Análise da viabilidade da utilização de visão computacional no levantamento de tráfego para o dimensionamento de pavimentos asfálticos

Abstract

Given the great importance of road transportation for both freight transport and people in Brazil, it is essential that the Brazilian road network connects several points of the country and mainly meets the quality requirements. For this, it is necessary study the area of pavement design and develop technologies that allow a better survey and study of traffic. The traffic study is important in order to determine the value "N", the number of passes of a standard axle by a highway, which is used to determine the thickness of each layer of road pavements. Currently the two main ways of surveying traffic are manual counting and the use of sensors, which, despite being efficient, have a relatively high cost and / or do not allow to control traffic evolution over long periods. The present work had the objective of studying the feasibility of the use of computer vision, a field of artificial intelligence, in the survey of traffic through recordings of the flow of vehicles of a highway. It was concluded at the end of the project that the quality of the wheel and axle model has a direct impact on the approaches used by the program, the development time and, consequently, the results obtained. Based on the results and the different possibilities of improvement thought and presented throughout the project, it is necessary that further studies in this area being developed to define its viability.