In order to classify pedestrians from the mixed multi-objective traffic scene quickly and accurately, this paper proposes an intelligent pedestrian recognition method based on the cascade classifier structure. Using the “from coarse to fine” strategy, a double-layer hierarchical series combination classifier is designed. HGA-BP classifier with two-layer structure is used for pedestrian recognition. Firstly, the candidates are extracted by combining the basic characteristics of the target object shape, in order to quickly eliminate most of the non-target areas, and then use the advanced features of the target to identify the candidate target areas after the processing of the previous classifier. Through the experimental analysis, this method can better classify and identify pedestrians and other negative moving objects, and count the number of pedestrians in the whole traffic scene accurately.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Research on Pedestrian Intelligent Recognition Method Based on Cascade Classifier Structure


    Contributors:
    Wang, Aili (author) / Li, Lu (author) / Dong, Baotian (author)


    Publication date :

    2020-09-01


    Size :

    354856 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    A network shaped cascade classifier based on potential functions for pedestrian detection

    Zhang, Zhongyan / Zhang, Baochang / Zhao, Kun et al. | IEEE | 2014


    A cascade classifier applied in pedestrian detection using laser and image-based features

    Premebida, Cristiano / Ludwig, Oswaldo / Silva, Marco et al. | IEEE | 2010


    Intelligent pedestrian traffic light based on face recognition

    CHEN HUA / DONG XIANGHUI / XIN MINGHAI et al. | European Patent Office | 2020

    Free access

    PEDESTRIAN RECOGNITION FOR INTELLIGENT TRANSPORTATION SYSTEMS

    Fernandez, D. / Parra, I. / Sotelo, M. A. et al. | British Library Conference Proceedings | 2005