High-Precision 3D-DIC Measurement Method Based on Improved Forward Newton Iteration.

Area of Science:

• Mechanical Engineering
• Optical Measurement
• Materials Science

Background:

• Traditional 3D Digital Image Correlation (DIC) methods face challenges with accuracy, robustness, and anti-noise performance.
• Feature-based or FFT search algorithms in 3D-DIC often compromise accuracy for speed, leading to errors like poor feature point matching.

Purpose of the Study:

• To develop an improved high-precision 3D-DIC measurement method addressing the limitations of traditional algorithms.
• To enhance accuracy, robustness, and computational efficiency in 3D-DIC measurements.

Main Methods:

• An improved 3D-DIC method utilizing an exhaustive search for an exact initial value.
• Employing the forward Newton iteration method with pixel classification and first-order nine-point interpolation for accurate sub-pixel positioning.
• Efficient calculation of Jacobian and Hazen matrices for improved performance.

Main Results:

• The improved method demonstrates high accuracy, with superior mean error, standard deviation stability, and extreme value performance compared to similar algorithms.
• Achieved a 3.8 times increase in computational efficiency compared to the traditional Newton-Raphson (NR) algorithm.
• Reduced iteration time in the sub-pixel iteration stage.

Conclusions:

• The proposed 3D-DIC method offers a simple, efficient, and highly accurate solution for precision measurement applications.
• It overcomes the accuracy and robustness limitations of traditional 3D-DIC techniques.
• The method shows significant potential for use in high-precision engineering and scientific fields.