The camera calibration is an important step in the visual processing of a soccer pitch. The camera is calibrated ahead of a match to obtain its initial position and calibration parameters. The camera operator can however manually aim and zoom the PTZ camera (pan-tilt-zoom) which leads to a change in extrinsic and intrinsic calibration parameters. The student will work on an algorithm that can automatically track points on the pitch and subsequently recalibrate the camera in real-time. The task will include research on the applicability of open-source PTZ simultaneous localization and mapping (SLAM) and its effectiveness on real data. Because of the real-time aspect, the algorithm will be implemented in a suitable programming language. In addition, a new method will be developed in cooperation with GOAL SPORT technology s.r.o. The algorithms will be then experimentally compared and included as a part of an application. The robustness of the solution will be provided by testing suites. The student will analyze the existing infrastructure and work on the possibility of the integration process and deployment.