Ultrasonic Phased Array Compressive Imaging in Time and Frequency Domain: Simulation, Experimental Verification and...

Area of Science:

• Non-destructive testing
• Ultrasonic phased array imaging
• Signal processing

Background:

• Phased array systems generate large datasets, posing challenges for material characterization.
• Compressive sensing (CS) offers a method to reduce data acquisition requirements.
• CS can potentially compress ultrasonic phased array images in time and frequency domains.

Purpose of the Study:

• To investigate the application of compressive sensing (CS) for data reduction in ultrasonic phased array imaging.
• To evaluate the reconstruction accuracy of CS in both time and frequency domains.
• To demonstrate the potential of CS for efficient material characterization.

Main Methods:

• CS framework applied to compress ultrasonic phased array images.
• Projection of images onto the Discrete Cosine Transform (DCT) domain.
• Verification using CIVA simulations and experimental data of artificial flaws.
• Case study involving inspection of an engine cylinder cavity with pit defects.

Main Results:

• CS accurately reconstructs array images with fewer samples than Nyquist requirements in simulations (time and frequency domains).
• Experimental verification shows considerable data reduction with preserved defects, but time-domain CS cannot break Nyquist limitation.
• Frequency domain CS successfully reconstructs images below Nyquist limits, achieving a breakthrough.
• Orthogonal Matching Pursuit (OMP)-based CS demonstrated reliable performance in a real-world engine cylinder inspection.

Conclusions:

• CS is a viable technique for significant data reduction in phased array-based material characterization.
• Frequency domain CS represents a breakthrough, enabling image reconstruction below Nyquist sampling rates.
• OMP-based CS provides a robust solution for real-world applications like engine cylinder inspection.