With regard to inverse synthetic aperture radar (ISAR) imaging, traditional range-Doppler (RD) algorithm is inapplicable to sparse aperture. Although compressive sensing (CS) algorithm can overcome this problem, the imaging resolution is not high enough. When deep learning (DL) is applied to ISAR imaging, some problems may also occur, such as network complexity, the selection of labels in network training, the influence of noise and sparse aperture, and the loss of weak scattering points in network test. In order to solve the above problems, a joint fast iterative shrinkage-thresholding algorithm (FISTA) and Visual Geometry Group Network (VGGNet) high-resolution imaging method is proposed in this paper. In the proposed method, FISTA is presented to reduce the impact of noise and sparse aperture. The high-resolution processing network (HRPN) is built based on VGGNet. Then, combined with peak extraction technology, the random ideal scattering points are utilized to construct the training/validation set. Meanwhile, the training process of HRPN is analyzed theoretically, and the network test strategy is designed by the differences between the test set and the training/validation set. Extensive experiments based on both simulated and measured data demonstrate that the proposed method has good imaging performance and small network complexity.

A signal time‐frequency results inversion method is proposed for extracting micro‐motion features of rotor targets under low signal‐to‐noise ratios (SNRs). In the case of low SNRs, the echo's energy of rotor targets is mainly concentrated in the flash's echo component. Conventional micro‐motion feature extraction of rotor targets primarily utilises the sinusoidal modulation feature in time‐frequency results, whose energy is much lower than the flash. Under low SNRs, the sinusoidal modulation in the echo's time‐frequency results will be submerged in the noise, making feature extraction challenging. A deep learning network is used to inverse the time‐frequency results containing sinusoidal modulation based on the flash's features in the time‐frequency results. Based on the inversion time‐frequency results, the GS‐IRadon algorithm is used to extract micro‐motion features, which can significantly reduce the times of IRadon transformations and improve feature extraction speed and accuracy. Through simulation and analysis, a novel method using a deep learning network like UNet can effectively inverse time‐frequency results under low SNRs, providing a new technical approach for micro‐motion feature extraction. Time‐frequency results inversion is a novelty method used to achieve micro‐motion feature extraction of rotor targets.

For inverse synthetic aperture radar (ISAR) imaging of targets with micromotion parts, the removal of micro-Doppler (m-D) effects is the key procedure. However, under the condition of a sparse aperture, the echo pulse is limited or incomplete, giving rise to the difficulty of eliminating m-D effects. Thus, a novel m-D effects removal algorithm is proposed, which can effectively eliminate m-D effects, as well as the interference introduced by noise and sparse aperture in the ISAR image of the main body. The proposed algorithm mainly includes two processing steps. First, combined with the cut-off point determined by the normalized singular value difference spectrum, the rough estimation of the main body is achieved by singular value decomposition (SVD). Then, the variational mode extraction (VME) is extended to complex variational mode extraction (CVME). The constrained variational problem constructed by bandwidth and spectral overlap constraints is solved by the alternating direction method of multipliers (ADMM), and the precise estimation of the main body is obtained. Experimental results based on both simulated and measured data demonstrate that the proposed algorithm can acquire the high-resolution ISAR image of the main body under noisy and sparse conditions.

The micro-motion demonstrates a detailed description of the target’s motion, modulating its radar signal in terms of echo, phase and imaging. In this article, we propose a micro-motion feature extraction method based on multi-domain parameter estimation and phase compensation. This method analyzes the periodic peak characteristics of rotor targets in both time and frequency domains. Then, by designing a sliding window peak search algorithm, multiple peaks are extracted, and these are used to estimate the time-domain peak frequency and frequency-domain spectral interval. By comparing the peak frequency in the time domain and the spectral interval in the frequency domain, the odd and even numbers of the rotor target blades are determined, and the rough parameter estimation of the blade number and rotational speed of rotor targets is obtained. Using the estimated parameters to construct compensation factors for phase compensation of the signal, the rotor target blade number is determined through the compensated signal, achieving a more accurate estimation of the blade number and rotational speed and further improving the feature extraction accuracy. The simulation results demonstrate that the method is capable of swiftly and efficiently extracting the features of propeller aircraft and helicopters and has good performance in the case of low SNR or clutter. The effectiveness of the method has also been verified through measurement data.

With regard to inverse synthetic aperture radar imaging with limited bandwidth and sparse aperture, it is a challenge to traditional range-Doppler (RD) algorithm. We proposed an innovative two-dimensional (2D) joint sparse imaging algorithm, namely, 2D fast orthogonal matching pursuit (2D-FOMP) algorithm. In the proposed algorithm, one-dimensional OMP (1D-OMP) is extended to 2D-OMP in the complex domain from three aspects of atom recognition, projection update, and residual update. Then, the equivalence between 1D-OMP and 2D-OMP is analyzed theoretically. Meanwhile, two strategies that multi atom recognition and matrix recursive update are added in 2D-OMP to further improve the reconstruction speed of 2D-FOMP. Experimental results based on both simulated and measured data demonstrate that the proposed algorithm has good imaging performance under noisy and sparse conditions.