Sistema de reconocimiento facial utilizando visión artificial basado en una arquitectura iot, para el conteo de usuarios en las unidades de la Cooperativa de Transporte Público “Unión Ambateña”

Abstract

Over the years, the transportation company Union Ambateña has seen it is necessary to use different control systems in each bus, in order to track the number of passengers who use the public transportation every single day and be able to know how much money they receive. However, it has been difficult to get accurate data due to limited technology. A system with IoT architecture is presented for detection, classification, and counting of people using the transportation service. An EZVIZ H6c 1080p Wi-Fi camera and an EXTREM 720p camera are used to capture real-time videos of people and faces inside each bus, which are processed on the NVIDIA Jetson Nano development board. The detection models count the total number of passengers boarding the bus and also detect and record the number of adults using the service. The system is developed with the YOLOv8 artificial vision algorithm, which handles adult detection, as well as detection of all types of people. The results of the user count obtained are displayed on the Node-RED platform, which is used with a local service to show the value of the final revenue report. The results are stored and managed on the Ubidots IoT platform, which allows for control of the transportation units and decision-making with real-time data.