Recovering Football Games Based On Multi-Cameras Calibration

The way to restore a three-dimensional football match is by utilizing the two-dimensional images, which including setting the video camera calibration, restoring the coordinates of the three-dimensional football match field and modeling the football circumstances. The model can be used in a lot of fields such as virtual reality, computer game design, etc. The most important is video camera calibration and it as well is the basis of restoring the coordinates of the three-dimensional football match.The process of the video camera calibration is to calculate the internal and external parameters of the shooting camera, which is the precondition of the camera imaging model. In this dissertation, it gives out a new way to calculate the shooting camera self-calibration for a football match. In detail, just by the camera self-calibration we can get the camera's internal and external parameters and structure information. The advantage of the method is that we can finish the visual task online without exact calibration devices. There are several ways to calculate some object's three-dimensional coordinates by using the solid images of the shooting object. The equivalent point can be found by match images,the 3-D coordinates can be restored by the fact that linearity 3-D measures law and relatively directional, absolute directional。The model of the football match includes mannequin model of football player and football field model. The mannequin framework model comprises human joints and bones. The bones can be deemed as rigid body and the whole human framework as the system of rigid body. To build the football match field model, some important points can be extracted by the cubic regular of the field. This dissertation restores the three-dimensional football match and the players. And the experiment data prove the hypothesis.