Super-resolution terahertz imaging algorithm based on blind deconvolution

Achieving a high resolution is crucial in the increasingly popular field of terahertz imaging, and diffraction-limited imaging remains a significant practical challenge. In synthetic aperture radar imaging, antennas are typically treated as point sources, and the effect of the antenna aperture is ignored. This study illustrates the impact of the applied antenna on imaging performance and provides a method for achieving super-resolution imaging through compensation within the aperture field of the antenna. Simulation experiments verify the robustness of the algorithm under low SNR conditions. The image contrast can be used for deconvolution kernel iteration. Iteration over target distances and deconvolution kernels accelerates the process of blind deconvolution imaging. By distinguishing from time-domain pulse and stepped-frequency continuous wave measurements, the proposed algorithm allows accurate calculation of the distance to the target under single-frequency measurement conditions. The algorithm improves the resolution from a theoretical value of 5.5-3 mm.