Video news programs broadcast much social information in time every day, which can tell the news accurately and lively. It is one of the most popular news broadcasting ways. So some companies, which provide TV programs for audience, offer the service of news video on demand. But the companies must extract the structure information of the news video in advance. This is the procedure of news cataloguing. News cataloguing is completed by hand at present, which takes much time and human resource. This dissertation does a lot of research work on video structuring technology, presents a scheme for automatic news cataloguing, and implements a system that can catalogue the news video automatically. The scheme in this dissertation catalogues the news video on the technology of codec, pattern recognition, image processing and audio information processing. The system depending on the powerful computing capability of computers has a good efficiency and practical value. The work of this dissertation includes the following aspects: (1) Doing lots of research work on current technology of video structuring, introducing the methods of video shots segmentation, key frame extraction and video scenes construction, and analyzing the advantages and faults of these methods. (2) Providing a scheme for automatic news cataloguing. The scheme finds the number of stories in a news program and the starting time of each story by compounding the information of the anchors, captions and voice silences in the video. The scheme has a good performance in news cataloguing, and can be used in practice. (3) Introducing a metadata set for the multimedia content, to describe the results of the video structuring. (4) Designing and developing a news video cataloguing system which can catalogue the news video automatically. The cataloguing results are stored in the metadata database, and can be exported to XML files. Other video news broadcasting systems can use these XML files to get the cataloguing information of the news video.
修改评论