AGT-CV: An Aerial-Ground Team Cross-View Dataset for Heterogeneous Robot Teams in Unstructured Environments

1 min read

AGT-CV is a real air-ground dataset from a Clearpath Husky (3D LiDAR, stereo, IMU, GPS) and an Autel EVO II drone (RGB, thermal, GPS) on forest trails, rocky paths, mud, snow and grass. It has over 13,000 synchronised frames from about 29 minutes of operation in four environments, with SAM 3 zero-shot masks and almost 8,000 hand-labelled images. It was recorded in early spring, so the drone can partly see the ground robot through sparse canopy.

Relevance

Real overlapping air-ground data in rough terrain is scarce, and 29 minutes is small but real. Useful for cross-view work in multi-robot multimodal perception; see aerial swarms and multi-robot systems.