A sparse reconstruction algorithm for ultrasonic images in nondestructive testing

Area of Science:

• Materials Science
• Non-destructive Testing (NDT)
• Ultrasonic Imaging

Background:

• Ultrasound imaging systems (UIS) are critical for non-destructive testing (NDT).
• Image quality in UIS is determined by hardware and image reconstruction algorithms.
• Existing algorithms have limitations in resolution and signal clarity.

Purpose of the Study:

• To introduce a novel image reconstruction algorithm for ultrasonic NDT.
• To enhance the resolution and signal-to-noise ratio (SNR) of ultrasound images.
• To evaluate the performance against traditional NDT imaging techniques.

Main Methods:

• Developed a new image reconstruction algorithm based on regularized least squares with an l1 regularization norm.
• Acquired A-scan signals using an ultrasonic imaging system with a monostatic transducer in pulse-echo mode.
• Tested the algorithm with simulated and real data, comparing it against B-scan, SAFT, ω-k SAFT, and RLS algorithms.

Main Results:

• The proposed algorithm achieved a significant resolution improvement of approximately 91% compared to B-scan using real data.
• Demonstrated superior performance in terms of signal-to-noise ratio (SNR) over traditional algorithms.
• Successfully reconstructed images of a point-like reflector.

Conclusions:

• The novel l1-regularized least squares algorithm offers substantial improvements in ultrasonic image reconstruction for NDT.
• This method enhances resolution and SNR, leading to more accurate defect detection.
• The algorithm presents a promising advancement for ultrasonic imaging applications.