Underwater environments present unique challenges for imaging due to factors such as light attenuation, scattering, and colour distortion. This research combines advanced CNN models like CBAM(convolutional Block Attention Mod-ule) and VGG16 with state-of-the-art object detection methods of CNN like YOLO or RCNN to enhance the visual quality of underwater images and to detect the objects based on an accuracy rate. Leveraging the various capabilities of the VGG16 model, pretrained on extensive datasets, the system efficiently restores degraded underwater images by capturing and learning intricate features. Integrating the CBAM model enhances this process by selectively attending to salient features while suppressing irrelevant ones, thereby refining the restoration results. Additionally, the combined architecture facilitates object detection within the restored images, enabling the identification and localization of submerged objects with high accuracy. Currently the work presents short review on the existing methods of underwater image restoration and a suggests method employing the CBAM(convolutional Block Attention Mod-ule) and VGG16 to overcome the prevailing challenges in underwater object detection. In future, the research aims to present a website that would be more useful for the students , researchers and the underwater explorers.

Convolutional Neural Networks (CNNs) VGG 16 Underwater Object Detection Image Restoration