Um Classificador Visual Autônomo de Densidade de Tráfego Urbano
Autores
5336 |
Gabriel Marques Rosario
|
2411,131
|
5337 |
2411,131
|
Informações:
Publicações do PESC
The concept of Intelligent Transportation Systems, which originated in the revolution in computing and communications occurred in the '90s, is being increasingly discussed in Traffic Engineering as a means to enhance existing transportation systems through the facilities offered by new technological tools. One of its possible applications consists in an adaptive signal control system capable of making decisions, in real time, based on the behavior of traffic in a city and then control, in an intelligent manner, the traffic signals timing, improving traffic flow and reducing congestion. However, this solution requires a means of extracting, from the traffic facilities being observed, information like flow rate or density, so it can, then, make its decisions. Taking into account the widespread use of common traffic cameras and the reasonable price at which a high computational power can be obtained, this work proposes a fast, efficient and low cost alternative to commercial solutions in determining the situation in a traffic facility. Purely based on image analysis, the developed solution, which works autonomously, requiring minimal human interaction, is able to determine, second after second, the density of a traffic stream. Moreover, a second method was developed using weightless neural networks for determining the current situation of a traffic facility based on the relationship between flow and density.