SEA-RAFT-DIC: displacement-field measurement under degraded speckle conditions based on a laplacian probability distribution

Digital image correlation (DIC) is widely used for full-field displacement measurement, but image degradation caused by occlusion, texture blurring, and speckle contamination, as well as rapid local correspondence variations induced by high displacement gradients and nonuniform deformation, can lead to matching failures and error amplification. Conventional DIC is constrained by the trade-off between spatial resolution and robustness, while existing deep learning-based methods generally lack explicit correspondence modeling and uncertainty constraints. To address these limitations, this study proposes SEA-RAFT-DIC, a DIC-oriented framework based on the global correlation and iterative update mechanisms of SEA-RAFT. A two-layer SmallStem, squeeze-and-excitation channel attention, and atrous spatial pyramid pooling are incorporated into the feature encoder to enhance local speckle details, informative channel responses, and multiscale displacement representation. A Mixture-of-Laplace Gradient Consistency Loss combines a mixture-of-Laplace probabilistic term, average endpoint error, and displacement-gradient consistency to model matching uncertainty while preserving displacement accuracy and structural continuity. A Hermite displacement-field speckle dataset and three degraded test sets containing occlusion, blurred textures, and salt-and-pepper noise are constructed. SEA-RAFT-DIC shows strong generalization even without degradation augmentation. After unified degradation-augmentation training, the full-field AEE values on the three test sets decrease to 0.2035 (occlusion), 0.1759 (blurred textures), and 0.1489 (salt-and-pepper noise), respectively, while the corresponding degraded-region values are 0.7953 (occlusion), 0.5343 (blurred textures), and 0.2748 (salt-and-pepper noise). A tensile experiment on a silicone plate with a prefabricated crack further demonstrates that SEA-RAFT-DIC is applicable to real high-displacement-gradient fields near crack tips and achieves favorable measurement accuracy, robustness, and practical potential under degraded speckle and complex deformation conditions.

Authors

Institutions

Publication Details

Journal
Optics & Laser Technology
Published
2026-10-17
DOI
https://doi.org/10.1016/j.optlastec.2026.116741
Primary Topic
Optical measurement and interference techniques
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

SEA-RAFT-DIC: displacement-field measurement under degraded speckle conditions based on a laplacian probability distribution

Dong Zhao, Jian Wei Zhao, Fanyu Wang, Lei Yan et al.
Optics & Laser Technology
Optical measurement and interference techniques
article

SEA-RAFT-DIC: displacement-field measurement under degraded speckle conditions based on a laplacian probability distribution

Dong Zhao, Jian Wei Zhao, Fanyu Wang, Lei Yan, Hongye Zhang
article en

Abstract

Digital image correlation (DIC) is widely used for full-field displacement measurement, but image degradation caused by occlusion, texture blurring, and speckle contamination, as well as rapid local correspondence variations induced by high displacement gradients and nonuniform deformation, can lead to matching failures and error amplification. Conventional DIC is constrained by the trade-off between spatial resolution and robustness, while existing deep learning-based methods generally lack explicit correspondence modeling and uncertainty constraints. To address these limitations, this study proposes SEA-RAFT-DIC, a DIC-oriented framework based on the global correlation and iterative update mechanisms of SEA-RAFT. A two-layer SmallStem, squeeze-and-excitation channel attention, and atrous spatial pyramid pooling are incorporated into the feature encoder to enhance local speckle details, informative channel responses, and multiscale displacement representation. A Mixture-of-Laplace Gradient Consistency Loss combines a mixture-of-Laplace probabilistic term, average endpoint error, and displacement-gradient consistency to model matching uncertainty while preserving displacement accuracy and structural continuity. A Hermite displacement-field speckle dataset and three degraded test sets containing occlusion, blurred textures, and salt-and-pepper noise are constructed. SEA-RAFT-DIC shows strong generalization even without degradation augmentation. After unified degradation-augmentation training, the full-field AEE values on the three test sets decrease to 0.2035 (occlusion), 0.1759 (blurred textures), and 0.1489 (salt-and-pepper noise), respectively, while the corresponding degraded-region values are 0.7953 (occlusion), 0.5343 (blurred textures), and 0.2748 (salt-and-pepper noise). A tensile experiment on a silicone plate with a prefabricated crack further demonstrates that SEA-RAFT-DIC is applicable to real high-displacement-gradient fields near crack tips and achieves favorable measurement accuracy, robustness, and practical potential under degraded speckle and complex deformation conditions.

Optics & Laser TechnologyVol. 204
Beijing Forestry University (CN)
Openalex Percentile: Top 16%
Optical measurement and interference techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.