Abstract

Improved hardware and processing techniques such as synthetic aperture sonar have led to imaging sonar with centimeter resolution. However, practical limitations and old systems limit the resolution in modern and legacy datasets. This study proposes using single image super resolution based on a conditioned diffusion model to map between images at different resolutions. This approach focuses on upscaling legacy, low-resolution sonar datasets to enable backward compatibility with newer, high-resolution datasets, thus creating a unified dataset for machine learning applications. The study demonstrates improved performance for classifying upscaled images without increasing the probability of false detection. The increased probability of detection was 7% compared to bicubic interpolation, 6% compared to convolutional neural networks, and 2% compared to generative adversarial networks. The study also proposes two sonar specific evaluation metrics based on acoustic physics and utility to automatic target recognition.
Original languageEnglish
Pages (from-to)509-518
Number of pages10
JournalJournal of the Acoustical Society of America (JASA)
Volume157
Issue number1
DOIs
Publication statusPublished - 23 Jan 2025

Data Availability Statement

The data used in this study belong to NATO STO CMRE. They were made available for this study with the permission of NATO STO CMRE. However, they cannot be made publicly available.

Funding

This work was performed under Project No. SAC000E03 of the STO-CMRE Programme of Work, funded by the NATO Allied Command Transformation and the UK Research and Innovation body (UKRI EP/S023437/1).

FundersFunder number
UK Research & InnovationEP/S023437/1

Fingerprint

Dive into the research topics of 'A diffusion-based super resolution model for enhancing sonar images'. Together they form a unique fingerprint.

Cite this