VisionAid — Assistive Navigation for the Blind
Wearable obstacle detection with on-device object recognition and audio guidance.
92%
Detection accuracy
18 FPS
Inference
A
Grade awarded
The build
A computer-vision FYP that actually works outside the lab. A Raspberry Pi wearable runs a quantised detection model on a live camera feed, fuses it with ultrasonic range data, and speaks obstacle warnings and scene descriptions through bone-conduction audio. Delivered with the working prototype, IEEE thesis and a viva walkthrough — the student scored an A grade.
This project was delivered under our FYP & Academic Projects service. Final year projects built properly — and explained properly.