Real-time Traffic Analytics from Live Camara Feeds

该系统将通过使用深度学习技术进行实时检测来检测和跟踪车辆，行人和jaywalkers。监控摄像头将安装在特定位置，将实时传输交通数据。将建立一个深度学习模型来检测车辆和行人，并以高精度对道路使用者进行分类。该系统不仅允许用户创建车道和人行横道来计算车辆和行人，还可以创建乱穿马路的边界来检测特定区域的乱穿马路者。Deepsort算法将用于识别和跟踪车辆和行人及其移动方向。该系统将允许用户添加和管理多摄像机流（即，不同角度）和特定摄像头的功能，如车辆计数、行人计数或乱穿马路。该系统展示了各种易于理解的仪表板可视化，以实时表示多个捕获的车辆、行人和乱穿马路者。此外，将开发一个警报系统，在发生感兴趣的事件时通过短信和/或电子邮件向指定人员发出警报。这些见解可以帮助系统用户做出明智的决策，以更好地进行交通管理，更安全的道路。

The proposed system will detect and track vehicles, pedestrians and jaywalkers by using deep learning technique for real-time detection. Surveillance cameras will be installed at specific locations which will stream the real-time traffic data. A deep learning model will be built to detect vehicles and pedestrians and classify road users at high accuracy. The system not only allows users to create lanes and crosswalks to count vehicles and pedestrians but also creates jaywalking boundaries to detect jaywalkers in specific areas. The Deepsort algorithm will be utilized to identify and track the vehicles and pedestrians as well as their direction of movement. The system will allow users to add and manage multi camera streams (i.e., different angles) and features for specific cameras such as vehicle counting, pedestrian counting or jaywalking.The system illustrates various easy-to-understand dashboard visualizations to represent the multiple captured vehicles, pedestrians and jaywalkers in real-time. Furthermore, a warning system will be developed to alerts designated persons through SMS and/or email when an event of interest occurs. These insights can the system users to make informed decision for better traffic management safer roads.