Researchers have developed a novel framework to identify "dark vessels" that disable their transponders for surveillance evasion. The system combines a multi-task deep learning model to predict vessel location, type, and dimensions with a k-nearest neighbors algorithm for gross tonnage estimation. This approach leverages heterogeneous image and tabular datasets, addressing the lack of public SAR datasets for end-to-end training. The published code aims to improve maritime security by enabling the identification of vessels that should be equipped with transponders but are not. AI
IMPACT Enhances maritime security by enabling the identification of vessels attempting to evade surveillance through transponder disabling.
RANK_REASON The cluster describes a research paper detailing a new method for vessel detection and gross tonnage estimation using AI.
Read on Hugging Face Daily Papers →
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- deep learning
- GitHub
- gross tonnage
- Hugging Face
- k-nearest neighbors algorithm
- synthetic aperture radar
- Georgia Tech
- PaltrinieriDavide
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →