Futbol Maç Analizi İçin Maç Yayın Görüntüsünden Kamera Kalibrasyonu

Abstract

Digital camera calibration provides the matrix required for position conversion using the positions of known points in the image frames. By using this matrix, the image frames are transformed into image frames that look at the field from the top at vertical angle, and thus a relationship can be established between the positions.

In-field analysis in football first converts all kinds of actions performed by the players on the field into data. These analyzes have a great importance in development of the deficiencies of the teams and football players by using the obtained data. Teams are preparing for their next matches with detailed analyzes of the matches they have played before. At this point, on-field analyzes with high accuracy and fast results will enable the teams to improve their game quality and to perform all kinds of movements in the field more efficiently.

In the presented study, there is a mathematical model that predicts the elements of the perspective transformation matrix for digital camera calibration and the position on the real football field of the points determined in the frames obtained from the football video images taken with a single camera, using computer vision algorithms. The aim is to create an environment that will enable rapid on-field analysis of football with the location information at the last point obtained.