车牌识别系统外文文献
车牌识别系统毕业论文
车牌识别系统毕业论文论文(设计)题目车牌识别系统——车辆牌照定位系统的设计与实现院系名称计算机科学与技术系专业（班级）计算机科学与技术摘要车牌识别系统作为智能交通系统的一个重要组成部分，在交通监控中占有很重要的地位。
车牌识别系统可分为图像预处理、车牌定位、字符识别3个部分，其中车牌定位作为获得车辆牌照图像的重要步骤，是后续的字符识别部分能否正确识别车牌字符的关键环节。
车牌定位系统实现对车辆牌照进行定位的功能,即从包含整个车辆的图像中找到车牌区域的位置，并对该车牌区域进行定位显示，将定位信息提供给字符识别部分。
在本文中作者分析出车辆牌照具有如下特征：（1）具有固定的长宽比；（2）车牌区域内部字符数目固定；（3）字符与背景之间存在很大的颜色差别；（4）对于含有车牌信息的灰度图像，其车牌区域边缘明显，灰度跳变大，相对于车牌以外区域，具有明显的特征等。
所以，一般基于图像处理的车牌定位系统是通过分析车辆牌照的某些特征来进行定位的。
针对车牌本身固有的特征，本文首先介绍了在车牌定位过程中常用的几种数字图像处理技术：图像的二值化处理、边缘检测和图像增强等。
其次介绍了现在常用的车牌定位方法，并对这些方法进行分析，总结出各种方法的优缺点，然后在此基础上提出采用带边缘检测的灰度图像行扫描投影方法对车牌进行定位，并使用VC++6.0编码实现车牌定位系统。
最后对该系统进行了测试，测试结果表明该系统具有良好的人机交互方式，具有较高的识别正确率和较快的识别速度，对用户给定的待测图像能够迅速准确地进行车辆牌照的定位并将定位结果显示给用户，该系统具有一定的实用价值。
关键词：车牌定位，灰度图像，行扫描，投影AbstractAs an important part of the Intelligent Transportation Systems, License Plate Recognition System plays an important role in traffic monitoring area. License plate recognition system can be divided into three parts, i.e., image pre-processing, license plate location and character recognition. The vehicle license plate location is an important procedure which is used to obtain a license image. It is also the key of the following character recognition system which can identify the correct license plate characters. License plate location system can perform the vehicle license location function, i.e., finding the location of the vehicle license in the image containing the entire vehicle license plate, positioning the plate region and then demonstrating the location information on the computer screen which will be transferred to the character recognition system.In this thesis, the author analyzes the vehicle license and finds that it has the following characteristics: (1) Fixed aspect ratio. (2) Fixed license plate characters number. (3) Great color difference between characters and background.(4) Obvious edge and great intensity change for grayscale images with registration information, and obvious characteristics compared with the outer plate region. Therefore, the majority of image-based positioning systems perform location function by analyzing some characteristics of the vehicle license.According to the own inherent characteristics of license plate, this thesis introduces many commonly used digital image processing techniques in the location process of license plate: binary image processing, edge detection and image enhancement, and so on. Then, we introduce the commonly used methods of license plate location. Further, we analysis these methods and summarize their advantages and disadvantages. Moreover, we propose locating plate by using the gray-scale image projection and line scanning method with edge detection. This system was implemented by using the VC++ 6.0. Finally, the experimental results indicate that the system has a good human-computer interaction, a better identification rate and higher speed. For images provided by users, the system can quickly and accurately locate the vehicle license and display the location results to the users. Therefore, this system has some practical values.Key words: license plate location, gray-scale images, line scan, projection目录摘要 (I)Abstract ................................................................................................................................................................ I I 目录 (III)第一章绪论 (1)1.1 课题的来源及意义 (1)1.2 课题主要研究的问题 (1)1.3 系统设计的目标及基本思路 (1)1.3.1 设计目标 (2)1.3.2 基本思路 (2)第二章车牌定位中常用的数字图像处理技术 (3)2.1 汽车牌照的特征 (3)2.2 数字图像处理技术概述 (3)2.3 DIB图像概述 (3)2.4 车牌定位中常用的数字图像处理技术概述 (4)2.4.1 图像二值化 (4)2.4.2 边缘检测 (4)2.4.3 图像增强 (5)第三章车牌定位方法研究 (6)3.1 车牌定位常用方法介绍 (6)3.1.1 基于纹理特征分析的定位方法 (6)3.1.2 基于数学形态学的定位方法 (6)3.1.3 基于边缘检测的定位方法 (6)3.1.4 基于小波分析的定位方法 (6)3.1.5 基于图像彩色信息的定位方法 (6)3.2 基于行扫描灰度跳变分析的车牌定位方法 (7)第四章车牌定位系统的设计与实现 (8)4.1 车牌定位系统系统分析 (8)4.1.1系统业务需求 (8)4.1.2系统用户需求 (8)4.1.3系统功能需求 (8)4.1.4 系统运行环境需求 (8)4.2 车牌定位系统的整体架构设计 (8)4.2.1 系统总体架构 (8)4.2.2 系统技术架构 (9)4.3 车牌定位系统的功能模块划分和实现 (10)4.3.1 系统的功能模块划分 (10)4.3.2 系统的功能模块实现 (11)第五章车牌定位系统的系统测试 (16)5.1 系统测试过程 (16)5.2 系统测试结果 (17)5.3 测试结果分析 (24)第六章技术要点回顾 (26)6.1 难度分析 (26)6.2 主要工作 (26)6.3 应用的主要技术手段 (26)6.4 存在的问题及展望 (27)结论 (28)参考文献 (29)致谢 (30)第一章绪论1.1 课题的来源及意义随着全球各国汽车数量的持续增加，城市的交通状况越来越受到人们的重视。
车牌识别外文文献翻译中英文
外文文献翻译(含：英文原文及中文译文)文献出处：Gao Q, Wang X, Xie G. License Plate Recognition Based On Prior Knowledge[C]// IEEE International Conference on Automation and Logistics. IEEE, 2007:2964-2968.英文原文License Plate Recognition Based On Prior KnowledgeQian Gao, Xinnian Wang and Gongfu XieAbstract - In this paper, a new algorithm based on improved BP (back propagation) neural network for Chinese vehicle license plate recognition (LPR) is described. The proposed approach provides a solution for the vehicle license plates (VLP) which were degraded severely. What it remarkably differs from the traditional methods is the application of prior knowledge of license plate to the procedure of location, segmentation and recognition. Color collocation is used to locate the license plate in the image. Dimensions of each character are constant, which is used to segment the character of VLPs. The Layout of the Chinese VLP is an important feature, which is used to construct a classifier for recognizing. The experimental results show that the improved algorithm is effective under the condition that the license plates were degraded severely.Index Terms - License plate recognition, prior knowledge, vehiclelicense plates, neural network.I. INTRODUCTIONV ehicle License-Plate (VLP) recognition is a very interesting but difficult problem. It is important in a number of applications such as weight-and-speed-limit, red traffic infringement, road surveys and park security [1]. VLP recognition system consists of the plate location, the characters segmentation, and the characters recognition. These tasks become more sophisticated when dealing with plate images taken in various inclined angles or under various lighting, weather condition and cleanliness of the plate. Because this problem is usually used in real-time systems, it requires not only accuracy but also fast processing. Most existing VLP recognition methods [2], [3], [4], [5] reduce the complexity and increase the recognition rate by using some specific features of local VLPs and establishing some constrains on the position, distance from the camera to vehicles, and the inclined angles. In addition, neural network was used to increase the recognition rate [6], [7] but the traditional recognition methods seldom consider the prior knowledge of the local VLPs. In this paper, we proposed a new improved learning method of BP algorithm based on specific features of Chinese VLPs. The proposed algorithm overcomes the low speed convergence of BP neural network [8] and remarkable increases the recognition rate especially under the condition that the license plate images were degrade severely.II. SPECIFIC FEA TURES OF CHINESE VLPSA. DimensionsAccording to the guideline for vehicle inspection [9], all license plates must be rectangular and have the dimensions and have all 7 characters written in a single line. Under practical environments, the distance from the camera to