Abstract:Synthetic aperture sonar (SAS) requires precise time-of-flight measurements of the transmitted/received waveform to produce well-focused imagery. It is not uncommon for errors in these measurements to be present resulting in image defocusing. To overcome this, an \emph{autofocus} algorithm is employed as a post-processing step after image reconstruction to improve image focus. A particular class of these algorithms can be framed as a sharpness/contrast metric-based optimization. To improve convergence, a hand-crafted weighting function to remove "bad" areas of the image is sometimes applied to the image-under-test before the optimization procedure. Additionally, dozens of iterations are necessary for convergence which is a large compute burden for low size, weight, and power (SWaP) systems. We propose a deep learning technique to overcome these limitations and implicitly learn the weighting function in a data-driven manner. Our proposed method, which we call Deep Autofocus, uses features from the single-look-complex (SLC) to estimate the phase correction which is applied in $k$-space. Furthermore, we train our algorithm on batches of training imagery so that during deployment, only a single iteration of our method is sufficient to autofocus. We show results demonstrating the robustness of our technique by comparing our results to four commonly used image sharpness metrics. Our results demonstrate Deep Autofocus can produce imagery perceptually better than common iterative techniques but at a lower computational cost. We conclude that Deep Autofocus can provide a more favorable cost-quality trade-off than alternatives with significant potential of future research.

AF-DCDU-Net: An Approach for ISAR Autofocus Imaging via Deep Convolution and Deep Unfolding Net

AF-AMPNet: A Deep Learning Approach for Sparse Aperture ISAR Imaging and Autofocusing.

3-D SAR Autofocusing with Learned Sparsity

Robust and Efficient ISAR Autofocusing Based on Deep Convolution Network.

AFnet and PAFnet: Fast and Accurate SAR Autofocus Based on Deep Learning

STLS-LADMM-Net: A Deep Network for SAR Autofocus Imaging

Deep Learning Approach for Sparse Aperture ISAR Imaging and Autofocusing Based on Complex-Valued ADMM-Net

High-Resolution ISAR Imaging and Autofocusing Via 2D-Admm-net

ISAR Imaging of Target Exhibiting Micro-Motion With Sparse Aperture via Model-Driven Deep Network

SAF-3DNet: Unsupervised AMP-Inspired Network for 3-D MMW SAR Imaging and Autofocusing.

Autofocus-Coupled UTAMP for Sparse Aperture ISAR Imaging

SAR Image Autofocusing Based on Res-Unet

An Autofocus-Coupled UAMP Method for InISAR Imaging

Deep Autofocus for Synthetic Aperture Sonar

Compressed Sensing Linear Array SAR Autofocusing Imaging via Semi-definite Programming

SAR Image Reconstruction and Autofocus Using Complex-Valued Feature Prior and Deep Network Implementation.

Deep Learning-Based Enhanced ISAR-RID Imaging Method

Single-Layer Network Realizes Sparse Aperture ISAR Imaging

Real-Time, Deep Synthetic Aperture Sonar (SAS) Autofocus

Computationally Efficient Sparse Aperture ISAR Autofocusing and Imaging Based on Fast ADMM.

High-Resolution ISAR Imaging by Exploiting Sparse Apertures