Image registration method for full-field deformation measurement

This study proposes an automated image alignment technique for measuring full-field deformation fields in materials. The Maximally Stable Extremal Regions (MSER) method is employed to extract distinct feature points from grating images, which serve as benchmarks for image alignment. The effectiveness of the alignment method is validated under both high-noise and low-noise conditions, with alignment errors maintained within ±0.1 pixels during image translation and stretching. Additionally, tensile tests on notched single-crystal nickel alloy specimens are designed and conducted. The results demonstrate that the proposed alignment method effectively reduces dislocation areas in images and enhances strain field resolution in experimental scenarios. The proposed image registration method can transform the sampling moiré method from a semi-automatic to a fully automatic deformation measurement method, while maximizing the effective measurement area near defects such as cracks. It can also be applied to other image-based analysis methods, such as digital image correlation, geometric phase analysis, and the moiré method.