Napredni sistem za praćenje pokretnih objekata u sportu

Abstract

This dissertation describes in detail the designed advanced tracking system for moving objects in sports. New methods, new algorithms for detecting the position and movement of objects in space are described and introduced, and a new approach to the use of technologies is created, new algorithms and applications are developed to increase the accuracy and reliability of tracking systems even when using equipment with lower technical characteristics. For the first time, experimental testing and comparative analysis of the experimental results of two different software frameworks with algorithms, using different video cameras, as well as different developed software applications for detection and 3D simulation and graphical representation of object tracking results were performed. The idea for the implementation of technical-technological systems for tracking movable objects in sports is to monitor and detect the position and trajectory of the moving object in real time. These collected data are processed with precisely set methods, and in the end, as a result of the whole process, to produce photorealistic 3D visualization, a simulation that shows the predicted motion path of the observed object in space in order to make a precise and objective judge's decision. The focus of the research is the application of new algorithms for the collection, detection and graphical simulation of the position of objects in space with the development of new original methods for improving the 3D graphical representation without the flickering of objects on the screen, increasing the accuracy of 3D graphic animation up to 10 parts of a millimeter, which leads to a more precise positioning of the moving object, with optimization of motion detection for lower quality hardware. Emphasis is placed on practical system testing on a number of different types of video cameras and a number of software applications in different programming languages with a large number of open-source motion detection algorithms. The innovativeness of the proposed system is to improve the process of determining the position of the observed objects in space with simple new techniques and improve the graphical representation with less use of computer resources. The introduction of new algorithms and new methods is the result of mathematical calculations and a new engineering approach to designing such systems.