A traffic video retrieval system and its realization method

Abstract

The present invention provides a traffic video retrieval system and its implementation method. Among them, a traffic video retrieval system, including: video modules of embedded geographical information, for establishing the corresponding relationship between geographical information and recording video; geographical informationThe storage module is used to alternate geographical information with the corresponding videos; the video retrieval module of geographical information is used to carry out the corresponding retrieval of the video with geographical information through the index table;The retrieved video with geographical information is performed accordingly; the video monitor is used to recording pavement video data and performing corresponding monitoring operations;The video file corresponding to the range is played and the retrieved video files are played.

Description

technical field [0001] The invention relates to the technical field of video retrieval, in particular to a traffic video retrieval system and an implementation method thereof. Background technique [0002] With the development of computer technology and network technology, the construction of information highway, and the promotion and application of multimedia, various video materials are continuously produced, and more and more video databases have been established, and digital libraries, Many new service forms and information exchange means such as digital museum, digital TV, video on demand, distance education, and telemedicine. As an important means for people to obtain video information, video retrieval technology has been paid more and more attention, which has greatly promoted the development of video search industry. [0003] Content-based video analysis is to extract the main content from all frames and describe the video content in a structured manner from bottom ...

AI Technical Summary

Problems solved by technology

[0014] At present, video retrieval systems generally adopt content-based retrieval methods, which mainly use the basic features of video images and combine various visual features for joint retrieval. Since people's understanding of video content is based on the existing knowledge of humans, However, these low-level features cannot reflect empirical knowledge, how to effectively solve the problem of "semantic gap" (Semantic Gap) between low-level video features and high-level abstract concepts understood by humans, and how to effectively use a large number of semantics contained in video streams Information realization of video retrieval is the key and difficult problem of current research

Images