Computer vision guidance system for indoor navigation of visually impaired people

Abstract

Visually Impaired (VI) and blind people suffer from reduced mobility, as they cannot detect the terrain and their environment. They always need assistance and walking support systems in their daily life. Solutions have been proposed many decades ago and are rapidly improving nowadays due to the technology evolution and integration. A large number of assistance aids have been deployed in real life situations whereas other concepts remained as research ideas. This paper describes a new approach of an ambient navigation system that would help the visually impaired or blind person to move freely indoor (house, office, etc.) without the assistance of anyone. The system is composed of IP cameras attached to the ceiling of each room and the smart phone of the subject is used as human machine interface (HMI). Frames are sent to a computer that analyzes the environment, detects and recognizes objects. A computer vision guidance algorithm is designed to help the user reach his destination (or his personal item) with obstacle detection. The system is commanded by voice messages via a simple mobile application. Feedbacks (alerts, route) are voice messages returns by the application to the user. This system provides a reliable solution to assist those users in their indoor navigation providing them a correct route with obstacle avoidance.