The invention provides a vehicle violation detection method based on deep learning. The method comprises the following steps: collecting a road surface solid line scene data set; marking a background region, a yellow solid line region and a white solid line region of the pavement solid line scene data set; pixel values are set to distinguish the background area, the yellow solid line area and the white solid line area to generate a mask label image; performing data enhancement expansion processing on the mask label image to generate a training set; inputting the training set into a segmentation network with a global pooling module for training to obtain a pavement solid line segmentation model; inputting the vehicle real-time image into the road surface solid line segmentation model and the vehicle target detection model, and outputting a yellow solid line area, a white solid line area and a vehicle target; the position relation of the vehicle target relative to the yellow solid line area and the white solid line area is detected, and a vehicle violation detection result is output; according to the vehicle violation detection method based on deep learning, the image segmentation effect is improved, and the accuracy of vehicle violation output is improved.

    本申请提供了一种基于深度学习的车辆违规检测方法,包括步骤:采集路面实线场景数据集;对路面实线场景数据集的背景区域、黄实线区域、白实线区域标注;通过设置像素值区分背景区域、黄实线区域、白实线区域生成mask标签图像;将mask标签图像进行数据增强的扩充处理,生成训练集;将训练集输入带全局池化模块的分割网络中进行训练,得到路面实线分割模型;将车辆实时图像分别输入到路面实线分割模型以及车辆目标检测模型,输出黄实线区域、白实线区域以及车辆目标;检测车辆目标相对于黄实线区域、白实线区域的位置关系,输出车辆违规检测结果;本申请的基于深度学习的车辆违规检测方法提升了图像分割效果差以及输出车辆违规的准确率。


    Access

    Download


    Export, share and cite



    Title :

    Vehicle violation detection method based on deep learning


    Additional title:

    一种基于深度学习的车辆违规检测方法


    Contributors:
    CHEN YAQIONG (author) / LUO FUZHANG (author) / ZHU GUANGQIANG (author) / GONG XIAOLONG (author) / LI XIAOKAI (author) / CHEN HAITAO (author) / LAI SHIWU (author) / ZHAN YANBO (author)

    Publication date :

    2022-11-11


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    G06V / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS



    Traffic Rules Violation Detection using Deep Learning

    Tonge, Aniruddha / Chandak, Shashank / Khiste, Renuka et al. | IEEE | 2020


    Traffic Monitoring and Violation Detection Using Deep Learning

    Sargar, Omkar / Jain, Saharsh / Chittupalli, Sravan et al. | Springer Verlag | 2023


    Traffic Rule Violation Detection System: Deep Learning Approach

    Kathane, Mandar / Abhang, Shubham / Jadhavar, Abhishek et al. | Springer Verlag | 2022


    DEEP LEARNING BASED TRAFFIC VIOLATION ENFORCEMENT SYSTEM AND METHOD

    JUNG BOOEUN / CHOI HEUNGMOOK / KIM YOUNGKYU et al. | European Patent Office | 2020

    Free access

    Vehicle model filtering-based traffic violation detection method

    CAO QUAN / WANG YU | European Patent Office | 2015

    Free access