CMC-RMA: Compressed Multi-Coset Range Migration Algorithm for MIMO-SAR

Anstract- This paper introduces the Compressed Multi-Coset Range Migration Algorithm (CMC-RMA) for near-field mi-crowave imaging. CMC-RMA reduces the SAR acquisition time by compressed scanning, wherein a MIMO radar captures measurements by intermittently skipping blocks during SAR acquisition. The reconstruction of the microwave image typically involves utilizing the 2D Fourier Transform (FT) of the measurements. However, this process is impeded by the compressed scanning mechanism. Thus, to estimate the 2D FT, we employ a multi-coset framework making use of intermittent scanning and formulate a compressed sensing-based framework by leveraging block-sparsity constraints. Classical RMA is then employed on the estimated 2D FT to reconstruct the microwave image of the scene. Subsequently, a denoising convolutional neural network (DnCNN) is used to enhance and denoise this reconstructed microwave image to obtain the high-resolution image of the scene. Comparative analysis with existing methods using real experimental data reveals a significant 4dB increase in PSNR for identical compression factors, confirming the effectiveness of the proposed approach.