An Effective Smart Goggle to Address the Problem Faced by the Visually Impaired People

Abstract

According to our research and the studies, there are around 4.95 million people who are blind and at the same time70 million people who are partially visually impaired in India. Without vision, it becomes difficult for the blind people to find their way without encountering obstacles. Avoiding obstacles can be uncomfortable, clumsy, and inaccurate, even with aids such as walking sticks. People with vision or vision problems have to deal with multiple issues in their daily lives, as modern assistive devices often do not meet consumer needs in terms of price and level of support. Our project presents a new design of assistive glasses for the visually impaired person. It is intended to assist in multiple daily tasks utilizing a portable design format. The overall build cost is kept low by using a Raspberry Pi 4 single board computer for processing and a camera for image acquisition. Our proposed glasses use Machine Learning (ML) to extract information from images and verbally communicate what the wearer is seeing through the audio output. As a result, ML-enabled glasses enable blind and partially sighted users to find personal items at home, read documents at workplace, take public transportation without any overhead, recognize friends, and many more such applications, can make them more independent, freer, and more aware of their surroundings. We made use of YOLOv3 algorithm for object detection and got the accuracy of 94.3% and which was able to detect majority of the objects in the surroundings of the user.