With the great success of deep learning in the field of machine vision, crowd density statistics based on machine vision have become more and more important in today's world. In order to calculate the density of people flow in a certain area, this paper combines the YOLOv4 objective detection algorithm and the DeepSORT multi-objective tracking algorithm, and uses the Kalman filter algorithm and the Hungarian algorithm for trajectory prediction and data association. And it is proposed to carry out density statistics through the Collision-line counting method. This method uses the ID information and trajectory information of pedestrians in the objective tracking process to count the number of people and directions passing through a certain set line segment, and indirectly count the density of pedestrian flow in the corresponding area. This method can also be used in conjunction with multiple cameras at multiple entrances and exits in a certain area, and has the characteristics of flexible operation and high precision. In addition, this method can also be used to detect traffic flow.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Collision-Line Counting Method Using DeepSORT to Count Pedestrian Flow Density and Hungary Algorithm


    Contributors:


    Publication date :

    2021-10-20


    Size :

    2742460 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English




    Bicycle Theft Detection by using YOLOv5s and DeepSORT algorithm

    ripa, Thamina Sultina / Hasib Khan, Tashfia Farah / Rokunuzzaman, Md. et al. | IEEE | 2024


    TrackNCount - Intelligent Vehicle Tracking, Counting, and Speed Estimation Using YOLOv8 and DeepSORT Algorithms

    Sathiyaprasad, B. / Reddy, Paluru Ambareesh Ram / Venkatesh, Digumalla Manisri | IEEE | 2025


    Traffic Flow Detection Based on Improved YOLOv5 and DeepSORT

    Lin, Xujie / Sun, Huadong / Chen, Ziyang | IEEE | 2024