Research on internal defect detection of multilayer media based on phase coherence far-focused pixel-based imaging method.

In ultrasonic inspection of internal defects within multilayer media, traditional amplitude-based imaging methods frequently suffer from low signal-to-noise ratio (SNR) and inconsistent defect characterization, particularly for identically sized flaws at varying depths, primarily due to interface reflections and material attenuation. To address these challenges, a phase-coherent far-focused pixel-based (PC-FPB) imaging algorithm is proposed in this paper. This method constructs coherent imaging signals using a vector coherence factor and introduces a novel scan-line-based coherence modulation strategy. A dynamic threshold is statistically determined from the distribution of maximum coherence factors across scan lines and applied through a multiplicative fusion mechanism to selectively enhance coherent defect signals and suppress background noise. Experimental validation on multilayer structures demonstrates that the imaging intensity and SNR for deep-seated defects are significantly enhanced by the PC-FPB imaging algorithm. Comparative results further indicate that it outperforms conventional amplitude-based methods in the detection of deep defects. For single-hole imaging, the defect amplitude was improved by 9.8 dB and the image SNR increased by 7.7 dB. In multi-hole imaging, the maximum quantification error for defects with a diameter of 1 mm was recorded at 0.245 mm2, demonstrating a high degree of consistency in the imaging of identically sized defects at varying depths. Furthermore, the average SNR of defects at various depths increased by 15 dB and the average lateral resolution improved by 53.4%. The findings confirm that PC-FPB effectively overcomes the limitations of traditional methods and enhances defect detection in multilayer structures.