Navigating Alert for Visually Impaired Using Computer Vision Aided System

Abstract

Covid-19 pervaded the entire world very soon after the outbreak, and it is still present and people are still suffering. In a medically advanced world, where technology is futuristic, but the fact that the virus has still caused deaths of over 3 million people all over the world reportedly and over 180 million affected cases. While the governments of all countries have been trying their best to keep the situation in control, and as the virus continues to evolve and keeps evolving, various countries have been engaged in manufacturing vaccines and have even successfully produced them. The only thing left for us to do, is to, at the least, not catch the virus from other people who are already infected. And the way to do that is to practice social distancing. While it is not something normal people can't do, yet in actuality some don't do, for them it becomes more difficult to maintain such social distance when they are outside, on the streets because not everyone is going to practice social distancing, and it is not doable for them to know if someone is standing at a very close distance. The proposed mobile application is going to target this issue and help the visually impaired to maintain social distancing to keep themselves safe in this pandemic. This application will help resolve the issue for visually impaired people by detecting a person on behalf of them, so that they can become alert and do what is required. The main objective of proposed application is to help the visually impaired to maintain social distance without having to rely on other people for help or asking someone to accompany them wherever they go using OpenCV, Yolo v4 for human detection.