| Network traffic content-based video retrieval is combined with video retrieval technology, network technology and database technology.This technology will be applied to traffic video system and aimed to built a content-based traffic video retrieval system, wich will promote the entire transport.The content-based video retrieval process was analysed based on network applications. The correlation algorithm and realization skill were also studied in this paper. According to the characters of the traffic video, the research skill integrate with color and shape characters method was deep studied. Then from the perspective of network applications, given a architecture design of distributed video database retrieval system. Lastly, designed a user search interface for the transport filed.The main study of this article can be listed as follows:(1) Retrieved from the user research and network capabilities to achieve, combined the characteristics of the traffic video, the content-based video retrieval on how to achieve the network application environment was deeply studied and discussed.(2) Because of the poor results in the single feature video retrieval approach, the method of fution color and shape characteristics of the image was studied. This method was used in the video retrieval system. The experimental result show that it is better to use the method of fution color and shape feature than the method of signle feature video retrieval.(3) According to content-based video retrieval technology, the transportation video database mangement system and network database architecture databases were combined. The database structure and network structures were analyzed, the B / S three-tier network architecture was also considered, a architecture design method of distributed traffic video database retrieval system was presented.(4) Finally, a system's user function search module on.NET platform was developed. This search methods system mainly account of the video characteristics compared to the traditional retrieval methods. According to a hierarchical video retrieval needs, a sample retrieval, object retrieval, video segment retrieval, video browsing, and keyword search were provided for user, meet the user's personalized seared need. And then a variety of search methods were described in detail, the flow chart and page structure diagram of each search method were presented. |
