交通流的流体力学模型参数获取方法研究-固体力学专业论文.docxVIP

Abst Abst ract This subject is mainly concerned with how to get the hydromechanical model parameters about traffic—flow．As a marginal subject，the theory of traffic—flow is described by physics and mathematics．From macroscopieal point of view,the current theory of traffic．flow uses hydromechanical model to express traffic condition， including two important factors：flux and velocity．An accurate analysis oftraffic‘flow must depend on a lot of reasonable data．The paper discusses digital image processing as a method of getting model parameters． Firstly,the paper studies picture segmentation through background subtraction． A key issue lies in self-adaptability ofbackground model．Ifthe issue is ignored，a bad noise influence will have on image segmentation．Accordingly,the author takes advantage of mathematical morphology to remove noise．Then moving objects are be confirmed by comparing a threshold． Secondly,the author can get some valuable information，which includes segmentation window and the position of centroid．A method named kalman filter, which is a good mathematical tool to predict，is employed to trace moving object·As a result，the author will get two parameters：flux and velocity of car． Finally the solution through software is also come up with in this paper．The system Can inspect 6 Car ways at the same time，count the number of moving Cars and measure the velocity and flux of car．Moreover,the system can also set and change virtual winding freely and tulle parameter and sensitivity． The outcome of this subject may convert into an approach by hardware，which directly apply in video monitoring．It has some virtues such as reasonable arithmetic， fine precision，high velocity and good real·time． Key words：traffic-flow；digital image process；hydromechanical model；imag。 tracing Il 独创性声明本人声明所呈交的论文是我个人在导师指导下进行的研究工作及 独创性声明 本人声明所呈交的论文是我个人在导师指导下进行的研究工作及 取得的研究成果．尽我所知，除了文中待别加以标注和致谢的地方外， 论文中不包含其他人已经发表或撰写过的研究成果，也不包含为获得 北京工业大学或其它教育机构的学位或证书而使用过的材料．与我一 同工作的同志对本研究所做的任何贡献均已在论文中作了明确的说明 并表示了谢