vehicles and the inclined angles are constant, so all characters of the license plate have a fixed width, and the distance between the medium axes of two adjoining characters is fixed and the ratio between width and height is nearly constant. Those features can be used to locate the plate and segment the individual character. B. Color collocation of the plateThere are four kinds of color collocation for the Chinese vehicle license plate .These color collocations are shown in table I.TABLE IMoreover, military vehicle and police wagon plates contain a red character which belongs to a specific character set. This feature can be used to improve the recognition rate.C. Layout of the Chinese VLPSThe criterion of the vehicle license plate defines the characters layout of Chinese license plate. All standard license plates contain Chinese characters, numbers and letters which are shown in Fig.1. The first one is a Chinese character which is an abbreviation of Chineseprovinces. The second one is a letter ranging from A to Z except the letter I. The third and fourth ones are letters or numbers. The fifth to seventh ones are numbers ranging from 0 to 9 only. However the first or the seventh ones may be red characters in special plates (as shown in Fig.1). After segmentation process the individual character is extracted. Taking advantage of the layout and color collocation prior knowledge, the individual character will enter one of the classes: abbreviations of Chinese provinces set, letters set, letters or numbers set, number set, special characters set.(a)Typical layout(b) Special characterFig.1 The layout of the Chinese license plateIII. THE PROPOSED ALGORITHMThis algorithm consists of four modules: VLP location, character segmentation, character classification and character recognition. The main steps of the flowchart of LPR system are shown in Fig. 2.Firstly the license plate is located in an input image and characters are segmented. Then every individual character image enters the classifier to decide which class it belongs to, and finally the BP network decides which character the character image represents.A. Preprocessing the license plate1) VLP LocationThis process sufficiently utilizes the color feature such as color collocation, color centers and distribution in the plate region, which are described in section II. These color features can be used to eliminate the disturbance o f the fake plate ’ s regions. The flowchart of the plate location is shown in Fig. 3.Fig.3 The flowchart of the plate location algorithmThe regions which structure and texture similar to the vehicle plate are extracted. The process is described as followed:Here, the Gaussian variance is set to be less than W/3 (W is the character stroke width), so 1P gets its maximum value M at the center of the stroke. After convolution, binarization is performed according to a threshold which equals T * M (T<0.5). Median filter is used to preserve the edge gradient and eliminate isolated noise of the binary image. An N * N rectangle median filter is set, and N represents the odd integer mostly close to W.Morphology closing operation can be used to extract the candidate region. The confidence degree of candidate region for being a license plate is verified according to the aspect ratio and areas. Here, the aspect ratio is set between 1.5 and 4 for the reason of inclination. The prior knowledge of color collocation is used to locate plate region exactly. The locating process of the license plate is shown in Fig. 4.2) Character segmentationThis part presents an algorithm for character segmentation based on prior knowledge, using character width, fixed number of characters, the ratio of height to width of a character, and so on. The flowchart of the character segmentation is shown in Fig. 5.Firstly, preprocess the license the plate image, such as uneven illumination correction, contrast enhancement, incline correction and edge enhancement operations; secondly, eliminating space mark which appears between the second character and the third character; thirdly, merging the segmented fragments of the characters. In China, all standard license plates contain only 7 characters (see Fig. 1). If the number of segmented characters is larger than seven, the merging process must be performed. Table II shows the merging process. Finally, extracting the individual character’ image based on the number and the width of the character. Fig. 6 shows the segmentation results. (a) The incline and broken plate image, (b) the incline and distort plate image, (c)the serious fade plate image, (d) the smut license plate image.where Nf is the number of character segments, MaxF is the number of the license plate, and i is the index of each character segment.The medium point of each segmented character is determined by:(3)where 1i Sis the initial coordinates for the character segment, and 2i S is thefinal coordinate for the character segment. The d istance between two consecutive medium points is calculated by:(4)Fig.6 The segmentation resultsB. Using specific prior knowledge for recognitionThe layout of the Chinese VLP is an important feature (as described in the section II), which can be used to construct a classifier for recognizing. The recognizing procedure adopted conjugate gradient descent fast learning method, which is an improved learning method of BP neural network[10]. Conjugate gradient descent, which employs a series of line searches in weight or parameter space. One picks the first descent direction and moves along that direction until the minimum in error is reached. The second descent direction is then computed: this direction the “ conjugate direction” is the one along which the gr adient does not change its direction will not “ spoil ” the contribution from the previous descent iterations. This algorithm adopted topology 625-35-N as shown in Fig. 7. The size of input value is 625 (25*25 ) and initial weights are with random values, desired output values have the same feature with the input values.As Fig. 7 shows, there is a three-layer network which contains working signal feed forward operation and reverse propagation of error processes. The target parameter is t and the length of network outputvectors is n. Sigmoid is the nonlinear transfer function, weights are initialized with random values, and changed in a direction that will reduce the errors.The algorithm was trained with 1000 images of different background and illumination most of which were degrade severely. After preprocessing process, the individual characters are stored. All characters used for training and testing have the same size (25*25 ).The integrated process for license plate recognition consists of the following steps:1) Feature extractingThe feature vectors from separated character images have direct effects on the recognition rate. Many methods can be used to extract feature of the image samples, e.g. statistics of data at vertical direction, edge and shape, framework and all pixels values. Based on extensive experiments, all pixels values method is used to construct feature vectors. Each character was reshaped into a column of 625 rows’ feature vector. These feature vectors are divided into two categories which can be used for training process and testing process.2) Training modelThe layout of the Chinese VLP is an important feature, which can be used to construct a classifier for training, so five categories are divided. The training process of numbers is shown in Fig. 8.As Fig. 8 shows, firstly the classifier decides the class of the inputfeature vector, and then the feature vector enters the neural network correspondingly. After the training process the optimum parameters of the net are stored for recognition. The training and testing process is summarized in Fig. 9.