3D Visual Grounding Dataset for Multi-Platform Robotics

3EED is a cross-platform, multi-modal dataset for 3D visual grounding, featuring synchronized RGB, LiDAR, and language annotations. It spans vehicle, drone, and quadruped robot platforms. The dataset was created by author RRRong and was last updated in October 2025.

Use Cases

• Train models to ground language queries in 3D space using synchronized RGB and LiDAR data.
• Evaluate visual grounding performance across different robotic platforms like vehicles and drones.
• Develop multi-modal perception systems that link textual descriptions to specific 3D point cloud regions.
• Benchmark cross-platform generalization for 3D scene understanding tasks.

Strengths

• Multi-modal data with synchronized RGB, LiDAR, and language annotations.
• Cross-platform coverage spanning vehicle, drone, and quadruped robots.

Limitations

• Specific row count, sample size, and dataset scale are unknown.
• Geographic and temporal coverage details are not provided.

Provenance

Source

RRRong via HuggingFace.

Freshness

Last updated October 2025.