Smart Ocean Cleanup: An AI-Integrated Autonomous System for Marine Waste Management

Marine pollution is a critical environmental concern, with plastic waste, oil spills, and industrial pollutants contaminating ocean surfaces. Conventional cleanup methods such as manual collection and mechanical skimmers are inefficient, labor-intensive, and unable to adapt in real time. To overcome these limitations, this research presents Smart Ocean Cleanup: An AI Integrated Autonomous System for Marine Waste Management, a novel framework using IoT enabled sensors, deep learning based pollutant classification, and autonomous robotic cleaning mechanisms for efficient waste removal. The system employs EfficientNetB7, DenseNet, YOLOv8, and CNN LSTM for high precision pollutant detection, classification, and real time monitoring. EfficientNetB7 classifies waste types like plastics and oil spills from captured images, while DenseNet enhances classification accuracy. YOLOv8 detects floating waste and guides the robotic cleaner for effective waste collection. CNN LSTM predicts pollution spikes by analyzing historical sensor data, allowing proactive intervention. IoT sensors including MQ135, Turbidity, pH, DS18B20, GPS, and Ultrasonic sensors continuously monitor ocean conditions and transmit data via LoRaWAN/MQTT to a cloud dashboard for live analysis. The system uses an Arduino microcontroller, with AI inference performed on Jetson Nano for edge processing and AWS SageMaker for cloud based analytics. An AI powered robotic boat fitted with floating nets, oil absorption pads, and a microplastic filtration unit autonomously navigates polluted areas for effective cleanup. Performance evaluation shows EfficientNetB7 achieves 96.11 % accuracy, YOLOv8 95.99 %, and CNN LSTM 94.85 % demonstrating the system's high effectiveness in pollutant detection, classification, and predictive analysis.