Sistema de Visión Artificial para la Detección y el Reconocimiento de Señales de Tráfico basado en Redes Neuronales

Published in: Innovation and Development for the Americas: Engineering, Education, Research and Development: Proceedings of the 8th Latin American and Caribbean Conference for Engineering and Technology
Date of Conference: June 1-4, 2010
Location of Conference: Arequipa, Peru
Authors: José R. Escorcia Gutierrez
Casimiro. A. Rocha
Refereed Paper: #158

Abstract

Nowadays, the activities that go unnoticed by the man or the lack of information needed to understand situations are the causes of accidents at large. This project aims to guide men in their activity of driving through artificial vision. The vision has been sold to work in many disciplines such as biomedicine, face recognition, fingerprint recognition, and so on.
For this project, it has been created a detection system based on neural networks and moment invariants capable of learning and train with some traffic signals with the goal of assisting the driver not to commit a breach or in the worst case an accident, the neural network will be able to recognize a traffic signal some distance so that the driver has a priori knowledge of this.
The system was implemented on MATLAB and this shows better performance than others based in fuzzy logic and basic image processing.