Desarrollo de un prototipo de semáforo inteligente con visión por computador

Abstract

The technical project was developed through the approach of making the prototype of an intelligent traffic light with computer vision; with a car detection module, through which two-way cars at an intersection are counted in real time, to give priority to the road with the largest number of vehicles. With this information, with the help of a traffic light controller module, the time of change of light is determined dynamically, giving priority to circulate, to the road with the largest number of vehicles, reducing traffic congestion and delay in mobility at that intersection. To program the vehicle counting and traffic light control modules, the Python language was used, essential for object recognition, and to develop the vehicle detection module; one of its libraries called OpenCV. Capturing and processing the images obtained by the camera is able to detect the vehicle among the other objects for counting and with said information through the traffic light controller module distribute the time for the change of light, the data obtained is stored in a postgres database. To carry out the project, in addition to the research, computer vision has been applied mainly, which is a branch of study of Artificial Intelligence, and which, added to the programming implemented in the modules, makes the developed prototype differ from a conventional traffic light.