Authors: N. Patchiraja and A. Solairaja, Assistant Professor, Dept. of ECE, PSN Institute of Technology and Science, Tirunelveli, Tamil Nadu, India
Editors: Dr. Thirumurugan Shanmugam and Dr. Shweta A. Bansal
ISBN: 978-81-963849-7-5
DOI: https://doi.org/10.59646/csebookc17/004
Abstract:
Many of the components that make up the ocean’s environment are unavoidable, and they’re increasingly being recognised as important economic and therapeutic resources. Due to difficulties in access, much of the ocean remains unknown. The problem of access to the deep ocean has been solved by the creation of remotely operated vehicles (ROVs). Thanks to advancements in underwater photography technology, we now have access to a wealth of high-quality still and moving underwater imagery. Due to concerns with the colour cast, lighting issues, blurred features, and lower contrast in the obtained underwater picture, this image data could not be used directly in further research. Due to the ocean’s water molecules scattering light at different wavelengths, a colour cast occurs as one descends further into the ocean. Light pollution from suspended particles in the water is an issue when using artificial lights on AUVs. The aforementioned mentioned problems cause the obtained photos to be of low quality. Without picture enhancement, it’s difficult to get reliable results from analyses like segmentation and classification. The problem of low picture quality has been effectively addressed by Generative Adversarial Networks (GANS). This work provides a foundation for future research into GAN design to create a new GAN specifically for underwater picture enhancement by discussing the different GAN architectures already in use for this purpose.