Abstract: A video-based embedded vehicle detection system is presented. The scheduling of image collection is controled by FPGA and image is processed by DSP. In this system, the alternative background update method is used and the moving objects detection method is improved. It is applied to obtain traffic data for urban traffic signal control machine, and has better performances on function and maintenance compared with traditional detection methods such as inductance coil detection.

Key words: vehicle detection, background subtraction, embedded system, programmable logic device

摘要： 设计一种基于视频技术的嵌入式车辆检测系统。该系统采用DSP+FPGA硬件结构，通过FPGA进行图像采集控制，DSP进行图像处理，实现对运动车辆的检测。程序设计采用选择性背景更新法提取背景，并对运动目标的连通域检测算法进行改进。该系统应用于城市交通信号机的交通信息采集，在功能、可维护性等方面优于感应线圈等传统的检测方式。

关键词: 车辆检测, 背景差法, 嵌入式系统, 可编程逻辑器件

CLC Number: 

• U391.41