Efficient Online Sparse Target Reconstruction for Stochastic Radiation Radar Based on Progressive Staring Sparse Pruning Approach

Stochastic radiation radar (SRR) utilizes stochastic electromagnetic wave signals to form a radiation field randomly distributed in space, offering advantages such as non-reliance on motion for high resolution imaging. However, existing SRRs typically acquire batched data through long-duration staring observations of the target, followed by target reconstruction to obtain high-resolution imaging results, leading to low imaging efficiency. Depending on the data acquisition mode of the stochastic radiation radar, this paper proposes a Progressive Staring Sparse Pruning (PSSP) Approach. On the one hand, a staring sparse pruning framework is proposed to construct an active set of target grids, which utilizes the structural feasibility gate and incremental significance characteristic of the closed-form solution. On the other hand, an adaptive progressive termination criterion is established to reduce computational complexity, which adopts an adaptive relative-coefficient criterion and a progressive active-set-size criterion to avoid focusing only on a few repeatedly active grids and to reduce unnecessary echo-sample processing. Simulation results demonstrate that the proposed method fully exploits the advantages of stochastic radiation radar staring observation, substantially improving computational efficiency while guaranteeing imaging performance.