Implementació d'una aplicació mòbil per facilitar el reconeixement de semàfors als vianants invidents

Abstract

[en] Despite the efforts that have been made in recent decades, blind and visually impaired people still face barriers in large cities which impede their ability to move independently around the streets. In terms of accessibility of urban space, one common problem they face is traffic signal controlled pedestrian crossings. While many of these have an audible signal, in large cities there are many that do not incorporate these devices or they are broken, leaving the blind and visually impaired with no indication of when it is safe to cross. This problem is more pronounced in small cities where the use of the audible signal is not so common and as a result there are fewer crossings accessible to blind and visually impaired people. In recent years there has been a huge increase in the amount of smartphones or intelligent phones per inhabitant. It is hard to meet someone who does not own one and use it for leisure, communication or productivity purposes. My objective is to take advantage of the availability of these devices in order to solve the issue of pedestrian crossing accessibility faced by blind and visually impaired people. This project will analyse in detail the problems faced by these users when they are confronted by a traffic signal controlled pedestrian crossing without audible signal. An application for android devices will be developed which is both accessible and tailored to their needs. Using improvements in Deep Learning, the app can detect objects in real time allowing it to recognize the traffic signals at pedestrian crossings and indicating their state to the user.