CUAS Training Data
Counter-UAS systems are classifiers. Their field performance depends on whether they've seen the adversary's current platform, in the current environment, at the current range, under the current countermeasures. We build the datasets that close that gap.
Collection Dimensions
- Target class: Group 1, 2, and 3 UAS captured across the classes your program cares about
- Flight profile: trajectory (looming, crossing, receding, diving), velocity, formation
- Environment: 9+ environment types — urban, rural, open field, forested, littoral, and more
- Lighting & obscurants: daylight, dusk, dawn, low-light, haze, precipitation
- Pixels-on-target: across the range of detection distances your deployment expects
- Scene activity: clean background through cluttered, multi-object scenes
- Horizon interaction: sky, mid-horizon, terrain-background target framing
Typical Use Cases
- Develop and benchmark counter-drone detection algorithms
- Train and validate computer vision models
- Support airspace security research and development
- Stress-test existing algorithms and CUAS systems
- Expand training datasets
- Accelerate Counter-UAS product development
- Protect U.S. warfighters