Blind Smart- Smart Mobile App to Find Crosswalks for Visually Impaired

Abstract

A new framework for crossover detection and localization is proposed. Proposed model extracts satellite pictures captured from Google Maps. To find crosswalks we use popular machine learning algorithm which uses images obtained from google application. Smart phone with GPS hold by visually impaired person is used to initialize the extraction of pictures from Google Maps, also to assist its user by providing audio feedback of the closest detected crossover. Proposed model provides better results for low resolution pictures also.Proposed model also suggests shortest path to reach destination with minimum number of cross walks.