【作者】 梁栋；

【导师】 潘孟贤； 高隽；

【作者基本信息】 合肥工业大学 ， 信号与信息处理， 2002， 硕士

【摘要】 随着社会经济的发展，交通拥挤和堵塞现象日趋严重，交通污染和事故越来越引起社会的普遍关注。这种情况下智能交通系统ITS应运而生，其核心技术是电子技术、信息技术、通信技术和系统工程。目前各国都在积极致力于本国ITS的研究和构架。 车牌作为汽车的标示具有唯一性，知道了车牌号，则车辆的所有信息，如车种、车主等便一目了然，如果能实时识别路网上所有正在运行的汽车车牌并将其汇总，可以为交通流诱导系统、交通控制和管理系统提供最详尽的信息，在大型停车场的管理系统、公共安全、交通管理及有关军事部门有着特别重要的实际运用价值，所以车牌自动识别系统成为智能交通系统中的一个研究热点，正日益受到人们的重视。 影响车牌自动识别系统识别精度的一个重要因素就是车牌区域定位的准确程度。这是因为由摄像机得到的原始图像尺寸较大，所需存储空间较大，又有众多的干扰区域存在，在图像中直接对车牌字符进行识别是十分困难的，如能准确定位车牌区域和切分字符，可以减少存储容量，避免干扰，进行准确的字符识别，从而提高整个识别系统的工作效率和识别精度。因此车牌区域定位分割的算法研究一直都是各国学者研究的热点问题。 本文通过对大量资料的搜集、整理，总结了近年来国内外在车牌定位分割领域的最新研究成果和最新进展，对车牌区域的各种固有特征和目前的车牌定位分割技术进行了系统的研究和探讨。在前人工作的基础上，首先研究了形状特性在车牌区域定位上的应用，提出了基于形状特性和反Hough变换的车牌区域定位算法，给出了车牌区域的形状特性在Hough空间的表现形式，详细论述了反Hough变换的原理及应用，并根据其原理在Hough空间滤除干扰，在图像空间重建车牌区域。此外针对倾斜车牌区域的定位分割问题，提出了基于灰度变化特性和方向场计算的定位算法，其中详细讨论了数学形态学算子在目标区域粗定位中的应用，重点介绍了方向场计算理论和实现方法，并将其应用于目标区域倾斜角度的检测，根据检测得到的结果进行了倾斜车牌区域的准确定位和校正，为后续的字符分割打下良好的基础。更多还原

【Abstract】 With the development of the world economy,traffic jam,pollution and traffic accident are more and more serious. Thus Intelligent Transport System (ITS) comes out and now many countries focus their attention on the study and constitution of ITS.As the mark of vehicle,license plate is unique,if we recognize all the vehicles in the path network in real time and gather the information,the most detailed information for ITS can be obtained. So license plate recognition system (LPRS) has important application in many places. At present,LPRS has become a research hotspot in ITS.The accurate degree of license plate location is an important parameter of LPRS. It is difficult to recognize the character directly in original image because the original image has big dimensions,which need big memory,and at the same time,there are many disturbed areas in the image. However,if we can locate license plate and segment character precisely,thus the capacity of storage can be reduced and the disturbance can also be avoided,and so the character will be recognized accurately,which can ultimately improve the efficiency and precision of the recognition system. Therefore,the study on the algorithm of license plate location is always the hotspot problem.In this paper we have summarized the latest research achievements and development of license plate location and discussed the intrinsic characteristic of license plate. On the basis of previous work,firstly,we studied the representation of shape characteristic in Hough space,put forward the algorithm based on shape characteristic and Inverse Hough Transform (IHT) to locate and reconstruct license plate. We elaborated the principle and application of IHT,and on this basis we got rid of the disturbance in Hough space and reconstructed license plate in image space in the end. Moreover,aiming at the location of inclined license plate,we have proposed the algorithm based on gray changing characteristic and Orientation Field. In this section,we discussed in detail the application of Mathematical Morphology operator in rough locating objective region and introduced the theory and realization method of Orientation Field. Then we used it to detect inclined angel of objective region,finally we can precisely locate and emendate the inclined license plate based on detection result,and so we provided the favorable foundation for segmenting character.更多还原