(a) Training process(b)Testing processFig.9 The recognition process3) Recognizing modelAfter training process there are five nets which were completely trained and the optimum parameters were stored. The untrained feature vectors are used to test the net, the performance of the recognition system is shown in Table III. The license plate recognition system is characterized by the recognition rate which is defined by equation (5).Recognition rate =(number of correctly read characters)/ (number of found characters) (5)IV. COMPARISON OF THE RECOGNITION RA TE WITH OTHER METHODSIn order to evaluate the proposed algorithm, two groups of experiments were conducted. One group is to compare the proposed method with the BP based recognition method [11]. The result is shown in table IV. The other group is to compare the proposed method with themethod based on SVM [12].The result is shown in table V. The same training and test data set are used. The comparison results show that the proposed method performs better than the BP neural network and SVM counterpart.V. CONCLUSIONIn this paper, we adopt a new improved learning method of BP algorithm based on specific features of Chinese VLPs. Color collocation and dimension are used in the preprocessing procedure, which makes location and segmentation more accurate. The Layout of the Chinese VLP is an important feature, which is used to construct a classifier for recognizing and makes the system performs well on scratch and inclined plate images. Experimental results show that the proposed method reduces the error rate and consumes less time. However, it still has a few errors when dealing with specially bad quality plates and characters similar to others. This often takes place among these characters (especially letter and number): 3—8 4—A 8—B D—0.In order to improve the incorrect recognizing problem we try to add template-based model [13] at the end of the neural network.中文译文基于先验知识的车牌识别Qian Gao, Xinnian Wang and Gongfu Xie摘要- 本文介绍了一种基于改进的BP（反向传播）神经网络的中国车牌识别（LPR）算法。
文献综述车牌识别
文献综述1前言近几年来，随着汽车的数量猛增，智能型交通体系(ITS——Intel ligent Tran sporta tion S ystem)便成为未来交通监管系统的主要发展趋势，所谓智能交通系统是在较完善的基础设施(包括道路、港口、机场和通信)之上将先进的信息技术、通信技术、控制技术、传感器、计算机技术和系统综合技术有效的集成，并应用于地面运输系统，从而建立起在大范围内发挥作用的，实时、准确、高效的运输系统[1~2]。
行驶车辆的车牌实时识别尤其是智能运输系统研究的重要组成部分。
车牌识别系统是对公路上配置的摄像头拍摄的照片进行数字图像处理与分析，综合应用大量的图像处理最新成果和数学形态学方法对汽车图像进行平滑、二值化、模糊处理、边缘检测、图像分割、开运算、比运算、区域标识等，利用多种手段以提取车牌区域，进而达到对汽车牌照的精确定位并最终完成对汽车牌照的识别。
车牌识别系统的用途很多，如高速公路电子收费站、公路流量控制、公路稽查、失窃车辆查询、监测黑牌机动车、监控违章车辆的电子警察等公路监管场合，以及停车场车辆管理、出入控制等需要车牌认证的场合都要应用车牌识别系统，尤其在高速公路收费系统中，实现不停车收费技术可提高公路系统的运行效率，由此可见车牌识别系统具有不可替代的作用，因此对车牌识别技术的研究和应用系统的开发具有重要的现实意义。
2 车牌识别技术研究现状车牌识别系统要综合应用多种手段提取车牌区域，对汽车牌照的精确定位并最终完成对汽车牌照的识别。
因此车牌识别系统要应对多种复杂环境，如车流量高峰期、照射反光、车牌污染等。
利用模拟人脑智能A NN，在识别车牌时能进行联想记忆与推理，能够较好地解决字符残缺不完整而无法识别的问题。
实时的车牌识别系统 中英文
VISL 项目在完成了02年一种实时车牌识别（LPR）的系统由酒吧，母鸡罗恩指导单位约哈难埃雷兹该系统一个典型的模式：摘要这个项目的目的是建立从汽车板在门入口处时，例如A区牌照时停车一个真正的应用程序，它已承认。
该系统具有视频摄像机的普通PC机，渔获量的视频帧，其中包括一个明显的汽车牌照和处理它们。
一旦发现车牌，它的数字确认，并显示在用户界面或数据库核对一。
形象的重点是设计一个单一的算法车牌从用于提取，分离板的特点及识别单个字符。
背景：目前已在实验室过去类似的项目。
包括项目实施的整个系统。
这个项目的目的首先是改善方案的准确度，并尽可能其时间复杂度。
该实验室的所有项目在过去。
根据精度不佳的测试中，我们就程序设置的45个影像，我们用我们的成功，并只有在特定的条件感到满意。
出于这个原因，除了再次从非常罕见的情况下，整个程序写。
简要说明执行情况：我们的车牌识别系统可大致分为以下框图。
框图全球系统。
另外这个进程可以被看作是减少或地方的牌照抑制有害信息从携带信息的信号，这里是一个视频序列包含大量无关信息的特点，形式抽象符号的研究。
光学字符识别（OCR）已采用神经网络技术，采用神经元在输出层的前馈网络的3层，200个神经元在20输入层，中间神经元在10层，。
我们保留了神经网络数据集图像用在项目的先例，其中包括238位第我们的算法的详细步骤说明如下图：框图程序的子系统。
这里介绍捕获帧的一个给定的产出上面所述的主要步骤：示例捕获帧黄色区域捕获的帧过滤捕获帧地区扩张黄色车牌区域确定氡角度的变换板的使用改进的LP地区调整唱片轮廓-列和图调整唱片轮廓-线条和图唱片作物灰度唱片唱片二值化，均衡使用自适应阈值二进制唱片归唱片确定使用的LP水平轮廓图像总和先决行归唱片轮廓调节字符分割使用的山峰到山谷方法扩张型数位影像调整数字图像水平轮廓-线和图调整的数字图像轮廓调整大小的数字图像OCR的数字识别的神经网络方法工具该方案实施开发了基于Matlab。
车牌照识别系统设计与实现毕业设计论文
车牌照识别系统设计与实现Design and Implementation of Car License Plate Recognition System毕业论文（设计）原创性声明本人所呈交的毕业论文（设计）是我在导师的指导下进行的研究工作及取得的研究成果。
据我所知，除文中已经注明引用的内容外，本论文（设计）不包含其他个人已经发表或撰写过的研究成果。
对本论文（设计）的研究做出重要贡献的个人和集体，均已在文中作了明确说明并表示谢意。
作者签名：日期：毕业论文（设计）授权使用说明本论文（设计）作者完全了解**学院有关保留、使用毕业论文（设计）的规定，学校有权保留论文（设计）并向相关部门送交论文（设计）的电子版和纸质版。
有权将论文（设计）用于非赢利目的的少量复制并允许论文（设计）进入学校图书馆被查阅。
学校可以公布论文（设计）的全部或部分内容。
保密的论文（设计）在解密后适用本规定。
作者签名：指导教师签名：日期：日期：注意事项1.设计（论文）的内容包括：1）封面（按教务处制定的标准封面格式制作）2）原创性声明3）中文摘要（300字左右）、关键词4）外文摘要、关键词5）目次页（附件不统一编入）6）论文主体部分：引言（或绪论）、正文、结论7）参考文献8）致谢9）附录（对论文支持必要时）2.论文字数要求：理工类设计（论文）正文字数不少于1万字（不包括图纸、程序清单等），文科类论文正文字数不少于1.2万字。
3.附件包括：任务书、开题报告、外文译文、译文原文（复印件）。
4.文字、图表要求：1）文字通顺，语言流畅，书写字迹工整，打印字体及大小符合要求，无错别字，不准请他人代写2）工程设计类题目的图纸，要求部分用尺规绘制，部分用计算机绘制，所有图纸应符合国家技术标准规范。
图表整洁，布局合理，文字注释必须使用工程字书写，不准用徒手画3）毕业论文须用A4单面打印，论文50页以上的双面打印4）图表应绘制于无格子的页面上5）软件工程类课题应有程序清单，并提供电子文档5.装订顺序1）设计（论文）2）附件：按照任务书、开题报告、外文译文、译文原文（复印件）次序装订3）其它摘要汽车牌照自动识别系统是智能交通系统的重要组成部分，是高科技的公路交通监控管理系统的主要功能模块之一，汽车牌照识别技术的研究有重要的现实应用意义。
汽车车牌识别系统毕业论文(带外文翻译)解析
汽车车牌识别系统---车牌定位子系统的设计与实现摘要汽车车牌识别系统是近几年发展起来的计算机视觉和模式识别技术在智能交通领域应用的重要研究课题之一。
在车牌自动识别系统中，首先要将车牌从所获取的图像中分割出来实现车牌定位，这是进行车牌字符识别的重要步骤，定位的准确与否直接影响车牌识别率。
本次毕业设计首先对车牌识别系统的现状和已有的技术进行了深入的研究，在此基础上设计并开发了一个基于MATLAB的车牌定位系统，通过编写MATLAB文件，对各种车辆图像处理方法进行分析、比较,最终确定了车牌预处理、车牌粗定位和精定位的方法。
本次设计采取的是基于微分的边缘检测，先从经过边缘提取后的车辆图像中提取车牌特征，进行分析处理，从而初步定出车牌的区域，再利用车牌的先验知识和分布特征对车牌区域二值化图像进行处理，从而得到车牌的精确区域，并且取得了较好的定位结果。
关键词：图像采集，图像预处理，边缘检测，二值化，车牌定位ENGLISH SUBJECTABSTRACTThe subject of the automatic recognition of license plate is one of the most significant subjects that are improved from the connection of computer vision and pattern recognition. In LPSR, the first step is for locating the license plate in the captured image which is very important for character recognition. The recognition correction rate of license plate is governed by accurate degree of license plate location.Firstly, the paper gives a deep research on the status and technique of the plate license recognition system. On the basis of research, a solution of plate license recognition system is proposed through the software MATLAB,by the M-files several of methods in image manipulation are compared and analyzed. The methods based on edge map and das differential analysis is used in the process of the localization of the license plate,extracting the characteristics of the license plate in the car images after being checked up for the edge, and then analyzing and processing until the probably area of license plate is extracted，then come out the resolutions for localization of the car plate.KEY WORDS：imageacquisition，image preprocessing，edge detection，binarization，licence，license plate location目录前言 (1)第1章绪论 (2)§1.1 课题研究的背景 (2)§1.2 车牌的特征 (2)§1.3 国内外车辆牌照识别技术现状 (3)§1.4车牌识别技术的应用情况 (4)§1.5 车牌识别技术的发展趋势 (5)§1.6车牌定位的意义 (6)第2章MATLAB简介 (7)§2.1 MATLAB发展历史 (7)§2.2 MATLAB的语言特点 (7)第3章图像预处理 (10)§3.1 灰度变换 (10)§3.2 图像增强 (11)§3. 3 图像边缘提取及二值化 (13)§3. 4 形态学滤波 (18)第4章车牌定位 (21)§4.1 车牌定位的主要方法 (21)§4.1.1基于直线检测的方法 (22)§4.1.2 基于阈值化的方法 (22)§4.1.3 基于灰度边缘检测方法 (22)§4.1.4 基于彩色图像的车牌定位方法 (25)§4.2 车牌提取 (26)结论 (30)参考文献 (31)致谢 (33)前言随着交通问题的日益严重，智能交通系统应运而生。
