Electronic System for Signage Detection

The present work focuses on the need to give independence, security and generate a better lifestyle for people with visual disabilities, through their journeys abroad. The work seeks to be an aid so that people with visual disabilities can move better, recognizing useful signage according to standardized colors for prevention, obligation and information, for their safety and performance, such as pedestrian traffic lights, bus stops, pedestrian crossings, in addition, to evaluate the image detection methods using Machine Learning with deep learning techniques through the haar-cascade model to give a better recognition response to the user. The input to the system is a continuous video sequence, which analyzes and provides the user with an audible output by recognizing the different traffic signs at 2 meters. This process is based on an embedded system, which consists of a Raspberry Pi single board computer, Raspberry camera and headphones, the system was designed to be a low cost tool with a rechargeable battery that can be adapted to the white cane for the support and autonomy of people with visual disabilities.