Monitoring video-based rapid suspicion tracking method

Abstract

The invention discloses a monitoring video-based rapid suspicion tracking method. The method comprises the following steps of: obtaining a video of a venue or nearby in an appointed time period; eliminating ineffective parts in the video; appointing a possible suspicion in the video by criminal investigation policemen according to experiences; obtaining features of the suspicion from an appointedsuspicion image; analyzing a video image input in a system; recording the analysis result and marking the analysis result in the video; outputting the marked suspicion image as a front-calibrated image; watching and confirming the marked video section by the policemen and recording the confirmation result; displaying the whereabouts of the recorded suspicion on a digital map; and for the confirmedsuspicion, comparing and querying information of the suspicion in a database. According to the method, a plurality of criminal cases in which a certain suspicion appears can be put together, and a plurality of criminal cases in which a plurality of persons appear also can be put together, so that the policemen can be helped to solve the cases and rapidly arrest the suspicions.

Description

technical field [0001] The invention relates to the field of video image processing, in particular to a surveillance video-based fast tracking method for suspects. Background technique [0002] At present, the "Skynet Monitoring System" has realized 24-hour monitoring of main roads, key units, and hot spots in various sub-districts of the city. The correlation analysis between video images and cases has become the most powerful means for the public security system to find evidence and crack cases. At present, the search for a specific target with the help of the monitoring system is mainly completed by manually analyzing massive video files. The practice of manual retrieval has proved that the analysis of image data can reveal the traces and patterns of criminal suspects' actions, which plays a key role in the detection of cases. At the same time, the data volume of video materials is massive, and the manual retrieval and analysis method not only consumes manpower and finan...

AI Technical Summary

Problems solved by technology

[0003] At present, domestic surveillance video-based fast-tracking methods for suspects include the following existing technologies: 1. Face recognition technology, which requires a clear face image in the video, and has high requirements for the quality of the original image data , generally not easy to satisfy; 2. Based on single image processing technology, this method is based on the single processing of a single image, without emphasizing the advantages of video over a single image, a lot of useful information will be lost; the above The two technologies are currently the mainstream good total technologies in the market, but their limitations make their recognition rate and speed low, which cannot meet the automatic processing of massive video data in practical applications.

Images