汽车牌照识别系统的车牌定位技术研究外文资料翻译(适用于毕业论文外文翻译+中英文对照)
建立一个自动车辆车牌识别系统车辆由于数量庞大的抽象,现代化的城市要建立有效的交通自动系统管理和调度.最有用的系统之一是车辆车牌识别系统,它能自动捕获车辆图像和阅读这些板块的号码在本文中，我们提出一个自动心室晚电位识别系统，ISeeCarRecognizer,阅读越南样颗粒在交通费的注册号码.我们的系统包括三个主要模块：心室晚电位检测，板数分割和车牌号码识别。
在心室晚电位检测模块，我们提出一个有效的边界线为基础Hough变换相结合的方法和轮廓算法.该方法优化速度和准确性处理图像取自不同职位。
然后,我们使用水平和垂直投影的车牌号码分开心室晚电位分段模块.最后，每个车牌号码将被OCR的识别模块实现了由隐马尔可夫模型。
该系统在两个形象评价实证套并证明其有效性是适用于实际交通收费系统。
该系统也可适用于轻微改变一些其他类型的病毒样颗粒。
一.导言车牌识别的问题是一个非常有趣，且困难的一个问题.这在许多交通管理系统中是非常有用的。
心室晚电位识别需要一些复杂的任务,如车牌的检测，分割和识别。
这些任务变得更加复杂时，处理各种倾斜角度拍摄的图像或含有噪音的图像。
由于此问题通常是在实时系统中使用，它不仅需要准确性,而且要效率.大多数心室晚电位识别应用通过建立减少一些复杂的约束的位置和距离相机车辆，倾斜角度。
通过这种方式，车牌识别系统的识别率已得到明显改善.在此外，我们可以更准确地获得通过一些具体的当地样颗粒的功能，如字符数，行数在一板,或板的背景颜色或的宽度比为一板高。
二.相关工作心室晚电位的自动识别问题在20世纪90年代开始就有研究。
第一种方法是基于特征的边界线。
首次输入图像处理，以丰富的边界线的一些信息如梯度算法过滤器,导致在一边缘图像.这张照片是二值化处理，然后用某些算法，如Hough 变换，检测线。
最终,2平行线视为板候选人［4] ［5］。
另一种方法是基于形态学[2］。
这种方法侧重于一些板块图像性质如亮度，对称，角度等。
车牌识别系统的研究背景意义及国内外研究现状
车牌识别系统的研究背景意义及国内外研究现状1车牌识别系统的背景1.1 车牌识别系统的背景及研究意义1.2 车牌识别系统简介2 车牌识别系统的国内外现状3车牌识别难点1车牌识别系统的背景1.1 车牌识别系统的背景及研究意义随着经济社会的迅猛发展，人们的生活水平的提高，机动车辆的数量也越来越多。
为了提高车辆的管理效率，缓解公路上的交通压力，我们必须找到一种解决方案。
而作为汽车“身份证”的汽车车牌，是在公众场合能够唯一确定汽车身份的凭证。
我们可以以此为依据，设计一种车牌识别系统监控各个车辆的情况。
为此，我国交通管理部门对汽车车牌的管理非常重视并制定了一套严格的管理法规。
其中对汽车车牌的制作、安装、维护都要求由制定部门统一进行管理。
在此基础上，如果研制出一种能在公众场合迅速准确地对汽车牌照进行自动定位识别的系统(CPR)，那么这将是一件非常有意义的工作，并将极大地提高汽车的安全管理水平及管理效率。
车辆牌照定位与识别是计算机视觉与模式识别技术在智能交通领域应用的重要研究课题之一, 该技术应用范围非常广泛, 其中包括: (1) 交通流量检测；(2)交通控制与诱导；(3) 机场、港口等出入口车辆管理；(4) 小区车辆管理；(5) 闯红灯等违章车辆监控；(6) 不停车自动收费；(7) 道口检查站车辆监控；(8) 公共停车场安全防盗管理；(9) 计算出行时间；(10) 车辆安全防盗、查堵指定车辆等。
其潜在市场应用价值极大，有能力产生巨大的社会效益和经济效益。
如图1所示，LPR[1]的部分应用：图1 LPR在收费口、道路监控和停车管理中的应用近些年,计算机的飞速发展和数字图像技术的日趋成熟，为传统的交通管理带来重大转变。
先进的计算机处理技术，不但可以将人力从繁琐的人工观察、检测中解放出来，而且能够大大提高其精确度，汽车牌照自动识别系统就是在这样的背景与目的下进行开发的。
汽车牌照自动识别系统(VLPRS)是对由公路上配置的摄像头拍摄的照片进行数字图像处理与分析，综合应用大量的图像处理最新成果和数学形态学方法对汽车图像进行平滑、二值化、模糊处理、边缘检测、图像分割、开运算、闭运算、区域标识等多种手段以提取车牌区域,进而达到对汽车牌照的精确定位并最终完成对汽车牌照的识别。
车辆图像预处理和车牌定位的方法研究 中英文对照翻译
The Method Research of Vehicle Image Preprocessing and LicensePlate LocationAbstract—Aiming at the characteristics of vehicle images, this paper presents a method about vehicle image preprocessing and license plate location. The image preprocessing mainly includes graying the image, detecting the edge on image, median filtering and taking binaryzationon image. The license plate location consists of locating upper and lower boundary, locating left and right boundary for the image after preprocessing and finally the license plate region located. All the above is the foundation for the subsequent license plate recognition. A large number of experiments have proved that this method has the good image preprocessing effect, high accuracy rate, location speed and the good practical value.Keywords-Preprocessing; Median filtering; Binaryzation; Edge detection; License plate locationI. I NTRODUCTIONVehicle license plate recognition system based on vehicle license for the specific target is dedicated computer vision system[1]. It is one of the important research topics about computer vision and pattern recognition technology in the field of intelligent transportation applications. Vehicle license identification is the general composed by the following process: image acquisition, image preprocessing, license plate location, character segmentation, character recognition[2]. The correct rate of the last process has a direct impact on the next process. Since the original image from the acquisition card includes the vehicle license plate, the car itself, and automotive background image, it is necessary to remove these non-licensed images in order to extract the correct regional license and for the foundation of the license plate character recognition . In the actual system, due to natural changes in day and night illumination, vehicle own movement, the camera angle of observation, collecting images of the equipment itself and other factors influence, the image obtained is not always very satisfactory, there is a wide range of noise. Therefore, it is necessary to make the license plate image pre-processing for improving image quality, layingfoundation for the subsequent license plate recognition.II. I MAGE PREPROCESSINGImage preprocessing is an essential process in license plate recognition system, and the quality of preprocessing directly affects the location. The image preprocessing in this article includes image grey, edge detection, median filter andbinaryzation. The following gives the detail statement on the preprocessing process.A. Image greyingAll vehicle images acquired through camera and image card are color image, and image format is not same. The commonly used image forms are JPEG and BMP. If treatment with acquired image directly, not only the image format is complex, moreover the computation data quantity is extremely huge, such license plate location cannot satisfy the request for fast and real-time. Therefore, the color image need to be formatted processing, transforming the JPEG image or the BMP image to DIB (Device Independent Bitmap) which favors the computer to process. Then, using R, G and B tricolor weighted average method process the DIB image, processing function is shown as equation (1):F(x, y) =0.299*R(x, y) +0.587*G(x, y) + 0.114*B(x, y) (1)R(x, y), G(x, y) and B(x, y) are R, G and B tricolor component of the input color image [3]. The color image transforms to grey image by equation (1) processing, the result is shown in Fig. 1.(a) Original image(b) Grey imageFigure 1. A contrast between original image and grey imageB. Edge detectionEdge is the most basic feature of the image, so the edge indicates step change of the grey level on its surrounding pixels. Chinese vehicles license plate region has big color contrast between license plate bottom and license plate character. The license plate is composed of 7 characters with rich edge information in turn which are the Chinese character, the letter and the Arabic numeral, and the character in license plate region and background have obvious edge in the entire picture, also have many edge. This is one of the basic characteristics that license plate region distinguishes from other region in the vehicles picture, and it is also the fundamental basis of this algorithm. Commonly used edge detection operators have Prewitt operator, Sobel operator, Canny operator, LOG operator, Roberts operator and other operator. Prewitt operator and Sobel operator are first-order differential operator, the former is the average filter, the latter is the weighted average filter ，while the image edge detected by the two methods may be better than two pixels. The Canny method uses first derivative as the foundation to judge edge points. It is one of the best traditional first-order differential operators in the detection of step edge. The shortcoming is smoothing out some details [4]. LOG operator uses Gaussian function to smooth image first, then uses Laplace transform to process image, and this method processing image edge is insufficiently clear and the speed needs to be improved. The localization using Roberts operator is quite precise, but more sensitive to noise. The experiment indicated that using the Prewitt edge detection operator can better stand out the edge characteristic of license plate, and speed is faster. Fig. 2 is several imagesafter process by different edge detection operator.