湖南省交通管理工作中的信息采集与分析

发布时间：2018-06-07 01:28

本文选题：交通管理 + 信息采集　；参考：《湖南大学》2015年硕士论文

【摘要】：随着社会私有车辆的逐步增多,公路上的车辆超速、超载以及日常丢车的现象越来越严重,车辆牌照是车辆的唯一身份标识,因此车牌识别功能的应用范围相当广泛。本文研究的工作包括:实现抓拍闯红灯信号交通违法行为的功能和记录所有经过的车辆并进行黑名单实时比对报警功能和车辆号牌自动识别功能;实现违章识别与判断和处罚功能,可以自动识别车辆是否按交通规则行驶,并记录车辆违章证据,做到执法有据;实现了信息查询功能,可将检测到的信息以数据库形式存储起来,然后提供查询功能,允许用户输入多样化的查询条件进行信息的汇总和对比分析。经过系统需求分析和设计实现,该系统己经开始部署运行,目前运行状态良好。经过测试和实验,本系统能够正常检测车辆信息并进行信息传输和维护以及查询。本文创新之处在于:在现有的车牌识别聚类算法中,K均值、C均值聚类车牌图片识别算法在软件工程领域具有较大应用优势。但是上述算法产生的聚类中心点数目往往过多且中心点的选取上往往会产生局部最优解的问题。针对这两方面的问题,本文结合遗传算法的相关理论,采用双重遗传算法优化中心点的选取过程,并提出了DGC均值模糊聚类算法。该算法首先采用第一重遗传算法实现对聚类中心的优化,通过遗传种群群体个体的分析来确定遗传中心的个数;其次,本文算法以个体为中心进行一次聚类,得出第二重遗传算法个体适应度,从而确定聚类中心的全局最优解。
[Abstract]:With the increasing number of private vehicles in the society, the phenomenon of overspeed, overload and daily loss of vehicles on the highway becomes more and more serious. Vehicle license plate is the only identification of vehicles, so the license plate recognition function is widely used. The work of this paper includes: realizing the function of snapping red light signal traffic violation and recording all passing vehicles and carrying out real-time blacklist comparison alarm function and automatic identification function of vehicle number plate; Realizing the function of identifying, judging and punishing violation of regulations, it can automatically identify whether the vehicle is driving in accordance with traffic rules, and record the evidence of violation of the regulations of the vehicle, so as to enforce the law according to the law, and realize the function of information inquiry. The detected information can be stored in the form of a database, and then the query function can be provided, allowing users to input a variety of query conditions for the summary and comparative analysis of the information. After system requirement analysis and design, the system has been deployed and running well. After testing and experiment, the system can detect vehicle information and carry out information transmission, maintenance and query. The innovation of this paper lies in: in the existing license plate recognition clustering algorithm, the K-means and C-means clustering license plate image recognition algorithm has a great application advantage in the field of software engineering. However, the number of cluster centers generated by the above algorithms is often excessive and the problem of local optimal solution is often caused by the selection of center points. In order to solve these two problems, combined with the theory of genetic algorithm, this paper uses double genetic algorithm to optimize the selection process of center point, and proposes the DGC mean fuzzy clustering algorithm. Firstly, the first genetic algorithm is used to optimize the cluster center, and the number of the genetic center is determined by the analysis of the individual of the genetic population. Secondly, the algorithm takes the individual as the center to cluster once. The individual fitness of the second genetic algorithm is obtained, and the global optimal solution of the clustering center is determined.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP391.41;TP311.13

【相似文献】