Vehicle access control using machine learning

Abstract

Being able to apply machine learning techniques involves using technology that aims at innovation and that provides companies with certain strategic advantages over the competition they face. The purpose of this research is to develop a pattern recognition system to identify vehicle license plates that uses machine learning techniques to control the entry of vehicles into University facilities. The methodology used is Scrum, allowing cyclical incremental work "sprints", in which there were short times where a number of tasks were carried out, achieving planning and control of the development of the project, fulfilling scheduled tasks, intervening three roles the scrum master, the product owner and the team. The solution is built by a set of small services, to guarantee reusability, scalability, flexibility and all the advantages that a microservices architecture offers us. The results of the technological development in the research made it possible to demonstrate that it is possible to identify and recognize license plates accurately and quickly using machine learning models and algorithms, managing to overcome difficulties that include various image conditions, accurate detection and segmentation of license plates, as well as the system adaptability to various fonts, sizes and styles of license plates.