DarkVesselNet: Multi-Modal Remote Sensing and Trajectory Reasoning for Dark Vessel Detection

2026-05-29

AI Summary

This study addresses the challenge of detecting illicit “dark vessels”—ships that deliberately disable their Automatic Identification System (AIS) broadcasts—by proposing an end-to-end differentiable detection framework that fuses multi-source remote sensing data with AIS trajectory information. The approach uniquely integrates TGARD-inspired gap-aware detection with a Pi-DPM-inspired anomaly head, incorporating Sentinel-1 SAR imagery, Sentinel-2 optical data, geospatial foundation models, and trajectory inference modules. Key technical components include SAR speckle filtering, optical band ratios, Haversine distance computation, and cross-sensor registration to enable precise identification. Implemented as an open-source Python package and deployed on Hugging Face Spaces, the system has undergone software-level validation, demonstrating the efficacy of its sensor fusion pipeline and anomaly detection methodology.

Abstract

Dark vessel detection requires fusing what vessels report through AIS with what satellites observe through radar and optical sensors. DarkVesselNet is a multi-modal remote sensing stack that combines Sentinel-1 SAR, Sentinel-2 optical imagery, geospatial foundation model backbones, AIS trajectory reasoning, TGARD-style gap detection, and a Pi-DPM-inspired anomaly head. The repository exposes the system as a tested Python package and a public Hugging Face Space. The paper presents the sensor stack, backbone abstraction, fusion path, anomaly head, and current validation. The evidence currently available is software-grounded: tests for SAR speckle filtering, optical band ratios, Haversine distance, TGARD gap emission, sensor coregistration, backbone token shapes, and differentiable anomaly scoring.

Problem

• dark vessel detection
• multi-modal remote sensing
• AIS
• satellite imagery
• anomaly detection

Innovation

• multi-modal fusion
• dark vessel detection
• geospatial foundation model
• trajectory reasoning
• anomaly detection