Real-time pedestrian detection and re-identification method and device

Abstract

The invention provides a real-time pedestrian detection and re-identification method and device, which are used for monitoring specific pedestrians in multiple paths of pedestrian video streams in real time, and are characterized by comprising the following steps: performing face detection on each path of pedestrian video stream by using a target detection model to obtain pedestrian faces; calculating the similarity between the pedestrian face and the specific pedestrian face database face by using a similarity discrimination model to obtain a face similarity value; when the face similarity value is greater than a specific face threshold value, determining that the pedestrian face is the specific pedestrian face; cutting a specific pedestrian screenshot from the pedestrian video stream, and correspondingly storing the specific pedestrian screenshot and the corresponding ID to obtain a specific pedestrian library; performing pedestrian detection by using the target detection model to obtain a to-be-recognized pedestrian, and cutting out a screenshot of the to-be-recognized pedestrian; based on the to-be-recognized pedestrian screenshot and the specific pedestrian library, calculating a pedestrian similarity value through the similarity discrimination model; and when the pedestrian similarity value is greater than a predetermined specific pedestrian threshold, determining that the to-be-identified pedestrian is a specific pedestrian.

Description

technical field [0001] The invention belongs to the field of data identification, and in particular relates to a real-time pedestrian detection and re-identification method and device. Background technique [0002] Target detection technology is mainly used to detect targets in images or videos, and the detection content includes the category of the target and the coordinates of the target. In recent years, the target detection algorithm based on deep learning has made great progress. The more popular algorithms can be divided into two categories. One is the R-CNN algorithm based on Region Proposal (including R-CNN, Fast R-CNN, FasterR-CNN, etc.), they are Two-stage algorithms, which need to generate target candidate frames in advance, and then classify and return the candidate frames. The other type is One-stage algorithms such as YOLO and SSD, which can use a convolutional neural network (CNN) to directly predict the categories and positions of different targets. The Two...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above problems, a pedestrian re-identification method and device are provided to solve the problem of low pedestrian re-identification rate caused by pedestrians changing clothing accessories, seasonal dressing styles and other differences through real-time updated pedestrian database. The present invention Adopted the following technical solutions:

Images