Heat haze removal in digital image correlation using a grayscale weighted average algorithm

Digital image correlation has emerged as a vital tool for quantifying full-field deformation in high-temperature material characterization. Nevertheless, measurement accuracy faces critical challenges from heat haze induced image distortions. This study presents a novel Grayscale Weighted Averaging algorithm integrated with Zero-Mean Normalized Cross-Correlation to mitigate these artifacts. Through systematic numerical simulations and experimental validations under controlled thermal loading, the Grayscale Weighted Averaging algorithm enhanced method demonstrates superior performance: 5% PSNR improvement enhancement compared to conventional methods. Simulative tensile strain analysis reveals 28.5% reduction in relative error and 50% decrease in standard deviation of displacement measurements. Experiments in simulated operating environment of aero-engine at 1423K and industrial furnace environments at 573K confirm the algorithm's robustness, achieving 49.1% artifact suppression while maintaining displacement field regularity at 1423K. The variance coefficient of strain measurements decreases by 23.2% under severe heat haze conditions. This optical optimization strategy significantly advances high-temperature DIC metrology through its dual capability of image quality enhancement and measurement uncertainty reduction, particularly valuable for high-temperature digital image correlation.