We devised an objection detection system using image processing and deep learning which is to be implanted on walking sticks for the visually disabled. It detects the objects in front of it, and also how far they are from the person. This information was fed back to the person in form of audio for them to be able to perceive it. We also created a stabilsing system which helped our sensors and cameras remain still and stable while the walking stick changed its angles.
Just after about 24 hours of building this project we realised that we had multiple glitches in running constant video feedback through our algorithms. We solved this problem by using better deep learning algorithms and were also able to add facial recognition.
Discussion