Improved measurement of the glue layer in composite material by using sparse deconvolution.

Area of Science:

• Physics
• Materials Science
• Signal Processing

Background:

• Terahertz (THz) technology offers superior penetration and resolution for inspecting thin layers in multilayered materials.
• Identifying echo arrival times in THz signals is difficult due to incident pulse shape and system noise.

Purpose of the Study:

• To develop and evaluate sparse deconvolution algorithms for precise THz echo time identification.
• To improve the accuracy and efficiency of THz-based material inspection.

Main Methods:

• Investigated two terahertz signal sparse deconvolution algorithms.
• Utilized the circulant structure of the convolution matrix to reduce computation time.
• Employed group velocity dispersion to mitigate time-varying pulse effects with low computational cost.

Main Results:

• Achieved accurate pinpointing of echo arrival times from experimental THz data.
• Reduced computation time to hundreds of milliseconds for the deconvolution methods.
• Demonstrated minimal computational expense for the group velocity dispersion method.

Conclusions:

• The developed algorithms enhance the accuracy of identifying thin glue layers using THz signals.
• Fast computation times and high confidence make these algorithms suitable for real-time industrial inspection.
• These advancements facilitate improved quality control in manufacturing processes.