Prototipo de un sistema web para el conteo de vehículos en tiempo real utilizando la librería de visión artificial OpenCV

Abstract

This document details the implementation process of a web system prototype for counting vehicles in real time using OpenCV computer vision library. The prototype architecture includes the following modules: video acquisition and transmission, reception and processing, storage and reporting. The acquisition and transmission module captures video through IP cameras and sent it to a server hosted on the Internet. The reception and processing module processes the captures in real-time and performs vehicles counting using an algorithm developed with the OPENCV library. The storage module persistently stores the result of the vehicle count, along with the time and date. The reporting module provides graphical interfaces accessible by web that allows CRUD operations over users, initial system configuration parameters and cameras; also reports vehicle count from each camera geo localized on a map. Finally, the results of the effectiveness of the prototype are shown.