Computer vision / edge perception

O.R.B.I.T.

Low-Cost 360° Perception and Ranging

Low-cost perception research combining stereo cameras, YOLO, ranging, bearing, and persistent target tracking.

GitHub

Perception system

O.R.B.I.T. (Optical Ranging, Bearing & Identification Telemetry) is research toward a modular, low-cost 360° perception sensor that can detect, classify, range, bear, track, and export real-world targets in real time.

Every detection leaves the pipeline as a normalized target record: identity, timestamp, class, confidence, bearing, and range. Bench prototypes have progressed from pipeline simulations to a calibrated stereo YOLO rig with persistent tracking and Kalman-smoothed depth.

Ranging is now validated end to end. Next come a single-sensor slice, a 360° prototype, and then productization.