Smart Navigation Aid for Visually Impaired Person using a Deep Learning Model

People who are visually impaired face difficulties while walking in an unknown street. This paper suggests a Smart Navigation Aid for Visually Impaired for the ease of walking in unfamiliar conditions. This system is built to help a person get a sense of his surrounding objects and make decisions according to the alerts given by the system. The main motive is to guide navigation through sidewalk, to help in crossing roads by understanding the traffic lights along with pedestrian crossing signals, detecting potholes and alert the user about the surrounding physical environment. The system can be used by attaching it to a cane or can be worn around the waist with a belt. The set-up is built such that it does not cause any hindrance in walking.