基于YOLO模型的车流量实时采集系统研究

Research on Real-time Traffic Flow Collection System Based on YOLO Model

对于一座现代化城市来说,合理的交通规划是一个城市高效运行的关键,作为交通规划的关键信息的城市车流量信息,原本需要人工进行识别、获取、验证的提取方式,随着计算机视觉技术的蓬勃发展弊端尽显,终将退出历史的舞台。为了提高城市车流量信息的准确性和及时性,利用现有的计算机技术设计一种基于YOLO模型的车流量实时采集系统。该系统基于YOLO视觉检测模型,采用DeepSORT算法对检测到的目标车辆进行跟踪识别、判断车辆的运行状态、实现当前路段的车流量统计、对已记录车流量信息进行可视化展示以及数据输出等。该系统可以有效地代替传统消耗人力的死板工作,实现自动化数据收集以及道路交通情况的快速监测。该系统操作简单,交互性强,为城市的交通管理和交通规划提供准确实时的信息数据。

For a modern city,reasonable traffic planning is the key to efficient operation of a city.As the key information of traffic planning,urban vehicle flow information originally needs manual identification,acquisition and verification of extraction methods,with the vigorous development of computer vision technology,will eventually withdraw from the stage of history.In order to improve the accuracy and timeliness of urban vehicle flow information,a real-time vehicle flow acquisition system based on YOLO model is designed by using the existing computer technology.Based on the YOLO visual detection model,the system uses DeepSORT algorithm to track and identify the detected target vehicles,judge the running status of vehicles,realize the traffic flow statistics of the current road section,visually display the recorded traffic flow information and output data.The system can effectively replace the traditional labor-consuming rigid work,and realize automatic data collection and rapid monitoring of road traffic conditions.The system is simple and interactive,and provides accurate and real-time information data for urban traffic management and traffic planning.