Underwater Imaging

Underwater imaging research focuses on overcoming the challenges of light attenuation, scattering, and biofouling to capture high-quality images and videos for various applications, including marine biology, infrastructure inspection, and autonomous navigation. Current research emphasizes the use of deep learning models, particularly Vision Transformers and diffusion models, to enhance image quality through dehazing, color correction, and artifact removal, often incorporating physical models of light propagation in water. These advancements are improving the accuracy of underwater visual odometry, enabling more efficient seafloor mapping, and facilitating improved target recognition in sonar imagery, ultimately leading to a more comprehensive understanding of underwater environments and improved capabilities for underwater robotics.