Tsallis Entropy-Based Nonparametric Focusing Algorithm Under Dual-Stage Processing Framework for THz SAR Imaging

Terahertz Synthetic Aperture Radar (THz SAR) imaging can provide higher resolution, but significantly affected by tiny motion error. Here, THz SAR image focusing-oriented dual-stage processing framework and Tsallis entropy-based nonparametric image focusing algorithm are proposed. Firstly, by distinguishing fundamental principles from specific application and implementation particulars, the proposed framework integrates the model constructed by multiple THz SAR image defocusing factors and the dual-strategy THz SAR image focusing algorithm driven by different defocusing factors. In Stage One, we construct a general platform vibration model based on THz SAR platform characteristics and phase error model related to other non-ideal factors, which are suitable for diverse THz SAR imaging scenarios. In Stage Two, we investigate the dual-strategy focusing algorithm based on specific simple harmonic model and image quality measurement criterion to address different defocusing factors. Secondly, as a specific implementation of the above framework, we construct the THz SAR signal model by considering the multi-component time-varying amplitude vibration and other non-ideal factors, and further propose Tsallis entropy-based THz SAR image focusing method by equivalent nonlinear equations system conversion. To better correct the complex and irregular phase error, the first and second partial derivatives of the Tsallis entropy in phase error estimation are precisely derived. By extending Shannon entropy, Tsallis entropy can further provide a tradeoff between image focusing efficiency and accuracy through changing its non-extensive parameter. Finally, the simulated and real-measured data verify the feasibility of the proposed framework and method by comparing with different methods in complex THz SAR imaging scenes.