The document presents a proposed system for classifying underwater objects using sonar signals. The objectives are to classify identified underwater objects, implement a framework for mine detection, and use an AI-based approach. The proposed model would use a dataset of sonar signals from mines and common objects to train a machine learning model. Data preprocessing would be performed to handle missing/noisy data. The trained model would then classify new sonar signals as either mines or common objects. Background research identified gaps in previous work such as reliance on image processing for classification which provides low accuracy. The proposed system aims to classify objects based on sonar signals to safely detect mines and protect submarines/ships.