Texture-driven fringe pattern evaluation in shearography based on gray-level co-occurrence matrices

Abstract Speckle shearography is a powerful non-contact optical method for full-field strain measurement, yet its performance can be constrained by fluctuations in laser intensity. In this study, a Michelson-interferometer-based shearography system integrated with a coaxial circular optical attenuator is developed to regulate optical density (OD) and quantitatively evaluate fringe quality. By varying light transmittance while maintaining constant source coherence and imaging parameters, we analyze how fringe contrast and measurement reliability evolve under different shearing amounts and loading intensities. The gray-level co-occurrence matrix (GLCM) is employed to extract four statistical texture features, i.e. contrast, correlation, homogeneity, and energy, from both experimental and simulated fringe patterns. Results reveal that a moderate OD value of around 0.15 optimizes fringe visibility and measurement accuracy by balancing transmittance and noise suppression. The combined use of GLCM-based descriptors and heatmap visualization establishes a robust, objective framework for fringe-pattern assessment, providing new guidance for intelligent optical quality control in shearographic systems.