THz-ViSAR-Oriented Fast Indication and Imaging of Rotating Targets Based on Nonparametric Method

Rotating targets/components often carry beneficial information for identifying high-value targets in video synthetic aperture radar (ViSAR). However, unlike linear motion models, rotating motion models involve more parameters to be estimated, rendering parametric-based methods computationally intensive. Besides, such micro-motion features are relatively insignificant in the microwave band. To conquer the aforementioned problems, this article proposes the nonparametric indication and imaging of rotating targets under terahertz (THz) ViSAR. Taking rotating corner reflectors (RCRs) as the object, the Doppler prior and imaging characteristics (DPIC) are first determined and regarded as the theoretical basis. Combining DPIC and Doppler sensitivity of THz band, RCRs can be easily indicated by the nonparametric time–frequency transform. Thus, range cells of RCRs can be deduced intuitively by the visual saliency, followed by the separation of the residual background and defocused target area. After clutter suppression, two signal of interests (SoIs) of RCRs are extracted from salient points, where unwrapped phases and azimuth spectrums are utilized for fast estimation of phase errors and additional Doppler centroids (DCs), respectively. Finally, the refocused target image is stitched into the residual background. To validate the effectiveness of the proposed method, the airborne field experiment of RCRs is conducted in THz-band for the first time. Moreover, the anti-noise ability, relocation capability, and applicable speed interval of the proposed method are demonstrated, along with its extended application potential in terms of the ground moving target imaging.