SmartView: Automatic Monitoring Applied to Major Sports Competitions
Due to the large audiences attending sports venues, such as stadiums or gyms, several problems arise. These problems are difficult to deal with efficiently when the entire process of analysis is performed manually, as they can range from detection of conflicts occurring in the bleachers and the identification and location of the individuals involved, to the delimitation of regions of the bleachers or group of individuals that should be monitored with greater attention due to their improper behavior. In order to assist in the monitoring and, consequently, the safety of the fans present in sports competitions, this project aims to employ computer vision techniques to automate the resolution of the aforementioned problems and provide relevant information to those responsible for monitoring the fans in sports facilities, increasing precision and efficiency in decision making. As a result, security agents will only receive segments of videos containing situations classified as of interest from the point of view of monitoring environments and surveillance, which may require some type of human intervention. In order to obtain the information necessary for the satisfactory execution of the monitoring system, the project focused on obtaining robust solutions through the use of accurate visual characteristics descriptors, which are used to reduce the propagation of errors in the detection, recognition and re-identification of people, as well as in the recognition of their actions. The project also focused on the development of efficient methodologies that will allow the processing of large amounts of visual data.
Contributions generated from this research…