In the previous year, a project was carried out on the topic of AI-based football analysis. However, this project only used one camera as a data source. In the course of this, an AI model for object recognition was created. This model is used by our extended version.Read more about the pilot project here
This project aims to extend the pilot project to use multiple cameras instead of relying on one camera to capture the whole scene. By using multiple data sources the accuracy of the produced results of the AI should be increased.
In order to achieve this aim, solely relying on AI won't bring satisfying results. The biggest component needed is the data fusion process which captures the retrieved data from every data source. It then derives the given data into a new set of information to generate game statistics.