Multiframe resolution recovery of radar imagery: Towards super-resolution sensing

The aim of this study is to address a new approach and develop the relevant technique for super-resolution (SR) feature-enhanced recovery of microwave remote sensing (RS) imagery. The challenging proposition is twofold. First, we adapt the SR multi-scale iterative reconstructive (MSIR) image post-processing method for solving the inverse problem of recovery of the speckle corrupted low resolution RS images employing iterative projections onto the nested refined resolution frames. Second, we unify the modified RS-adapted MSIR method with the Descriptive Experiment Design Regularization (DEDR) high-resolution RS image enhancement technique for attaining the overall SR recovery with considerably enhanced resolution performances. Algorithmically, the MSIR processing loop is performed via the Fourier transform (FT) or wavelet transform (WT) of the input image. Different wavelet dictionaries were examined in order to approach the most speeded-up WT-based iterative MSIR-level image recovery.