Entropy weighted TOPSIS optimized interferometric synthetic aperture focusing technique (InSAFT) for laser-ultrasonic...

Laser-ultrasonic synthetic aperture focusing technique (SAFT) imaging typically suffers from a high noise level, and its imaging quality further depends on manually specified parameters such as the excitation-detection distance and the imaging frequency band. In this study, an interferometric SAFT (InSAFT) imaging method optimized by entropy weighted TOPSIS is proposed for laser-ultrasonic defect imaging. Inspired by the interferometric processing concept of InSAR, the conjugate product of two-channel complex frequency-domain SAFT images is constructed to obtain their relative phase information, and a normalized frequency-domain coherence coefficient is used to characterize the phase consistency between the two channels. The coherence coefficient is then used to weight the reconstructed image and suppress low-coherence noise. Furthermore, to eliminate the subjectivity of parameter selection, four evaluation metrics, covering defect detectability, indication compactness, defect-background separability, and overall artifact level, are combined with the entropy weighted TOPSIS method to objectively rank the candidate configurations of the excitation-detection distance, detection spacing, and imaging frequency band. Numerical simulations and laser-ultrasonic full matrix capture (FMC) experiments on aluminum plates containing horizontal hole defects were conducted to validate the proposed method. The results show that the optimized InSAFT suppresses background noise while preserving the defect response. In the experiment, the pDB increased from 6.07 for conventional PSM to 11.52, the API decreased from 0.45 to 0.42 mm², the gCNR increased from 0.33 to 0.39, and the AE decreased from 0.17 to 0.10. Compared with manually selected InSAFT parameters, the TOPSIS-selected configuration achieved better values for all four evaluation criteria. The proposed method provides an objective and automated approach for improving laser-ultrasonic defect imaging.