This work deals with camera calibration in general views of football match broadcast sequences. In this context, to calibrate a camera, which is determining the camera real-world location and where it is pointing at, is essential to conduct high-level analysis of football scenes. However, the majority of camera calibration algorithms aim to get the best possible accuracy, which is time-consuming, and has high computational costs. Our proposal trades accuracy for efficiency, which is a requirement for scene analysis algorithms, keeping the accuracy as high as possible. In short, the algorithm uses the Hough transform to detect the field lines and circles painted on the grass. Then, it applies specific algorithms, depending on the semantic contents of the image, to obtain a set of calibration references. These references will be finally used to compute the parameters of a pin-hole camera model, hence completing the camera calibration process. The starting point of this project was the Master Thesis1 of Jordi Pont, which constituted a proof of concept for the use of the Hough transform information to tackle this problem. In this work, the calibration of scenes in which the goal is displayed achieves about 60 per cent correct calibrations without the use of temporal redundancy, and a new algorithm has been designed to calibrate midfield scenes which achieves about 70 per cent correct calibrations.