Removing fluid lensing effects from spatial images

TL;DR

This paper introduces a machine learning model that effectively removes fluid lensing distortions from underwater images, enhancing remote sensing of aquatic ecosystems.

Contribution

A novel machine learning approach is proposed to correct fluid lensing effects, improving the clarity of images of underwater ecosystems for remote sensing applications.

Findings

Significant reduction of lensing distortions in test images

Improved image stability and clarity demonstrated

Potential for enhanced remote sensing of aquatic environments

Abstract

Shallow water and coastal aquatic ecosystems such as coral reefs and seagrass meadows play a critical role in regulating and understanding Earth's changing climate and biodiversity. They also play an important role in protecting towns and cities from erosion and storm surges. Yet technology used for remote sensing (drones, UAVs, satellites) cannot produce detailed images of these ecosystems. Fluid lensing effects, the distortions caused by surface waves and light on underwater objects, are what makes the remote sensing of these ecosystems a very challenging task. Using machine learning, a proof of concept model was developed that is able to remove most of these effects and produce a clearer more stable image.