Crack Characterization Using Multi-Scale Scattering Matrix Denoising Extrapolation and Optimal Scale Selection
e-Journal of Nondestructive Testing · 2026-07 · 31 (7)
DOI: 10.58286/33195 ↗At a glance
Journal article in e-Journal of Nondestructive Testing, 2026-07. Written by Yiliang Hu and 1 co-authors.
Abstract (original language)
The ultrasonic scattering matrix shows strong potential for small-defect characterization. This paper proposes a novel processing framework for multiscale scattering matrix denoising, extrapolation, and adaptive crack parameter regression. The multi-scale scattering matrix denoising extrapolation method can process scattering matrices measured within a small angular range to generate low-noise scattering matrices at various larger scales. Optimal scale selection method adaptively emphasizes informative scales for robust estimation of crack size and orientation. Simulation studies demonstrated that large-scale matrices ([-75°, 75°] and [-90°, 90°]) were most frequently selected as optimal, with average proportions of approximately 62% and 26%, respectively, indicating that wide-angle reconstructions retain more stable scattering features and improve noise robustness. Crack dimensions (MAE: 0.043λ–0.048λ, R² > 0.98) and orientation angles (MAE: 3.34°–7.80°, R²: 0.971–0.784) were estimated with high accuracy. Experimental validation on six aluminum cracks with sizes 0.8λ and orientation angles 0°–75° achieved RMSEs of 0.097λ for size and 3.76° for orientation, confirming the method’s practical robustness and effectiveness in ultrasonic non-destructive testing (NDT).
Cite
APA
Hu, Y., & Bai, L. (2026). Crack Characterization Using Multi-Scale Scattering Matrix Denoising Extrapolation and Optimal Scale Selection. e-Journal of Nondestructive Testing, 31(7). https://doi.org/10.58286/33195
Vancouver
Hu Y, Bai L. Crack Characterization Using Multi-Scale Scattering Matrix Denoising Extrapolation and Optimal Scale Selection. e-Journal of Nondestructive Testing. 2026;31(7). doi:10.58286/33195
BibTeX
@article{Hu2026,
author = {Hu, Yiliang and Bai, Long},
title = {Crack Characterization Using Multi-Scale Scattering Matrix Denoising Extrapolation and Optimal Scale Selection},
journal = {e-Journal of Nondestructive Testing},
year = {2026},
volume = {31},
number = {7},
doi = {10.58286/33195}
}
EUVIMED