Multisource big data fused video monitoring pedestrian identity identification method

Abstract

The invention discloses a multisource big data fused video monitoring pedestrian identity identification method, which uses the law of consistency of multisource data object space-time behavior to identify the physical identity of a pedestrian by virtue of a communication cell phone number of a monitory point. The main steps includes: observing the number of times of reappearance of cell phone numbers on a monitory point where a specific suspicious pedestrian on a travel path; according to an identification accuracy requirement, determining an appropriate number of observation monitory points; determining the number of pedestrians that reappear including the specific suspicious pedestrian and the number of cell phone numbers that reappear of each monitory point on the travel path; calculating pedestrian and cell phone number reappearance probabilities, and combined probabilities of the two; and ranking according to the combined probabilities, and selecting combinations which are ranked in the front to be output. The method provided by the invention has the advantages of simple calculation, high identification efficiency, strong universality and the like.

Description

Technical field [0001] The invention belongs to the technical field of video surveillance, and relates to a surveillance video pedestrian identification method, in particular to a multi-source big data fusion video surveillance pedestrian identification method. technical background [0002] In recent years, with the increasing complexity of social security, video surveillance systems have been widely deployed in urban areas, and urban video surveillance systems have increasingly played an important role in maintaining social stability and fighting crime. The huge video information collected by the video surveillance system has become the first-hand information source for the public security organs to solve the case. Surveillance video captures the image information of monitored targets such as pedestrians and vehicles, but the image information itself does not directly reflect the physical identity of the monitored target, such as the person's name, ID number, hometown and other ...

AI Technical Summary

Problems solved by technology

For monitoring probes deployed in outdoor places such as streets, because they do not have the conditions to obtain registration information, registration information cannot be used to achieve the purpose of identification. However, today when mobile phones are widely used, pedestrians’ mobile phone call records can help to identify, for example, a certain In the monitoring video of an open-air place, only one pedestrian appears at a specific time. If the nearby base station monitors the mobile phone communication at the same time point, the mobile phone number is very likely to belong to this person, because the positioning of the mobile phone already has good accuracy. , my country has adopted a real-name registration system for mobile phones, and it is easy to restore to a real person through a mobile phone number

[0006] However, in reality, whether it is registration information or communication records, there is often a many-to-many relationship with the targets in the surveillance video. For example, the base station has multiple phone numbers, and there are multiple pedestrians in the surveillance video.

Images