Nonlinear forward and reconstruction framework for terahertz computed tomography.

Area of Science:

• Physics
• Imaging Science
• Materials Science

Background:

• Terahertz computed tomography (THz-CT) utilizes non-ionizing terahertz waves for industrial non-destructive testing.
• Current THz-CT methods often assume straight-line propagation, ignoring significant wave refraction in inhomogeneous materials, thus limiting reconstruction accuracy.

Purpose of the Study:

• To develop a nonlinear forward modeling and reconstruction framework for THz-CT.
• To address the limitations of existing methods by incorporating the effects of THz wave refraction.

Main Methods:

• An iterative ray propagation algorithm based on Snell's law was developed to model nonlinear THz propagation paths, including unknown refractive indices and interfaces.
• A derivative-free differential evolution algorithm was employed to solve the ill-posed, nonconvex, and nonsmooth inverse problem of refractive index reconstruction.
• Transmission coefficient recovery was formulated as a total variation (TV)-regularized optimization problem solved using the alternating direction method of multipliers (ADMM).

Main Results:

• The proposed nonlinear framework accurately computes nonlinear THz paths after recovering refractive index and interface geometry.
• Experimental results on real THz datasets showed significant improvements in reconstruction accuracy compared to linear-ray-based approaches.

Conclusions:

• The developed nonlinear forward modeling and reconstruction framework effectively overcomes the limitations of linear assumptions in THz-CT.
• This advancement offers enhanced accuracy for non-destructive testing applications utilizing THz imaging.