CarSight: real-time vehicle identification for theft mitigation

Dissertação de mestrado em Informatics Engineering ; Despite efforts to combat vehicle theft, this type of crime presents a significant threat to global security, where in 2022 Portugal occupied 12th place in the European ranking of countries with the most vehicle thefts. When detecting a stolen vehicle, traditional techniques are still used, such as manual identification carried out by the police, these methods usually require a lot of resources and are time-consuming. The application of Artificial Intelligence to combat car theft presents an innovative approach to public security. AI systems are capable of quickly and accurately identifying stolen vehicles, which speeds up the response of security forces and makes the streets safer. Along with this major problem, there is the Tidy City platform that proposes to equip urban vehicles on the island of Madeira, such as garbage trucks or delivery vehicles, with mobile devices to automatically detect and classify infrastructure problems, one of the most used devices being cameras, as a result of their use thousands of images are generated daily, these can be used to find vehicles. Therefore, the main objective of this project is to develop an artificial intelligence system capable of identifying a specific vehicle based on its license plate, having greater efficiency than the traditional methods currently used, thus reducing the number of stolen cars and increasing the number of vehicles. Cars recovered from theft. The Tidy City platform must be integrated into the system so that its images can be consumed, The system that was developed during this project can be divided into two main components. The first component consists of recognizing number plates, while the second component aims to recognize and identify the different characters present on the plates. To perform license plate recognition in images, two models were trained using YoLoV8. It was necessary to select a data set large enough to cover the various variations present in images from the Tidy City ...