Virtual Booth

This project presents an innovative Smart Sea-Garbage Vessel integrated with Artificial Intelligence (AI) for autonomous marine debris detection and collection in coastal and freshwater environments. The system combines a YOLO-based Convolutional Neural Network (CNN), ResNet-50 architecture, embedded hardware, and real-time computer vision for intelligent waste detection and navigation. Hyperparameter optimisation using a Genetic Algorithm (GA) improved classification accuracy from 0.866 to 0.902, with a cross-entropy loss of 0.406. Implemented on a Raspberry Pi 5 platform, the vessel successfully detected and classified floating debris under varying environmental conditions. This innovation offers significant potential for sustainable marine cleaning and intelligent environmental monitoring.