Reconocimiento y geo-posicionamiento de señales de transito

This degree work presents the development of a computer vision system that addresses the recognition and automatic geo-positioning signals vertical traffic regulations in the city of Cali. The system indexes each traffic signal in a database, so that in the future can be altered or damaged locate signals. Additionally, it is expected to contribute to the development of driver assistance modules in the prevention and care of road safety rules in Colombia. multilayer neural networks (ANN) and support vector machines (SVM): In this project, two pattern recognition techniques are compared. SVM were chosen because they generate better results. an application of computer vision which is responsible for the detection and real-time targeting traffic signal captured through a webcam developed. Segmented signal is subsequently recognized with support vector machines. Data latitude, longitude and altitude of the traffic signal sensing system come from the GPS a mobile device and are stored in a MySQL database through a system based on PHP, HTML and XML application, achieving its interpretation and visualization thanks to Google maps API that allows the Geo-positioning.

• Autores: González Correa, Luis Alberto; Amariles García, Sebastián; Elvira García, Felipe
• Fecha de publicación: 2014
• Institución: Universidad de San Buenaventura
• Repositorio: Repositorio USB
• Idioma: spa
• OAI Identifier: oai:bibliotecadigital.usb.edu.co:10819/3115
• Acceso en línea: http://hdl.handle.net/10819/3115
• Palabra clave: Reconocimiento; Geo-posicionamiento; Transit; Designate; Visión por computador; Visualización de la información; Diseño de software; Señales del tránsito
• License: http://purl.org/coar/access_right/c_abf2