Grid-based matching for full-field large-area deformation measurement

• The grid-based matching method has been systematically studied for full-field deformation measurement.
• The simple deviation metric can globally register deformed grids accurately and robustly.
• The constellation matching algorithm can match distorted targets in noisy and blurred images quickly, accurately and robustly by balancing global registration and local deformation constraints.

Grid-based measurement can facilitate metrology and inspection of flexible electronics manufacturing. Multiple fundamental difficulties, however, arise in the large-area and full-field deformation measurement of deformable grid patterns including noise, occlusions, and artifacts. This paper addresses one of the key issues in deformation measurement: the registration and matching of deformed grid patterns. The emphasis is on accurate and robust periodicity tracing registration and constellation matching algorithms for grid pattern fidelity. The registration algorithm uses deviation metrics in deformed grids to estimate global translation, rotation and scaling; the matching algorithm uses the constellation reference grid to mine buried deformed point patterns. Using synthetic data, the validity of the registration algorithm is proved by registering noisy deformed grid patterns with various distortion scales and transformations; the validity of the matching algorithm is proved by matching deformed grid point patterns with various distortion scales, extra point rates and missing point rates. Compared to established non-rigid registration and point pattern matching algorithms, our algorithms demonstrate higher speed, sub-pixel accuracy and robustness in the matching of highly-deformed and noisy grids.