Sea-Scan: High-Accuracy, ML-based Dark Vessel Detection and Localisation via Weakly Supervised DAS Monitoring
arXiv:2606.213262.1
Predicted impact top 94% in SD · last 90 daysOriginality Incremental advance
AI Analysis
It addresses the problem of detecting non-cooperative vessels for maritime security, offering a practical solution with high accuracy and low false alarms.
The paper presents an ML-based vessel detection and localization system using weakly supervised DAS monitoring, achieving a 97.8% detection rate at a 1.98% false-trigger rate for dark vessel detection.
We present an ML-based vessel detection and localization system, trained with weak supervision from imperfect AIS labels, that achieves a 97.8% detection rate at 1.98% false-trigger rate, successfully identifies dark-vessel events from unlabeled data.