(a) Prewitt operator(b) Sobel operator(c) Canny operator(d) LOG operator(e) Robert operatorFigure 2.Image contrast after process by several edge detection operator C. Median filteringMedian filtering method is a non-linear smoothing technique. It sets the grey level of each pixel to the middle value of all pixels’ grey level in a neighborhood window [5]. Median filtering method is a non-linear technique that based on a sequencing statistic theory. It can inhibit the noise effectively. The basic principle of median filtering is to replace the value of point in digital image or numerical sequences with the middle value of this point’s one neighborhood, so it can let the around pixel value close to this point’s value. Thus the isolated noise is eliminated. This method utilizes the two-dimensional sliding template of a certain structure, arranges the pixel in template according to the size of pixel value, then a rise (or drop) two-dimensional data array was produced. The output of two-dimensional median filtering result provided by equation (2):G(x, y) =med {F(x-k, y-l)} (2)F(x, y), G(x, y) is respectively for original image and the image after dealing with. W is a two-dimensional template. The result is shown in Fig. 3.Figure 3. Median filter imageD. Image binaryzationImage binaryzation processing is that setting gray value of pixels on image to 0 or 255, that is, the entire image presents tangible black and white effect [6]. We obtain the binaryzation image through selecting 256 brightness level of grey image by suitable threshold. Binaryzation image can still reflect the whole and partial characteristic of image. In digital image processing, binaryzation image holds the extremely important status. First, binaryzation reduces the amount of image data. Secondly, to highlight the outline of the interest goal, this is favor to further processing. Fig. 4 is the image after binaryzation processing.Figure 4. Image after binaryzation processingIII. LICENSE PLATE LOCATIONThe task of license plate location is to remove most unwanted background information from the whole image and find the license plate region with a small amount of redundant background. Because the license plate region contrasts to the background, the histogram of license plate image shows a bimodal shape after image preprocessing. The wave trough between two wave peaks corresponding to the gray level is selected as a threshold. Supposing the image is divided by F(x, y) and the gray level range is [Z1, Zk]. Fig. 5 shows there are two obvious wave peaks in gray levelZi and Zj, and in Zt there is a wave trough. By choosing Zt reasonably, B1 belt can contain grey level correlation to the background as far as possible, while the B2 band includes grey level correlation to the license plate as far as possible [7].Figure 5. Double peak of histogramA. Locating upper and lower boundaryOne characteristic of license plate image is the crowded characters in the internal, so the grey jump is extremely fierce. We find the possible location of license plate region by using row grey jump rule of grey image. We preserve this position and call it as the fake license plate region. Specific algorithm including following steps:Step1：Calculating the level histogram of image, and smoothing the level histogram with [1,1,1,1,1] / 5 operator.Step2：Searching the bottom edge distance of license plate from the image base, if 5 line which is predefined as 5 has satisfied the request continuously which the value is bigger than or equal to 10 pixels in histogram, and the value of current line differs above 4 pixels with the value of front the Nth line, then we believe that the bottom edge distance of license plate has founded. Current line minus 5, and locates the scan line to the summit of current peak. If the current line does not satisfy the condition, then continues to search upwardly until the top margin of image.Step3：Locating the current line to the bottom of up wave crest, if the peak bottom value is greater than the maximum value, then locating to the summit of current peak, and the summit for maximum value line; searching upwardly from the current line’s next line, if the value of search line is greater than the recorded maximum value, then setting the current line as maximum value and carrying on searching upwardly from it. Otherwise, if the current value is smaller than two-thirds of maximum value, or the current value is less than 5 pixels, or the license plate’s height is greater than 80 pixels,then we think the top margin of license plate has been founded.Step4：Check whether the license plate’s height complies with the requirement or not. If the license plate’s height is smaller than 40 pixels to continue search upwardly, otherwise the license plate region has been found, and precision positioning is from up and down location of license plate.After above 4 steps searching, the upper and lower boundary of license plate has been found. Location result is shown as in Fig. 6.Fig. 6. Locating upper and lower boundaryB. Locating left and right boundaryWe can find the left and right boundary by the rule of character change in license plate. Specific algorithm including following steps:Step1：To the vertical histogram, scanning from left to right, the points less than 4 pixels in the histogram are removed firstly.Step2：The current line is written for the nLeft, as the beginning of the peak.Step3：Adding the rows that greater than or equal to 4 pixels in the cumulative histogram up, and recording it as the width of peak: Pixel1Wide.Step4：Adding the rows that smaller than or equal to 4 pixels in the cumulative histogram up, and recording it as the width of trough: Pixel0Wide. If the width of current peak is less than 4 and the average peak height is less than one-sixth of height, and the supreme value is less than a quarter of height, then merging this peak into the trough of upper peak, and to Step 1.Step5：WaveCrestCount adds 1. WaveCrestCount is the number of peak.Step6：Repeating Step 1 until the current row number is greater than the image width. Step7：Statistic all peaks, when seven consecutive width of wave trough is smaller than the height of license plate, the wave peak of left side is regarded as the left edge distance of license plate.Step8：Counting backward continually, until meeting a width of wave trough is greater than the height of license plate. The start point of wave trough is regarded as the rightedge distance of the license plate.The result of locating left and right boundary is shown as in Fig. 7.Fig. 7. Locating left and right boundaryIV. CONCLUSIONSThis paper mainly researches the preprocessing of license plate image and license plate location. The preprocessing not only removes the noise in the image but also processes edge detection to the license plate image. After preprocessing, according to the characteristic of license plate image and the regularity of grey change，the boundary of license plate is located. The use of the methods are proposed in this article, in a variety of weather conditions and under the conditions of different backgrounds 200 license plate images are collected and implemented the automatic positioning of the plate. The method can be more rapid and effective to identify the license plate from the complex background noise. Its feature detection has good anti-interference effect, can meet the real-time system's demands and has good application prospects.车辆图像预处理和车牌定位的方法研究摘要—针对车辆图像的特征，本文提出了一种车辆图像预处理和车牌定位的方法。
