Defocusing Recovery Technology of Terahertz Image Based on 3-D PSF Simulations

Terahertz (THz) imaging has broad application prospects in nondestructive testing (NDT). However, compared to visible light, it has a low resolution and limited depth of field owing to its long wavelength. To address these limitations, a three-dimensional point-spread function based on a simulated THz time-domain spectroscopy system combined with the Lucy-Richardson method is proposed for image defocus restoration. The experimental results show that the contrast of THz images at different defocus positions is 2-5 times higher than that of the original image after restoration using the proposed method, and the resolution is improved by approximately 60-80%. Moreover, the maximum contrast enhancement at the boundaries of tiny features can reach 1.81 times. For a hole-defect sample, the relative error of the defect area at different depths is <5%. The proposed method solves the problem of defocus in THz images, enhances the resolution of these images, and facilitates NDT.