Road traffic sign is critical in informing people of the traffic situation and restricting drivers’ and pedestrians' behavior. Therefore, the real time autonomous detection and recognition ability of road traffic signs play an important role in fields such as navigation for the blind, driver warning system, and vehicle assisted driving. This study employs lightweight neural networks based on MobileNet and Yolo-V4 and compares their recognition speed and accuracy. The experimental results show that YOLO-MobileNet-V2 and YOLO-MobileNet-V3 had high recognition speed but lower accuracy. YOLO-V4-tiny and YOLO-MobileNet-V1 presented high accuracy, with mean average precision (mAP) of 90% or higher, while maintaining the detection speed over 60 frames per second (fps), meeting the real time standard in road traffic sign recognition. The study proves that YOLO-V4-tiny and YOLO-MobileNet-V1 can be employed to reduce the time of calculation and lower the requirement of hardware equipment by reducing the computational complexity. There-fore, this study provides certain practical reference in related fields.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Road traffic sign recognition based on lightweight neural network


    Beteiligte:
    Li, Le-yang (Autor:in) / Yue, Qin (Autor:in) / Luo, Rui-wen (Autor:in)

    Kongress:

    AOPC 2021: Optical Sensing and Imaging Technology ; 2021 ; Beijing,China


    Erschienen in:

    Proc. SPIE ; 12065


    Erscheinungsdatum :

    24.11.2021





    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Traffic Sign Recognition Method Based on Convolutional Neural Network

    Li, Jitong / Chen, Yuguang / Lin, Honghao et al. | ASCE | 2024


    A Novel Lightweight Traffic Sign Recognition Model Based on YOLOv5

    Li, Wenju / Zhang, Gan / Cui, Liu | ASCE | 2023


    German Traffic Sign Recognition Using Convolutional Neural Network

    Santosh, G V S Sree / Kumar, G Chaitanya / Sandeep, G et al. | IEEE | 2022