A method and system for moving target retrieval based on multi-video collision
A technology of moving target and retrieval system, applied in the field of video investigation, can solve the problems of low efficiency, easy to miss important information, time-consuming and labor-intensive, etc., and achieve obvious effect of improving efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Publication Date
- 2019-05-17
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
technical field
[0001] The present invention relates to the field of video investigation, and more specifically relates to a method and system for retrieving moving objects based on multi-video collision. Background technique
[0002] Moving targets, including people, vehicles, and objects, are the core information in surveillance video. Finding the same moving target that appears in the two videos is of great significance to the detection of the case and helps to quickly lock the suspect;
[0003] At present, this work has been done manually, which is very time-consuming and laborious, and the efficiency is not high. It is easy to miss important information, which is an important factor restricting the rapid detection of cases. Although there are technologies and application software for extracting snapshots of moving objects in videos, there may often be thousands of snapshots in each video. Without prior conditions, it is necessary to find the same moving object in multi...
Examples
Embodiment Construction
[0032] A method and system for retrieving moving objects based on multi-video collisions according to the present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0033] The following is the best example of a multi-video collision-based moving object retrieval method and system according to the present invention, which does not limit the protection scope of the present invention.
[0034] A moving target retrieval method based on multiple video collisions, comprising the following steps:
[0035] S1. Extract a snapshot of the moving target for each video involved in the collision;
[0036] S2, extracting its feature value for the moving target snapshot of each video;
[0037] S3. Two-by-two collisions are performed between the videos participating in the collision, and a moving target snapshot meeting the similarity requirement is obtained, and a target matching pair sequence is formed;
[0038] S4. Analyzing all targ...