IJSRP, Volume 7, Issue 11, November 2017 Edition [ISSN 2250-3153]
Mallawarachchi Yashas, Maalik K.S., Ashfaq M.Z.M., De Silva K.P.S.H., Ragulan.S.
Abstract:
"The Blind Guider" application is a real time application for detecting obstacles such as staircases, potholes, pedestrian crossings and moving vehicle detection and getting current location by GPS and giving directions to travel from place to place using the mobile phone camera with voice instructions. This Android application will warn of the obstacles in the road to the user. The system uses Haar-Cascade, BLOB and SURF algorithms for object detection. In order to identify key matching objects and for voice instructions the system uses a machine learning algorithm. Blind people use a white cane to navigate through roads and even indoors. The developed system is much more accurate than the white cane because the system will identify obstacles for the user within a certain distance. This application will provide assistance to the user when required. The mobile camera input is sufficient to commence the processing, the system will assist the user to navigate through obstacles in indoors and outdoors.