A CNN Based Underwater Aquatic Exploring System for Real Time Identification of Fish Species

The application of machine learning models is increasing rapidly, particularly solving challenges within underwater environments. This research paper introduces a machine learning model designed for the detection of underwater species and garbage waste in underwater environments.Existing systems are lacking due to the inadequate dataset real time classification and misclassifications of images. These systems typically identify only a small number of species because of the difficulties associated with varying underwater light conditions. This research proposes a new approach by utilizing the YOLO algorithm for the detection and classification of underwater images. The system functions by capturing underwater images, which are then preprocessed to improve the features. The dataset consists of 600 images, with 7 different classes and splitted into training and testing. The proposed method detects and classifies underwater objects with an overall precision of 95% to 97%. This research helps marine researchers and students in identifying underwater species and helps to improve underwater environments.