A people counting method and people counting device applied to entrance and exit monitoring

By processing monitoring videos using target detection and trajectory classification models, accurate number of people at entrances and exits can be counted, simplifying the detection process and improving accuracy.

CN117315539BActive Publication Date: 2026-07-21BOE TECHNOLOGY GROUP CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
BOE TECHNOLOGY GROUP CO LTD
Filing Date
2023-09-28
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing technologies for entrance and exit monitoring require calibrating cameras to draw sensing areas or lines, resulting in complex detection processes, susceptibility to errors, and high maintenance costs. Furthermore, these technologies can only count the number of people but cannot accurately count the number of individuals.

Method used

A pre-set target detection model is used to detect targets in the monitoring video. A pedestrian tracking sequence is formed through time-series tracking. A trajectory classification model is used to classify the movement trajectory and count the number of people entering and exiting.

Benefits of technology

It simplifies the pedestrian target detection process, accurately counts the number of people entering and exiting at entrances and exits, and solves the problems of complex detection processes, error-proneness, and the inability to count only the number of people.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117315539B_ABST
    Figure CN117315539B_ABST
Patent Text Reader

Abstract

The application discloses a people counting method and device applied to entrance and exit monitoring. The people counting method of one embodiment comprises the following steps: using a target detection model to sequentially perform target detection on collected images and output multiple target detection frames; performing time sequence tracking on target pedestrians corresponding to the target detection frames in each frame of the collected images respectively to form tracking sequences of the target pedestrians; in response to one target pedestrian leaving the monitoring video, obtaining a travel trajectory of the target pedestrian according to the tracking sequence of the target pedestrian; using a pre-set trajectory classification model to classify the travel trajectory and output a classification result, and counting the number of people entering and exiting an entrance and exit position according to a classification definition corresponding to the classification result. The embodiment provided by the application can simplify the pedestrian target detection process and accurately count the number of people entering and exiting by using the trajectory classification model based on the travel trajectory of the target pedestrian.
Need to check novelty before this filing date? Find Prior Art