Answers are grounded in this paper and cite their sources.
Digital image correlation (DIC) is widely used in fields such as experimental mechanics, material testing and industrial inspection due to its non‐contact, high‐precision and full field measurement characteristics. However, when measuring high‐temperature objects, heat haze can cause image dithering and blurring, seriously reducing the accuracy of DIC measurements. Existing research on heat haze removal for high‐temperature DIC measurement mostly relies on multi‐frame image acquisition and fusion to reconstruct clear speckle images, which greatly reduces the efficiency of DIC measurement and is not applicable to dynamic deformation measurement scenarios. To solve this problem, this paper proposes a heat haze removal network (HHR‐Net) dedicated to the preprocessing of high‐temperature DIC speckle images, which can output high‐quality clear speckle images with only a single frame of heat haze‐contaminated speckle image as input. Specifically, a multi‐scale wavelet Transformer (MWT) module is designed to effectively model the global spatial dependence of heat haze distortion while preserving the high‐frequency detail information of speckle images that is critical for DIC matching. In addition, a grouped channel‐space convolutional attention block (GCSAB) is proposed to adaptively extract heat haze distortion features in different image regions and complete the accurate reconstruction of clear speckle images. A physical experimental platform is built to generate a heat haze dataset. Simulation and physical experiments confirm the method effectively suppresses heat haze interference, improving speckle image quality and deformation measurement accuracy.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: