Pixel-level robust digital image correlation.

Area of Science:

• Optical Metrology
• Computational Mechanics
• Image Analysis

Background:

• Digital Image Correlation (DIC) is a key non-contact optical technique for measuring object deformation under stress.
• Existing DIC methods struggle with accuracy issues caused by factors like sharp displacement discontinuities, reflections, and image noise.
• These limitations hinder reliable full-field displacement and strain analysis in challenging scenarios.

Purpose of the Study:

• To develop a novel, robust, pixel-level Digital Image Correlation (DIC) method for in-plane displacement measurement.
• To address and unify solutions for common DIC accuracy problems, including discontinuities, reflections, and noise.
• To significantly enhance DIC measurement accuracy compared to traditional approaches.

Main Methods:

• A new subset-based, pixel-level robust DIC method is proposed.
• The approach minimizes a robust energy functional, adaptively weighting pixel differences for motion estimation.
• Local motion consistency is enforced to mitigate the influence of erroneous pixel motions.

Main Results:

• The proposed robust DIC method demonstrated improved displacement accuracy over the classic Newton-Raphson method in numerical and real-world experiments.
• The method successfully handled challenging conditions, including sharp displacement discontinuities, missing image information, reflections, and image noise.
• Reliable motion recovery was achieved even in image areas where classic DIC approaches failed due to severe decorrelation.

Conclusions:

• The proposed robust DIC method offers a significant advancement in accurate displacement and strain measurement.
• It provides a unified and straightforward approach to overcoming common limitations in optical metrology.
• This technique enhances the reliability of DIC analysis in complex and demanding applications.