Welcome to the PETS2025 Challenge!

The challenge has ended!

• See Leaderboard
• See Results summary paper : for the details of all submitted methods and full evaluation results. 

Join us in pushing the boundaries of maritime and coastal border surveillance.

PETS2025 brings together cutting-edge technologies and multi-authority collaboration to tackle real-world challenges in detecting, tracking, and geolocating targets across complex maritime environments. It covers critical tasks including:

1. Target Detection and Classification

Target detection and classification in maritime and coastal areas using multi-platform and multispectral sensors

RGB Ground Sensor	Thermal Ground Sensor
Ultra-Violet (UV) Ground Sensor	Short-Wave Infrared (SWIR) Ground Sensor
RGB UAV	Thermal UAV

Images from different sensors, including visible (RGB), thermal, UV, and SWIR ground sensors, as well as visible (RGB) and thermal UAVs, with examples of three object classes: person, vessel, and vehicle.

2. Long-Term Target Tracking 

Tracking multiple targets across diverse terrains using multi-platform and multi-spectral sensors

A sequence of visible ground sensor images showing the tracking of vessels and persons across different terrains, particularly "person (ID: 34)'' from the boat to the shore.

3. Target Geolocation Approximation

Approximating geolocations of specific targets using only imagery and telemetry data.

Example images showing a detected vessel and its geolocations over time, projected onto a map.

This challenge is sponsored by the EU-funded project EURMARS: An advanced platform to improve the EURopean Multi Authority bordeR Security efficiency and cooperation (https://eurmars-project.eu).