摘要
为了探测前车车距,采用机器视觉方式,以CCD构造单目成像系统采集前车图像,以检测车牌在图像中像素的数量方式实现车距的测量。系统以单目摄像头的成像模型,通过物像位置关系及大小关系,建立车牌图像与车距信息之间的模型;以CCD成像原理,建立车牌图像与传感器像元像素之间的模型。由此构建测距检测算法模型。采用MATLAB软件平台设计人机交互界面,对采集图像进行预处理提高图像的对比度;通过匹配连通域等算法,对车牌进行定位、分割,检测车牌水平方向像素数量;在此基础上,采取小孔成像原理计算车距的大小,并显示在软件平台上。实验证明了本系统能对前车图像进行分析从而计算出车距,该系统对3 m之外的车距检测平均误差为4%。
In order to detect the front vehicle distance,a monocular imaging system based on CCD was used to collect the vehicle image,and the vehicle distance was measured by detecting the number of pixels in the image.The model including license plate image and vehicle distance information was established by the position and size relationships between object and its image based on the image model of monocular camera.Meanwhile based on the principle of CCD imaging,the model between license plate image and sensor pixels was established.Then the model of ranging detection algorithm was constructed.By building a CCD imaging system,the license plate image of the measured vehicle was acquired.The man-machine interface was designed by using MATLAB,and pre-processings were carried out to improve the image contrast.By matching connected regions and other algorithms,the license plate was located and segmented,and the number of pixels in the horizontal direction of the image was detected.On this basis,the size of vehicle distance was calculated with the principle of small aperture imaging and was displayed on the software platform.Experiments proved that the system can analyze the image of the front vehicle and calculate the distance,the average error of the system is 4%for the distance detection beyond 3 meters.
作者
程瑶
赵雷
成珊
田又源
米曾真
CHENG Yao;ZHAO Lei;CHENG Shan;TIAN You-yuan;MI Zeng-zhen(College of Mechanical Engineering,Chongqing University of Technology,Chongqing 400050,China;The Tenth Research Institute of China Electronic Technology Group Corporation,Sichuan,Chengdu 610036,China)
出处
《计量学报》
CSCD
北大核心
2020年第1期11-15,共5页
Acta Metrologica Sinica
基金
国家自然科学基金(51505052,51505053)
重庆市科委基础与前沿研究一般项目(cstc2016jcyjA0497,cstc2015jcyjA40051)
重庆市教委科学技术研究资助项目(KJ1500935,KJ1500931,KJ1500617)
关键词
计量学
车距检测
机器视觉
图像处理
metrology
vehicle distance detection
machine vision
image processing
