dc.contributor |
Universitat Politècnica de Catalunya. Departament de Teoria del Senyal i Comunicacions |
dc.contributor |
Tarrés Ruiz, Francisco |
dc.contributor.author |
Enebral González, Javier |
dc.date |
2011-07-08 |
dc.identifier.uri |
http://hdl.handle.net/2099.1/12564 |
dc.language.iso |
eng |
dc.publisher |
Universitat Politècnica de Catalunya |
dc.rights |
Attribution-NonCommercial-ShareAlike 3.0 Spain |
dc.rights |
info:eu-repo/semantics/openAccess |
dc.rights |
http://creativecommons.org/licenses/by-nc-sa/3.0/es/ |
dc.subject |
Àrees temàtiques de la UPC::So, imatge i multimèdia::Creació multimèdia::Vídeo digital |
dc.subject |
Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Processament del senyal::Processament de la imatge i del senyal vídeo |
dc.subject |
Image analysis--Data processing |
dc.subject |
Multimedia systems |
dc.subject |
Tennis |
dc.subject |
Video descriptor |
dc.subject |
Video tracking |
dc.subject |
Scene analysis |
dc.subject |
Video indexing |
dc.subject |
Sports |
dc.subject |
Tennis |
dc.subject |
Vídeo digital |
dc.subject |
Imatges -- Processament |
dc.subject |
Tennis |
dc.title |
Automatic event detection for tennis broadcasting |
dc.type |
info:eu-repo/semantics/masterThesis |
dc.description.abstract |
Within the image digital processing framework, this thesis is situated in the
automatic content indexation field. Specifically during the project, different
methods and techniques will be developed in order to achieve event detection for broadcasting tennis videos.
Audiovisual indexation consists in the generation of descriptive tags based on the existing audiovisual data. All these tags are used to search the desired material in an efficient way. Televisions and other entities are looking for the improvement in operational efficiency. For this reason, the content indexation is a key factor and it is supposed to be automatic in the near future
In some sports as football where the demand of video highlights is high, a big amount of economic and human resources are available to do this task.
However, in other sports like tennis where the financial resources are not so high, automation becomes an important issue.
The report starts describing the method to separate the useful frames. These frames are those that provide information, in other words, they appear when the camera is focused on the tennis court. The following step is looking for the court and players location, with the aim of translating their perspective coordinates into the real world coordinates.
Depending on the distance travelled by the players, their movements and the duration of the shot, the implemented algorithm will be able to distinguish between different kind of shots, as aces, baseline rallies or net approaches.
The code has been developed in Matlab programming language.
The program has been tested with three tennis videos belonging to different surfaces: hard court, grass and clay. The results in terms of event detection and computing times will be detailed at the end of the report. |