Sparse Logistic Regression-Based One-Bit SAR Imaging

One-bit synthetic aperture radar (SAR) imaging has garnered significant interest due to its ability to lower the cost of storing enormous amounts of data during sampling and transmission, as well as the expense of analog-to-digital converters (ADCs). However, existing one-bit SAR imaging methods suffer from high computational complexity and artifacts in the resulting images. To address these problems, the sparse logistic regression (SLR) model solved by iterative hard thresholding (IHT) is applied to one-bit SAR imaging, and a new SLR-IHT imaging method is proposed. The SLR-IHT method models the one-bit SAR imaging problem as an SLR task and optimizes the solution using the IHT framework. By leveraging the joint sparsity of the real and imaginary components, the proposed method enhances imaging quality while effectively suppressing artifacts. To accelerate computation, the Armijo step size criterion is employed to adjust the step size and support set during the iterative procedure. Moreover, a theoretical investigation into the convergence properties of the proposed method was conducted. Extensive simulations and real data experiments are conducted to evaluate the performance of the SLR-IHT method. The results demonstrate its superiority over existing one-bit SAR imaging techniques in terms of imaging quality and computational efficiency.