Microscopic DIC measurement using an improved global registration image-stitching algorithm

Area of Science:

• Materials Science
• Mechanical Engineering
• Optical Metrology

Background:

• Microscopic Digital Image Correlation (DIC) is crucial for precise deformation analysis.
• Large-field microscopic measurements face challenges in balancing accuracy and computational efficiency.
• Existing methods struggle with accuracy and speed in extensive microscopic strain measurements.

Purpose of the Study:

• To develop an improved microscopic DIC method using a global registration image-stitching algorithm.
• To enhance strain measurement accuracy and computational efficiency in large-field microscopic applications.
• To extend the measurement range of DIC while maintaining high precision.

Main Methods:

• Capturing multiple adjacent microscopic images of structures.
• Stitching images using an improved global registration optimization algorithm for pre- and post-deformation analysis.
• Applying DIC to compute the deformation field on stitched images.

Main Results:

• The proposed method improved the average structural similarity index by 27% and reduced stitching time by 16%.
• Stitched image errors for displacements as small as 0.1 pixels were controlled within 0.02 pixels (average error and standard deviation).
• Micrometer-level strain error was maintained within 20 µε in practical measurements.

Conclusions:

• The improved DIC method effectively balances accuracy and computational efficiency for large-field microscopic measurements.
• The method significantly enhances the measurement range of DIC without compromising precision.
• Validated through simulations and experiments, the approach offers a robust solution for microscopic deformation analysis.