Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/dfki-ric/uxo-dataset2024/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

The UXO Dataset 2024 is a comprehensive multimodal dataset for acoustic and optical sensing of unexploded ordnance (UXO) underwater. It provides synchronized data from an ARIS 3000 imaging sonar and a GoPro Hero 8 camera, recorded in a controlled experimental environment. UXO Dataset Montage

​
The UXO Problem

During the 20th century, millions of tons of munition were dumped into the oceans worldwide. After decades of decay, the problems these unexploded ordnance (UXO) are causing are starting to become apparent.
In order to facilitate more efficient salvage efforts through autonomous underwater vehicles (AUVs), access to representative data is paramount. However, until now, such data has not been publicly available.

​
Dataset Purpose

This dataset is designed to:

Enable Research

Support research in acoustic and optical perception of UXO underwater

Facilitate AUV Development

Provide data for developing autonomous salvage systems

Provide Ground Truth

Include known and accurate transforms between sensors and targets

Open Access

Publicly available data for the research community

​
Recording Setup

The dataset was recorded using a custom-designed gantry crane system in a controlled underwater environment. This setup allowed for precise, repeatable trajectories while maintaining accurate position tracking. Recording Setup

​
Equipment

ARIS Explorer 3000 Imaging Sonar
  • High-resolution acoustic imaging
  • Precise frame-by-frame sonar data
  • Accurate metadata for each scan
GoPro Hero 8
  • Ultra HD (5.3k, 5312x2988) video recording
  • Synchronized with sonar frames
  • Multiple resolution options available
Custom Gantry Crane
  • Controlled trajectory execution
  • Precise position tracking via ROS
  • Repeatable motion patterns

​
Key Statistics

Trajectories

Close to 100 recorded trajectories

Frames

Over 74,000 frames captured

UXO Types

3 distinct types of UXO

Data Types

Multimodal: sonar, optical, pose, 3D models

​
Access

The dataset is publicly available and can be accessed at:
This dataset is intended for research purposes and was recorded in a controlled experimental environment.

Build docs developers (and LLMs) love

Get started for freeTalk to us