Sistema de monitoreo inteligente basado en MBSE, para optimizar el uso de los espacios en estacionamientos.

Abstract

Parking Manager is a multiplatform application developed under the model-based systems engineering (MBSE) methodology whose objective is the operation and monitoring of a parking lot using the concepts of vision and artificial intelligence. Parking lot monitoring requires two components, a predictive element and region of interest (ROI) information; For the development of the predictor element, the video game GTA V was used as a tool for simulation and generation of multiple videos and images of different types of parking lots in variable weather conditions, with the aim of building a database for training a neural network. convolutional (CNN) with the transfer learning (TL) technique using the Mobile Net V2 model; For the extraction of ROI information, a tool was developed that allows delimiting the spaces using image processing techniques and mathematical methods, consolidating said information in a standard data structure that is stored in a relational database whose engine is SQL Server through the consumption of a REST API developed in the .Net Framework. The operation consists of the management of information of vehicles that use the parking lot, by extracting the characteristics of the traffic license through an OCR, said information is sent to the API REST where it is stored and counted based on a defined rate. for charging for parking time. The application has a graphical interface (GUI) to carry out the operation and monitoring of the system, QT5 was used for the layout of the graphical interface (GUI) forms that are integrated with the Python programming language in the structure of the application. client, the use of open source software applied in a monitoring and control system assisted with Artificial Intelligence methods can be implemented with low-cost tools from existing technology and improve operational capabilities.