Drone Camera Action builds on-device AI that spots wildlife in live drone imagery — RGB and thermal — lets an operator confirm each sighting in one tap, and turns every flight into training-ready data.
Get in touch See the platformDrone, phone, and cloud working as one system — built for the field, where connectivity can't be assumed.
Raw RGB or thermal frames stream from the drone over the local link and render in real time.
Dual-model YOLOv8 inference runs entirely on the device and switches automatically with the active camera — no signal required.
The operator is notified the moment the model flags wildlife, with a bounding box over the feed.
A person in the loop confirms the species — capturing the frame and saving a trusted sighting.
Confirmed sightings upload to the backend as structured, geo-tagged records that feed the model training pipeline.
Every confirmed sighting becomes labelled data — so detection gets sharper over time.
Five steps, field to cloud.
The drone controller and the phone discover each other over Wi-Fi; the live feed starts automatically.
Operators fly the survey while the app renders the camera feed in real time.
On-device models flag wildlife live and draw detections over the feed.
One tap confirms the species and saves the sighting locally — even with no connectivity.
Back in range, sightings upload to the cloud as training-ready records.
The detection models are designed around the species that matter most to reserves and conservancies — each with its own on-feed marker.
Drone Camera Action (Pty) Ltd is a South African technology company developing AI-integrated drone software for wildlife monitoring and conservation. We pair drones, on-device machine learning, and a cloud training pipeline into one practical system for the people doing the work in the field.
Interested in a pilot, a partnership, or a demo? We'd love to hear from you.