Navigation And Obstacle Detection For Visually Impaired People

Abstract

In order to provide safety for the visually impaired people to have mobility in their way to get information of obstacle detection and warnings from the surroundings. To achieve it firstly need to get information of obstacle to be detected and localized it then send to visually impaired people using various modalities such as voice. Here we present an aiding system for visually impaired people based on tensor flow object detection model and Google speech's model. The proposal consists of two important components: Information of environment and Analysis representation. Initially it aims at capturing the environment with help of tensor flow object detecting model and analyzing it which is predefined obstacle for visually impaired people, the next component try to provide the obstacle's information under the form of speech to visually impaired people.