Deep Learning Based Underwater Image Detection and Processing for Oceanography

Le résumé fourni par la source

Recently underwater exploration is the growing research area. It plays a significant role in marine exploration and environmental monitoring for the sustainable development. This research interest results in rapid progress in the field of underwater image processing and analysis. However, images captured in underwater environment significantly differs from the terrestrial imaging. They face substantial challenges and distortions including illumination, noise patterns, disturbed contrasts and several other optical degradations. These challenges specially the optical degradations are mainly caused by the diffusion, absorption and scattering of light in water. The precision and effectiveness of underwater autonomous operations based on these images are greatly impacted by these challenges. Enhancement of these images can substantially improve their quality for human recognition and autonomous operations like object detection algorithms. This paper provides a cascaded framework based on integrated deep learning and image enhancement models to restore the underwater image quality and improve the detection accuracy compared with the conventional methods. Experimental results with the improved and enhanced images and an overall increase in object detection precision validated the accuracy of the proposed system.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.