Sparse deconvolution method for ultrasound images based on automatic estimation of reference signals.

Area of Science:

• Materials Science
• Acoustics
• Signal Processing

Background:

• Sparse deconvolution enhances temporal resolution in non-destructive testing (NDT).
• Traditional methods use standard reference signals, leading to inaccuracies due to spatial variations in acoustic properties.
• Manual reference signal measurement is inefficient for automated ultrasonic NDT.

Purpose of the Study:

• To develop a modified sparse deconvolution technique with automatic reference signal estimation.
• To improve the accuracy and efficiency of ultrasonic NDT.
• To enable regional sparse deconvolution for reduced computational load.

Main Methods:

• Automatic estimation of reference signals to minimize deviations.
• Decomposition of B-scan images into Regions of Interest (ROI) for regional processing.
• Implementation of sparse deconvolution on estimated reference signals.

Main Results:

• Alleviation of deviations caused by inaccurate reference signals.
• Improved accuracy of sparse deconvolution.
• Significant reduction in image dimensionality and computation time through regional processing.
• Maintained vertical resolution enhancement and noise suppression.

Conclusions:

• The proposed method enhances sparse deconvolution performance in ultrasonic NDT.
• Automatic reference signal estimation and regional processing improve accuracy and efficiency.
• The technique is effective across various scattering targets and materials.