An AI-Based Coastal Surveillance and Unauthorized Vessel Detection System Using IoT Sensors and Machine Learning

Small-scale fishermen operating near maritime borders often face safety risks due to navigation errors, harsh environmental conditions, and delayed emergency response. Traditional marine safety systems rely heavily on manual observation and basic alert mechanisms, which are often insufficient in real-time critical situations. This paper proposes an AI-assisted marine safety and border alert system that integrates sensor data with machine learning techniques to support timely decision-making and risk mitigation. The system processes inputs from salinity sensors, water-level sensors, and positional data to detect anomalous conditions such as potential border crossing, vessel instability, and emergency scenarios. A supervised machine learning model is employed to classify risk levels based on historical and simulated sensor data. Experimental evaluation using simulated maritime datasets demonstrates that the proposed approach improves early warning accuracy and reduces false alerts compared to static threshold-based systems. The proposed framework is intended as a humanitarian decision-support tool and can be extended for real-world deployment with minimal infrastructure.