摘要
为了避免交通追尾事故的发生,提出了一种基于现场可编程逻辑门阵列(FPGA)的嵌入式汽车测距方案。对该系统所采用的图像采集、图像处理、基于小孔成像原理的单目车距测量等算法进行研究。系统以友晶公司DE2FPGA开发板为硬件平台,通过对PAL制模拟摄像机输出的CVBS(Composite Video Broadcast Signal)视频流的采集、灰度化、滤波、边缘检测、膨胀、腐蚀、车牌定位、车牌大小的检测、液晶屏显示,实现了单目视觉汽车实时测距。整个系统使用Verilog硬件描述语言实现图像采集与显示、各种图像处理算法、车牌定位与车距测量算法,图像处理速度达到了每秒25frame/s,大大提高了系统的实时性。实验结果表明,系统所测试的距离误差最大仅为2.57%。该系统可用于汽车辅助驾驶,预防汽车追尾,也可用于智能机器人视觉。
In order to avoid rear-end collision,an embedded car ranging system based on the field-programmable gate array（FPGA） is presented and its applied algorithms such as image acquisition,image processing,monocular distance measurement based on small hole imaging and etc are investigated.The hardware platform for the design is Terasic DE2FPGA development board.The system uses a single camera to measure the distance in real time between two cars.It acquires the CVBS （Composite Video Broadcast Signal） video stream from PAL camera.The system detects the range by image graying,filtering,detecting edge,dilation,erosion,locating license plate,computing the size and liquid crystal display.The whole system uses Verilog hardware description language to realize image acquisition and display,all kinds of image processing algorithm,license plate location and distance measurement algorithm.The image processing speed reaches 25 frame/s which greatly improves the real-time performance of the system.Experimental results indicate that the maximum distance error is only 2.57 ％.The system can be used for the auxiliary driving to prevent the collision or intelligent robot vision.
出处
《液晶与显示》
CAS
CSCD
北大核心
2014年第3期422-428,共7页
Chinese Journal of Liquid Crystals and Displays
