Inversion Method Based on Generalized Plane Wave Spectrum in Terahertz Computed Tomography

Read full text
Authors: En Li, Chengyong Yu, Tianqi Feng, Yu Shi, Hu Zheng, Dakun Lai
Cited by: 0

Abstract

Terahertz computed tomography (THz-CT) is a popular technique for non-destructive testing. Traditional methods of THz-CT originated from X-ray CT (X-CT) that are both based on the Radon transform. However, those methods only consider the rays propagating along straight lines but neglect the scattering behavior of terahertz waves -- particularly refraction and corner scattering, which may result in poor reconstructions. At present, existing research has primarily focused on refractive behavior, but the impact of corner scattering on THz-CT cannot be ignored. This paper proposes an inversion method based on generalized plane wave spectrum (GPWS), and it can account for the corner scattering characteristics of layer structures that consist of (semi-)infinite layers. Considering the field differences arising from an ideal plane wave used in theoretical analysis vs a Gaussian beam employed in experiments, this paper proposes a field correction that effectively minimizes the discrepancies between these two conditions. To validate the proposed inversion method, a corner scattering field test system was designed and constructed. This paper prepared a PEC block as the basic model, and used the model to compare the proposed inversion method with the Order Subsets Expectation Maximization (OSEM) method. The accuracy of the proposed method is well validated by the experimental comparison. Additionally, this paper also prepared a PEC stripe and dielectric blocks wrapped in a foam box. The inversion results demonstrate that the proposed method is accurate for PEC samples and feasible for dielectric blocks.

Fields of Study & Subjects

Terahertz technology and applications Microwave Imaging and Scattering Analysis Numerical methods in inverse problems

Publication details

Journal / source
AIP Publishing
Publication date
11 March 2026
Language
EN