps_yy = duy/dy; eps_xy = 0.5*(dux/dy + duy/dx).
Run EXAMPLE1.m, EXAMPLE2.m, and EXAMPLE3.m in /examples to get started Each of the examples are explained in detail in the literature reference below.
Input data is included in /data
Results are stored in /results
Source code functions are available in /src
EXAMPLE1.m: CORRELATION OF OPTICAL IMAGES OF SPECKLED SAMPLES
Moisture-induced swelling gradients in growth rings at hygroscopic equilibrium are determined from optical images of speckled softwood samples Reference image: RH = 0%, Test image: RH = 95%
EXAMPLE2.m : CORRELATION OF NEUTRON RADIOGRAPHIES OF SOFTWOOD GROWTH RINGS
Moisture-induced swelling gradients in growth rings at hygroscopic equilibrium are determined from neutron radiographies Reference image: RH = 0%, Test image: RH = 95%
EXAMPLE3.m: CORRELATION OF NEUTRON RADIOGRAPHIES OF WOOD FIBER COMPOSITES
One-dimensional moisture diffusion in three-layer wood fiber composites is determined from neutron radiographies Reference image: t = 0 h (dry state), Test image: t= 17h
The author list for this code repository is:
Sanabria SJ^{a,*},
Lanvermann C^{b},
Franco M^{a},
Mannes D^{a},
Niemz P^{a},
The participating institutions are:
{a} Institute for Building Materials ETH Zurich, Stefano-Franscini-Platz 6, CH-8093, Zurich, Switzerland
{b} Neutron Imaging and Activation Group, Paul Scherrer Institute, CH-5232 PSI, Villigen, Switzerland
*Corresponding author: E-mail: sanse@stanford.edu
Please cite the original work in:
- Sanabria Martin, Sergio & Lanvermann, Christian & Michel, Franco & Mannes, David & Niemz, Peter. (2014). Adaptive Neutron Radiography Correlation for Simultaneous Imaging of Moisture Transport and Deformation in Hygroscopic Materials. Experimental Mechanics. 55. 403-415. 10.1007/s11340-014-9955-2. Preprint