Optical lens modeling and optimization with machine learning algorithm for underwater imaging

Underwater photogrammetry is a technique that can be used in the fields like biology, antiquities, and manufacturing with consumer-grade tools. However, because underwater photogrammetry includes unique optical phenomena, underwater camera operations diverge significantly from those on land. We created a unique underwater camera with lens based on the needs and specifications of the marine vessel. The camera can be used in water to capture the image. With conventional cameras, we had to comprehend and adjust the optical prior in order to expand underwater image restoration and enhancement. The traditional approach uses image transmission theory, which calls for mathematical computations to be made beforehand. This technique is very challenging and requires more time for estimation. Prior to calculating and evaluating the lens of the capturing device from which particular scenes were recorded, it is crucial to become familiar with MATLAB functions for prior estimation. The distance among camera and items are kept to a least, so we need a camera to accomplish this. Our suggested approach intended to get around these obstacles in order to create a high-resolution underwater imaging camera. We created a powerful extended depth of field camera on CodeV to model an optical lens that would produce excellent underwater imagery. We used a machine learning-based deblurring and enhancement algorithm as a form of extended optimization for the creation of optical lens. This novel optical development that can be used in water is anticipated to significantly improve the quality and structure by more than 68% and 56%, respectively. © The Author(s), under exclusive licence to The Optical Society of India 2023.