Smart Glasses for Visually Impaired Using Image Processing Techniques
P. N. Karthikayan, R. Pushpakumar
Abstract
In this world, approximately 300 million are visually impaired, smart glasses helps the people who are Visually Impaired to enhance their life. With the help of this device they can navigate independently, feel convenient in the society, and also feel secure. They have many problems in their life, some of them are they cannot navigate independently, cheated in terms of money, Bus or Train number identification etc. In this system two features are implemented they are Currency Recognition and Traffic Light Detection. A Machine Learning (ML) Model is configured which consists of Convolution Neutral Network (CNN), image processing techniques, and Object detection model. Currency recognition extracts both visible and hided marks on the money for effective classification. Traffic light detection uses an image processing algorithm with standard object recognition methods. This system helps the Visually Impaired to cross the road without other's help. For the first feature i.e. Currency recognition, the output is the value of the currency that person has in his/her hand and for the second feature i.e. Traffic light detection, the output is whether the traffic light is red or green, if it is green then the number of seconds that person has to cross the road.