基于MATLAB的车牌识别系统研究(课设参考文献)
1.2.3 车牌识别技术的发展趋势
5
上海交通大学硕士学位论文
绪论
车牌识别技术作为智能交通系统中的关键技术，在各国学者的共同努力下，已 经得到了长足的发展，并且已经得到了不同程度的实际应用，但目前还存在着种种 不足。
对于未来车牌识别产品的技术发展趋势， 汉王科 技智能 交通部 总经理 乔炬认 为。首先，由于市场需求不同，对识别产品的需求也有差异，因此就要求研发针对 不同细分市场的车牌识别产品。其次，随着算法的不断改进，基于视频触发技术的 车牌识别产品将得到大范围的应用，但是视频触发技术取代外触发装置尚需时日。 第三，现在的车牌识别系统设备过多，系统集成难度大，系统稳定性差，系统维护 是一个让人头疼的问题。随着技术不断进步，以往多个设备实现的功能可能由一个 设备实现。
为基础的车牌识别系统，识别率分别为 81.25%、85%、91.25%。日本对车牌图像的 获取也做了大量的研究，并为系统产业化做了大量工作。Luis [4]开发的系统应绪论
公路收费站，全天识别率达到了 90%以上，即使在天气不好的情况下也达到了 70%。 国外对车牌识别的研究起步早，总体来讲其技术已比较领先，同时由于他们车牌种 类单一，规范程度较高，易于定位识别，目前，已经实现了产品化，并在实际的交 通系统中得到了广泛的应用。由于中国车牌的格式与国外有较大差异，所以国外关 于识别率的报道只具有参考价值，其在中国的应用效果可能没有在其国内的应用效 果好，但其识别系统中采用的很多算法具有很好的借鉴意义。
上海交通大学硕士学位论文
绪论
1 绪论
1.1 研究背景
1990 年，美国智能交通学会 CITS America 提出了智能交通系统（ITS）的概念。 目前，智能交通系统已经在世界上经济发达国家的一些城市及高速公路系统中得到 了广泛应用。我国在该领域的研究起步较晚，但随着全球范围智能交通技术研究的 兴起及奥运会的成功举办，智能交通在我国也逐渐进入了应用阶段，相应的，我国 也加快了对智能交通技术研究的步伐，智能交通技术的研究现已进入快速发展期。
车牌识别英文文献2翻译
实时车辆的车牌识别系统摘要本文中阐述的是一个简炼的用于车牌识别系统的算法。
基于模式匹配，该算法可以应用于对车牌实时检测数据采集，测绘或一些特定应用目的。
拟议的系统原型已经使用C++和实验结果已证明认可阿尔伯塔车牌。
1.介绍车辆的车牌识别系统已经成为在视频监控领域中一个特殊的热门领域超过10年左右。
随着先进的用于交通管理应用的视频车辆检测系统的的到来，车牌识别系统被发现可以适合用在相当多的领域内，并非只是控制访问点或收费停车场。
现在它可以被集成到视频车辆检测系统，该系统通常安装在需要的地方用于十字路口控制，交通监控等，以确定该车辆是否违反交通法规或找到被盗车辆。
一些用于识别车牌的技术到目前为止有如BAM（双向联想回忆）神经网络字符识别[1]，模式匹配[2]等技术。
应用于系统的技术是基于模式匹配，该系统快速，准确足以在相应的请求时间内完成，更重要的是在于阿尔伯塔车牌识别在字母和数字方位确认上的优先发展。
由于车牌号码的字体和方位因国家/州/省份的不同而不同，该算法需要作相应的修改保持其结构完整，如果我们想请求系统识别这些地方的车牌。
本文其余部分的组织如下：第2节探讨了在识别过程中涉及的系统的结构和步骤，第3节解释了算法对于车牌号码的实时检测，第4节为实验结果，第5节总结了全文包括致谢和参考文献。
2.系统架构系统将被用来作为十字路口的交通视频监控摄像系统一个组成部分来进行分析。
图1显示了卡尔加里一个典型的交叉口。
只有一个车牌用在艾伯塔，连接到背面的车辆照相机将被用于跟踪此背面车牌。
图1 卡尔加里一个的典型交叉口系统架构包含三个相异部分：室外部分，室内部分和通信链路。
室外部分是安装摄像头在拍摄图像的不同需要的路口。
室内部分是中央控制站，从所有这些安装摄像头中，接收，存储和分析所拍摄图像。
通信链路就是高速电缆或光纤连接到所有这些相机中央控制站。
几乎所有的算法的开发程度迄今按以下类似的步骤进行。
一般的7个处理步骤已被确定为所有号牌识别算法[3] 共有。
车牌识别文献综述
1 前言随着我国汽车产业的飞速发展，大量在公共场合的汽车需要得到监管，为了更好地进行管理，必须对车辆进行一种智能化管理。
所谓的智能交通系统，是指在较完善的基础设施（包括道路，机场）之上将先进的通信技术和计算机技术和系统综合技术有效的集合并应用于地面交通运输系统，从而建立起来在大范围发挥作用的，准确，高速，实时的交通运输系统。
车辆牌照定位与识别是计算机视觉与模式识别技术在智能交通领域应用的重要研究课题之一，该技术应用范围非常广泛，其中包括：（1）交通流量检测；（2）交通控制与诱导；（3）机场、港口等出入口车辆管理；（4）小区车辆管理；（5）闯红灯等违章车辆监控；（6）不停车自动收费；（7）道口检查站车辆监控；（8）公共停车场安全防盗管理；（9）计算出行时间；（10）车辆安全防盗、查堵指定车辆等。
其潜在在市场应用价值极大，有能力产生巨大的社会效益和经济效益。
车牌识别系统作为智能交通系统最重要的子系统之一，是指对公路上配置的摄像头所拍摄的照片进行数字图像处理与分析对汽车图像进行平滑、二值化、模糊处理、边缘检测、图像分割等利用多种手段实现车牌定位，识别，分割最终完成对车牌的识别。
车牌识别系统的用途很多，在高速公路收费站、路口监测(电子警察)、大型停车场等场所具有广阔的发展前景。
由此可见，对车牌自动识别系统的研究有重要的现实意义。
2 国内外研究现状车牌自动识别系统起源于20 世纪80 年代，那时候主要是应用在被盗车辆的检测，还没有形成一套完整的识别系统。
到了80 年代，出现了一些用于车牌自动识别的图像处理方法，那时只是针对一些特定的问题采用简单的图像处理技术来实现。
到了20 世纪90 年代，随着计算机视觉技术的发展以及计算机计算性能的提高，才兴起了车牌的自动识别研究热潮，欧美的一些国家率先开始了车牌识别系统的研究工作。
国外学者具有代表性的研究工作有：A.S.Johnson 等在1990 年提出了利用图像分割来实现车牌自动识别，系统主要由特征提取、模板构造和字符识别三个部分组成，该系统可以实现车牌的自动识别，具有开创性的意义。
车牌识别英文文献1翻译
提出的模型本文的主要目的是要开发一个系统可以从没有虚假质量的复杂的现场图像中提取车牌号码，相机和车牌之间的距离，其中的相对车牌已被抓获在相机等。
对车牌识别系统概述见图1，在车辆图片由相机拍摄后，它会被传递到预先处理单位由系统作进一步处理。
其主要功能是消除图像采集子系统所造成的噪声，提升图像的其他两个子系统使用的功效。
图像将被板提取模块扫描以找到车辆的车牌。
下一阶段是对于车牌中字符的分割。
最后每个字符将传递给光学字符识别（OCR）模块来进行识别确定，最终结果将是ASCII字符和车牌号码。
3.1.预处理输入图像的最初处理是为了提高其质量，并为系统的下一阶段执行作准备。
首先，该系统使用的NTSC标准的方法将RGB图像转换为灰度图像。
G=0.299*R+0.587*G+0.114*B第二步，用中值滤波（5x5）对灰度图像进行处理，以消除噪音，同时也能保持图像的清晰度。
中值滤波是一种非线性滤波器，它用各像素5x5邻里的计算得出中位值来取代该像素的值。
3.2.板块提取板块提取处理包含五个不同的阶段进行，如图2所示；在这里每个阶段执行灰度图像分割过程以消除不属于车牌区域的多余的像素。
例如，水平定位阶段是负责寻找水平部分可能包含一个车牌。
在下面的讨论中，每一个阶段都会被细细展开讨论。
马来西亚车牌由一排白色字符在黑色的背景底色，所以我们可以说，车牌区域的特点是从一排黑色过渡到一排白色，反之亦然，这样转换被称为“边缘”。
从车牌字符到其背景在色彩强度上总的变化叫做边缘的强度。
最强边缘值，能够在从一个黑色像素过渡到一个白色像素或从白色像素变为黑色像素情况下被找到。
在理想的情况下，马来西亚的车牌是白色的字符绘制在黑色的背景上，因此这种搭配产生了高强度边缘值，用于查找出可能的板区域。
在本文中我们将使用Sobel算子来查找边缘。
Sobel运算在图片上执行一个2-D空间梯度测量。
通常它是用来寻找近似绝对的梯度幅度对在每一个点在输入的灰度图像上。
matlab在汽车牌照识别系统中的应用
MATLAB在汽车牌照识别系统中的应用摘要：汽车普及程度的提高使得智能交通技术迅猛发展，汽车牌照识别系统是智能交通管理领域和数字图像处理领域里的热点问题。
汽车牌照识别系统主要由图像预处理，边缘提取，字符定位，字符分割，字符识别几个部分组成。
本文介绍了MATLAB在汽车牌照图像处理识别系统中的部分应用。
该系统通过调用MATLAB函数，综合使用多种方法提高系统的有效识别能力。
关键词：MATLAB 边缘提取字符分割字符识别Application of MATLAB in License Plate Recognition SystemWANG Yanyan（School of Transportation Science and Engineering of BeihangUniversity, Beijing, 100191, China）Abstract: In this paper, CAN bus technology and SAE J1939 protocol are studied and a digital vehicle instrument solution based on Freescale's MC9S12HZ256 MCU is proposed. The message frame format and some engine-re lated parameters’ definition in SAE J1939 application layer protocol are introduced in detail. Stepper motor and its driver, also the methods for speed signal processing are introduced too. The hardware platform of digital vehicle instrument is composed by MCU, signal acquisition module and signal processing anddisplaying module. Data receiving and processing from CAN bus and sensors are accomplished by programming and vehicle condition can be reflected in real-time.Key words:digital vehicle instrument; MC9S12HZ256; stepper motor; CAN bus; SAE J1939引言车牌识别（License Plate Recognition System，LPRS）系统是计算机视觉与模式识别技术在智能交通领域应用的重要研究课题之一，是智能交通系统（Intelligent Transportation System，ITS）的重要组成部分。
《基于改进FasterR-CNN的停车场车牌识别及管理系统的研究与实现》
《基于改进Faster R-CNN的停车场车牌识别及管理系统的研究与实现》一、引言随着智能化和自动化技术的不断发展，停车场车牌识别及管理系统已经成为现代城市交通管理的重要组成部分。
为了满足高效、准确、便捷的停车需求，本文提出了一种基于改进Faster R-CNN的停车场车牌识别及管理系统。
该系统通过深度学习技术，实现了对车牌的快速、准确识别，并配合管理系统，实现了对停车场车辆的高效管理。
二、相关技术概述1. Faster R-CNN：Faster R-CNN是一种用于目标检测的深度学习算法。
它通过改进传统的RCNN系列算法，实现了更高的检测速度和准确率。
在车牌识别领域，Faster R-CNN具有良好的应用前景。
2. 深度学习：深度学习是机器学习的一个分支，通过模拟人脑神经网络的工作方式，实现对复杂数据的分析和处理。
在车牌识别和管理系统中，深度学习技术可以有效地提高识别准确率和系统性能。
三、系统设计与实现1. 车牌识别模块本系统采用改进的Faster R-CNN算法进行车牌识别。
首先，通过卷积神经网络对车牌图像进行特征提取；其次，利用区域推荐网络（RPN）生成可能的车牌区域；最后，通过分类和回归操作，实现对车牌的快速、准确识别。
为了提高识别准确率，我们还采用了数据增强技术，对车牌图像进行预处理和扩充。
2. 管理系统模块管理系统模块主要包括车辆信息管理、停车记录管理、费用结算等功能。
通过与车牌识别模块的接口连接，实现对车辆信息的自动录入和更新。
同时，管理系统还可以根据停车记录和费用结算情况，生成详细的报表和统计数据，方便管理人员进行查询和分析。
3. 系统实现系统实现主要包括软件设计和硬件设备选择。
软件设计采用Python语言和PyTorch框架进行开发，实现了车牌识别的算法和管理系统的功能。
硬件设备主要包括摄像头、计算机等，用于采集车牌图像和处理数据。
四、实验与分析1. 实验环境与数据集实验环境采用高性能计算机，配置了适当的GPU和内存资源。
《基于改进FasterR-CNN的停车场车牌识别及管理系统的研究与实现》
《基于改进Faster R-CNN的停车场车牌识别及管理系统的研究与实现》一、引言随着社会经济的快速发展和汽车保有量的迅速增长，停车场管理成为了一个重要的问题。
车牌识别技术作为停车场管理的核心技术之一，对于提高停车场的运行效率和管理水平具有重要价值。
传统的车牌识别系统通常基于手动输入或者固定模式匹配的方法，这些方法在面对复杂多变的环境时，识别准确率往往不尽如人意。
因此，本研究旨在通过改进Faster R-CNN算法，实现高效、准确的车牌识别及管理系统。
二、研究背景与意义Faster R-CNN是一种广泛应用于目标检测的深度学习算法，其通过改进区域建议网络（RPN）实现了较高的检测速度和准确度。
将Faster R-CNN应用于停车场车牌识别及管理系统，能够显著提高车牌识别的准确率和效率。
本研究将基于Faster R-CNN进行改进，以提高其在复杂环境下的车牌识别能力，为停车场管理提供更加高效、智能的解决方案。
三、系统设计与实现（一）系统架构设计本系统主要包含数据采集、预处理、改进Faster R-CNN车牌识别及管理模块。
