基于视频的车辆闯红灯检测系统设计

设计一种基于改进帧差法的车辆闯红灯检测系统。采用ARM＋FPGA的硬件平台,通过FPGA对视频图像进行实时采集,ARM进行图像处理,采用基于虚拟线圈的车辆检测方法,结合交叉路口的交通信号控制器,对违章闯红灯的车辆进行检测。在功能、安装及可维护性方面优于基于感应线圈的检测系统,在成本方面优于基于DSP的视频检测系统。

Design of video-based vehicle red-light violation detection system

A vehicle red-light violation detection system with improved frame difference was presented. A hardware architecture of ARM and FPGA was used in this system. Real-time image collection was controlled by FPGA and the image was processed by ARM. Using vehicle detection method with virtual coil and cooperating with traffic signal controller, this sys- tem could detect vehicles which violate red-light. It is superior to the detection system with induction coil in aspect of function and maintainability, and the cost is less than DSP-based system.