Resolution-enhanced radar/SAR imaging: an experiment design framework combined with neural network-adapted variational analysis regularization

Abstract

The convex optimization-based descriptive experiment design regularization (DEDR) method is aggregated with the neural network (NN)-adapted variational analysis (VA) approach for adaptive high-resolution sensing into a unified DEDR -VA-NN framework that puts in a single optimization frame high-resolution radar/SAR image formation in uncertain operational scenarios, adaptive despeckling and dynamic scene image enhancement for a variety of sensing modes. The DEDR -VA-NN method outperforms the existing adaptive radar imaging techniques both in resolution and convergence rate. The simulation examples are incorporated to illustrate the efficiency of the proposed DEDR-VA-related imaging techniques.

High-resolution recovery of spatial reflectivity maps in Harsh remote sensing scenarios: A metrically structured experiment design regularization approach

“…The problem model that we treat in this paper is structurally similar to the previous studies [8][9][10][11][12], thus the general phenomenological background and some numerical model details are repeated for convenience to the reader in subsections II.A and II.B. Next, in subsection II.C we specify a new aggregated − 2 1 metric structure in the image/solution space that composes the VA-structured base for the development of our new RED-VA image recovery framework.…”

“…Instead of the 2 only structured POCS developed in [9,10] or 1 only structured POCS addressed in [7,12], here we specify the balanced …”

“…The figure captions specify each particular simulated imaging method. Figure 2 reports the convergence rates evaluated in the commonly used resolution enhancement over noise suppression (RENS) metric [8,9] RENS [dB] = 10log 10 …”

“…The Bayesian minimum risk-based strategies [1,8] are infeasible due to the statistical model-level uncertainties. On the other hand, the celebrated descriptive experiment design regularization (DEDR) framework [8,10] is not properly adapted to tackle with the nontrivial prior information about the structural local statistics properties of the RS images. The recently developed compressed sensing (CS) approaches for sensor image enhancement [7,12,13] are also not particularly adapted for the considered here radar/SAR imaging in harsh environments.…”

“…The resolution enhancing recovery of the fine spatial spectrum pattern (SSP) of the scene reflectivity maps from such initial low resolution MSF imagery in harsh operational scenarios is a nonlinear ill-conditioned inverse problem due to the mentioned above operational scenario uncertainties [2][3][4]. A variety of approaches has been recently proposed to tackle with such class of problems [1,[3][4][5][6][7][8][9][10][11][12]. The Bayesian minimum risk-based strategies [1,8] are infeasible due to the statistical model-level uncertainties.…”