数据采集模块负责收集停车场中的车牌图像；预处理模块负责对图像进行去噪、增强等处理；改进Faster R-CNN模块则负责车牌的检测与识别；管理模块则负责将识别结果进行整合，并实现车辆进出场管理等功能。
（二）算法改进与优化本研究通过引入深度可分离卷积、轻量化网络结构等技术，对Faster R-CNN算法进行改进。
这些改进措施可以在保证识别准确率的同时，降低模型的计算复杂度，提高系统的实时性能。
此外，针对停车场环境下的车牌特点，我们还对算法的鲁棒性进行了优化，使其能够更好地适应光照变化、遮挡、模糊等复杂环境。
（三）系统实现与测试系统实现过程中，我们采用了Python作为主要编程语言，使用TensorFlow等深度学习框架实现Faster R-CNN的改进模型。
通过对大量停车场车牌图像进行训练和测试，我们验证了系统的有效性和准确性。
《基于改进FasterR-CNN的停车场车牌识别及管理系统的研究与实现》
《基于改进Faster R-CNN的停车场车牌识别及管理系统的研究与实现》一、引言随着智能化技术的不断发展，停车场车牌识别与管理系统逐渐成为智能交通系统的重要组成部分。
本文提出了一种基于改进Faster R-CNN的停车场车牌识别及管理系统，通过深度学习和计算机视觉技术，实现对停车场车牌的快速、准确识别，并实现高效的车牌管理。
二、背景及意义传统的车牌识别系统主要依赖于人工输入或简单的图像处理技术，存在识别速度慢、准确率低等问题。
而基于改进Faster R-CNN的停车场车牌识别及管理系统，通过深度学习技术，能够自动学习和提取车牌特征，实现快速、准确的识别。
此外，该系统还能实现车牌信息的高效管理，为停车场管理提供有力支持。
本文的研究与实现，对于提高停车场车牌识别的准确性和效率，推动智能交通系统的发展具有重要意义。
三、相关技术3.1 Faster R-CNNFaster R-CNN是一种常用的目标检测算法，具有较高的检测速度和准确率。
本文在Faster R-CNN的基础上进行改进，以提高车牌识别的性能。
3.2 深度学习深度学习是机器学习的一个重要领域，通过模拟人脑神经网络的工作方式，实现对复杂数据的分析和处理。
本文利用深度学习技术，实现对车牌特征的自动学习和提取。
四、系统设计与实现4.1 系统架构本系统采用分层架构设计，包括数据采集层、数据处理层、模型训练层和应用层。
数据采集层负责采集停车场车牌图像；数据处理层负责对图像进行预处理和特征提取；模型训练层利用深度学习技术训练车牌识别模型；应用层负责实现车牌识别及管理功能。
4.2 改进Faster R-CNN模型本文在Faster R-CNN的基础上，通过优化网络结构、调整参数等方式，提高模型的识别性能。
具体包括：引入更深的网络结构、改进特征提取方法、优化损失函数等。
4.3 车牌识别流程车牌识别流程包括图像预处理、车牌定位、特征提取和识别四个步骤。
图像预处理包括灰度化、二值化等操作；车牌定位通过改进Faster R-CNN模型实现；特征提取和识别则由深度学习模型完成。
自动车牌识别系统在英国的应用
自动车牌识别系统在英国的应用自动车牌识别(ANPR)这个概念最早出现在60年代和70年代早期，那时候采用的是光学字符识别（OCR）技术，就是用一台大型计算机利用磁性油墨来识别字符。
这些光学字符识别（OCR）字体在支票簿上已经使用了许多年.同样的技术还可以用来阅读车牌，但是第一台“实时”视频帧捕捉器直到80年代初期才开发出来。
第一套商业系统是由Cambridge Neurodynamics在90年代开发的，它依据的是神经网络方法（一种自我学习程序）。
最初的认为是，ANPR将主要对警察和高速公路应用领域带来好处，用于威慑恐怖袭击或者用它作为监控交通或识别超速车辆的工具。
英国第一套ANPR系统是在1997年由伦敦市警察局引入的。
安装的照相机用于监控进入或离开伦敦的车辆，主要目的就是为公众提供一个安全的环境。
今天，许多警察部队和港务局都在使用着ANPR系统。
ANPR正在成为一种被广泛接受的追踪车辆移动的工具。
怀疑论者仍然怀疑该系统的可靠性，但是精确度已经变得很出色。
在今天这个处理能力和视频帧捕捉技术快速发展的世界，从捕捉到图像处理的速度都很出色。
工作原理Blick Alfia公司使用的方法是智能模板比对法，它的工作原理是反复扫描一个图象，从而找出与预先设定的模板相匹配的形状。
如果扫描的是车牌，首先要找到一个矩形。
一旦找出这个形状，我们再使用模板找出与车牌上的内容形状类似的字母或数字字符。
已经建立的模板可以识别大多数的字符或字体，包括外国车牌或车牌上的其他可识别的字符。
进行这种比对的速度由个人计算机的处理能力决定。
我们相信，模板方法比神经网络方法更适合停车场领域的应用，而且从一开始就能确保更高的精确度，尤其是在停车场环境中，因为在被引导通过一个栏杆机时，我们经常要放慢行车速度，甚至要把车停住。
使用的照相机根据地点状况会有所不同。
例如在一个始终有光线的地下停车场，可使用标准的CCTV黑白照相机。
而在光线不断变化的地方(像露天平面停车场)，最好使用带红外源的智能照相机，红外线用于照亮反光的车牌，这样的话，无论在漆黑的晚上还是明亮的白天，不需要周围光线就能读取车牌。
- 1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。
Vehicle License Plate Recognition System Based on Digital Image ProcessingYao Yuan,Wu xiao-liDepartment of Computer Science and Engineering, Henan University of Urban Constructioneyaoyuan@Akf1l'll c l-This paper analyzes the basic method of digital videoimage processing, studies the vehicle license plate recognition system based on image processing in intelligent transport system, presents a character recognition approach based on neural network perceptron to solve the vehicle license plate recognition in real-time traffic flow. Experimental results show that the approach can achieve better positioning effect, has a certain robustness and timeliness.Keywonls-veltic/e license plate recognition; imllge processing,· "igilll/I'fJ ltologyI. INTRODUCTIONSince the 21st century, with social development and improvement of living standards, the number of vehicles is continuously increased, the traffic conditions is worsening, which brought huge pressures to the society and environment. Intelligent transport system is a real-time, accurate, and efficient transportation management system built based on a relatively perfect road infrastructure and by synthetically using the advanced electronic technology, information technology, sensor technology and systemic engineering technology in ground transportation[1] . d management . This system can solve the vanous roaproblems generated by the traffic congestion, thus receiving more and more attention.Vehicle license plate recognition is one of the key technologies in the intelligent transport system, while its development is rapid, has been gradually integrated into our real life. Vehicle license plate recognition system can carry out automatic registration, verification, monitoring and alarm management, is an important part of modem highway toll management system, highway speed automatic monitoring system, highway surveillance, parking automatic charging management and other fields.II. VEHICLE LICENSE PLATE RECOGNITION SYSTEM A. Vehicle license p late recognition system overviewVehicle license plate recognition system is mainly composed by hardware and software. The hardware part includes a control computer, one Ethernet camera, a UPS power supply and an interface control port. These sections ensure the car images intake and processing. The software part is divided into the Ethernet camera embedded front-end software and the processing software in the industrial computer.Vehicle license plate recognition system usually consists of data acquisition (license plate image acquisition), license plate extraction, and license plate identification several major components, the system architecture as shown978-1-4244-5540-9/10/$26.00 ©2010 IEEE in figure 1.Image extractionRecognitionresultsImage preprocessingInformation recognitionFigure I. Vehicle license plate recognition system structureIn the vehicle license plate recognition system, the image acquisition is completed mainly by the hardware, which is to extract the foreground image of the vehicle, to convert the camera's video signal to digital image signals to be sent to the computer for processing. Because the impact of the natural environment and the lighting conditions, there are many disturbances in the license plate images, which brings inconvenience to the positioning of the license plate, so in order to better extract the license plates, it needs to preprocess the license plate image to ensure the license plate location quality. VLP detection, this part is the core of the system, and the implementation of which affects the performance of the whole system, which is mainly to use pattern recognition[2] , digital image processing, information theory and other knowledge to position and extract the license plate in the license plate images. Character segmentation and recognition, when the plate has been successfully extracted, it needs to segment the characters in which, and use prior knowledge to identify them to get the final results.B. Key technologies o/license p late recognition1) Vehicle license regional positioning technology: it is to use the above characteristics to determine the true location of l icense plate. To accurately position the vehicle license plate .from the images obtained .from the natural scene is the key of the vehicle license plate recognition system, is also the most diffi c ult ste p.2) Vehicle license plate character segmentation technology: it is to divide the license plate region into a single character region .for the follow-u p license plate recognition module to ident!fY the single characters.J) Vehicle license plate character recognition technology: character recognition is the p rocess of confirming the Chinese characters, English letters and numbers on the license plate on the basis of the accuratel"J segmentationfor