A spatiotemporal relationship matching association recognition method for masking and disguising identities

Abstract

The invention discloses a temporal-spatial relationship matching association identification method for masking and disguised identity. Firstly, the matching probability of temporal-spatial relationship before the case is calculated by using the crime occurrence time, the disappearance time of the target and the travel time of the target from the disappearance position to the crime point. , to measure the matching degree of the spatio-temporal relationship between the target disappearance time and the case occurrence time; then use the target reappearance time, the case termination time and the travel time of the target from the case occurrence point to the reappearance location to calculate the matching probability of the spatio-temporal relationship after the case, and measure the target reappearance time. The matching degree of the time-space relationship between the current time and the termination time of the case; then, according to the matching probabilities of the above two kinds of space-time relationships, the joint matching probability is calculated to measure the matching degree of the pedestrian’s disappearance trajectory and the time-space relationship of the case; finally, the detection and identity of the disguised target based on the joint matching probability identify. The present invention introduces the idea of ​​temporal-spatial behavior association, and solves the problem that it is impossible to associate appearances to disguise pedestrians solely relying on visual attributes.

Description

technical field [0001] The invention belongs to the technical field of identity identification, and relates to a disguised identity association identification method, in particular to a spatio-temporal relationship matching association identification method of masking and disguised identity. [0002] technical background [0003] When criminals carry out serious criminal activities such as bank robbery, in order to avoid the identification of surveillance cameras, they often wear masks and camouflage, which makes the suspect identification technology based on face recognition invalid, but criminal investigators can still rely on face recognition. Facial recognition technology tracks criminals. Face recognition is based on cross-camera pedestrian re-identification technology, which comprehensively uses the characteristics of human body appearance, upper and lower body clothing, clothing, wearables, hairstyle, body shape, height, gender, age, walking posture, face and other cha...

AI Technical Summary

Problems solved by technology

[0005] The present invention introduces the idea of ​​spatio-temporal behavior correlation, and provides a spatio-temporal relationship matching and correlative identification method for masks and disguised identities. Through matching and analysis of spatio-temporal motion data, the unreasonable spatio-temporal behavior of the target can be found, and the problem can be solved by relying solely on visual attributes. Unable to correlate the problem of disguising pedestrians

Images