A Positionable Video Retrieval Method Based on Two-Dimensional Time Grid

Abstract

The invention provides a positionable video retrieval method based on a two-dimensional time grid. The method includes: according to the earliest start time T in all locatable video segments s , the latest end time T e And time resolution, combined with the start time and end time of each locatable video segment, calculate the row and column number information of each locatable video segment, and obtain a two-dimensional time grid; according to each locatable video segment in the two-dimensional time grid The row and column number information in the network generates its corresponding spatial point for each locatable video segment, and finally obtains the spatial point set P; uses the R-tree spatial index method to generate the R-tree spatial index structure corresponding to the spatial point set P; according to the locatable The start time and end time defined by the video retrieval conditions are positioned to the corresponding time grid in the two-dimensional time grid, and the corresponding spatial point is generated according to the row and column number information of the time grid, and the space is obtained by searching in the R tree spatial index structure point, the locatable video corresponding to the spatial point is obtained.

Description

technical field [0001] The invention relates to the technical field of geographic information and video retrieval, in particular to a two-dimensional time grid-based positioning video retrieval method, which can be used in the field of mass positioning video data management and analysis. Background technique [0002] At present, public safety has become one of the focal issues that governments at all levels pay attention to; the construction of video surveillance systems for public spaces has become a key link in ensuring public safety and deterring criminal behavior. With the continuous advancement of my country's Safe City, a large amount of video data has been accumulated all over the country; these videos are all locatable videos including shooting time and spatial location. How to efficiently retrieve the locatable video data at the corresponding time has become a fundamental problem in the management and analysis of large-scale surveillance video. Locable video is a v...

AI Technical Summary

Problems solved by technology

Among them, the current retrieval based on time information has the following problems: (1) Timestamps are mainly used for direct comparison, and the retrieval efficiency is low due to the complexity of timestamp data types; (2) One-dimensional timelines are mainly used, and absolute time is used Judging, failing to make full use of two-dimensional spatial semantics for time information retrieval of locatable videos; (3) The retrieval method is single and lacks a graphical time-based video retrieval operation method

Images