Vehicle Counting using Computer Vision: A Survey

The growth of vehicles on the roads of urban areas generates an increase in traffic density and congestion. In this situation, most cities in modern countries are starting to implement smart solution for solving this issue, which includes the traffic management systems. And inside the traffic management system, the vehicle counting becomes an important functionality since it allows knowing the traffic density and because the data it delivers allows planning and improving the structure of the roads within cities. Many of current vehicle counting solutions contemplate the use of sensors and radars; but considering that surveillance systems based on video cameras have currently been implemented on most roads within a city, those systems are being evolved to use of computer vision for counting vehicles. In this circumstance, this work reviews different previous works developed in this area and identifies the technological methods and tools used in those works; in addition, this work also presents the trends in this area. The most relevant articles were reviewed, and the results were summarized in tables and figures. Trends in the used methods are discussed in each section of the present work.