EBICS: Even Blind I Can See, A Computer Vision Based Guidance System For Visually Impaired

Abstract

Humans have five senses that help them to interact with the environment. Even losing one makes this interaction difficult. There are more than 63 million visually impaired people in India. Visually impaired people can range from being partially impaired to completely blind. These individuals face numerous difficulties in performing daily life tasks. Fortunately, assistive technologies have developed in the last few decades. This technology aims to enable people to live a healthy and independent life with equal opportunities in society. This paper describes the visual assistive system that ensures people live their lives to the fullest. This system helps people interact with the immediate environment in which they are living while taking care of them and keeping them safe from the dangerous elements in the surrounding. The proposed system uses a Raspberry Pi camera and Raspberry Pi 4 embedded system to interact with the environment and the user. We propose to aid the visually impaired community by employing a system that allows us to identify different objects and their inter-spacing distance between the object and the individual whilst simultaneously providing audio feedback. This is achieved by capturing the real-time frames and processing these images. The proposed system boasts an accuracy of 93.7%, thus outperforming similar products while being cheaper and using fewer resources simultaneously. Unlike other devices in the market. This result was achieved with no additional sensors helping reduce the cost of the proposed system significantly and making it affordable for the general public. It also reduces the point of failure as there is no additional hardware sensors other than a camera being used.