JATAYU
From seeing to responding.
An autonomous UAV project connecting human detection, tracking and stereo-vision navigation with hardware and control logic.
The problem
Perception is one part of a larger system. A detection needs to connect to depth information and a control pathway before it becomes useful to an autonomous vehicle.
How it was built
- 01Use Python, OpenCV and NanoDet for the perception pipeline.
- 02Connect stereo-vision depth estimation to navigation.
- 03Integrate MAVLink communication, Flask and Raspberry Pi hardware.
The takeaway
Applied AI needs an engineering context: what the model sees, what the system can infer, and how that information reaches a useful action.