the vehicle license p late character .III. VEHICLE LICENSE PLATE RECOGNITION SYSTEM BASED ON DIGITAL IMAGE PROCESSINGA. System designVehicle license plate recognition system structure as shown in figure 2:Vehicle license plate recognition systemImage extractionImage preprocessingRegional locationCharacterssegmentationCharacter recognitionI m age filteringI mage binarizationEdge detectionFeature extractionPattern matchingFigure 2. Vehicle license plate recognition system structurei) The regional location of v ehicle licenseIn the vehicle images, to position a car license plate is a difficulty of vehicle license plate recognition and image coding treatment. The positioning in the vehicle license is to extract the coordinates of the vehicle license plate area from the vehicle image, and then identify the license characters. It also needs to consider the distortion of the captured image, illumination uniformity, transmission impact and other reasons. If the captured image is vague, the license area is not obvious, and then the license area extraction will be greatly difficult. At present, the picture area extraction methods are static algorithm, character object extraction algorithm based on edge extraction and adaptive robust, the target search strategy algorithm based on color segmentation and other algorithms. The common starting point of those is to determine the location of the license through the characteristics of the license area.2) image p re processingIn order to identify the vehicle license from the video image, the original image is required to have appropriate brightness, higher contrast and can identify license plate video image. However, because that the camera usually works in an open outdoor environment, so the camera when shooting may affected by the distance between the camera and the license, their angle, the traffic speed, vehicle license plate clean degree and other factors, so that the license image may have blur, deflection, defects and other serious defects, therefore, it needs to preprocess the original image before the identification. The image preprocessing includes: image restoration, image enhancement, gamma correction, gray correction, color image grayscale, grayscale stretching, and other processes.J) Vehicle license plate character segmentation and processingAfter the license character image positioning and binary processing, the vehicle license is a level bar area only contains the license character, and to recognize the image characters, these characters need to be segmented from the binary license image. Because the weak stain, loss and other factors of the vehicle license may make the image have greater image noise, while the image binary process making some useful information lost, which resulted in the blur, even incompletion, adjacent characters adhesion of the license to be segmented, seriously affected the results of segmentation, therefore, the use of segmentation method with reference to the license character characteristics is effective.The key of the license plate recognition system is key character feature extraction. In other words, how to select the feature vector which is not only easy to extract, but also easy to identify, as well as has the feature vector as little as 'bl[6] d h h ..POSSl e ,an t e c aractenstlc parameters similar to the best sample characteristics, is the key of the feature extraction. Feature extraction and selection is crucial to the system identification, which basically determines the identification system performance and recognition accuracy, and even can affect the entire system identification effect. 4) Vehicle license p late character recognitionThe key of the license plate character recognition is the character feature extraction and pattern matching. When feature extracting and pattern matching there are mainly the following ways: one is the use of character structural features and transform for feature extraction, this method has high tolerance to the character incline and deformation, but the computation is huge, requires high computer performance. Another is the use of character statistical features for feature extraction, at present, most character recognition systems use this method, and when extracting the character features, also the character projector features and profile features can be used to composed the character feature vector for feature matching, thus the results have a high recognition rate.The specific processing flow of the system as shown in figure 3:Features information matchENDFigure 3. System specific processing flow chart Based on the hundreds of pieces of the vehicle licenseimages this paper carries out a license positioning and segmentation test, the results show that the correct rate can reach 96%. The automatic segmentation results can meet the requirements of character segmentation and recognition; andthe recognition range is accurate, the area size is appropriate, there is no missed part of the license[3] ; for the image without ideal light conditions, an image enhancement can be carried out once to make the dynamic range of image gray expanded and the contrast enhanced, and then for the image positioning and segmentation, thus, to improve the accuracy of image segmentation.B. System p eJjOrmance analysis1) Accuracy analysisIn order to achieve the purpose of real-time processing, the algorithms used in this system are not involved with complicated mathematical functions, and in such circumstances the system achieves good results, because the parts the system involved are more, so the output of each part can be the input of the next part. Linked together, the previous module error necessarily will lead to the latter modules error. Therefore, the system is a typical serial system, and the overall accuracy depends on the product of the various part accuracy.2) D!lJiculty analysisIn the image acquisition, the different object distances often result in different license plate sizes in the image. And the processing method of a fixed threshold adopted in the previous algorithm has not a universal adaptability. A fixed threshold can only handle a certain size of license plate images, but for other images with different sizes is helpless. Faced to a large number of license plate images with different sizes, to find a new algorithm with wider applicability is not easy. In actual image acquisition, the noise, light has a great influence on the image quality. A lot of random noise disturbance, different perspectives of the light, light, resulting in license plate light and dark gray irregular changes. The irregular and uncertain occurrence of the deformation, noise and other interference information all make the clarity of the captured license plate image greatly reduced.The difference of the angel when image collecting, the actual front license plate and the license plate incline will cause the captured license plate graphic to generate geometric deformation[5] . And the license plate graphic geometric deformation, the different degrees of the deformation, also make the license plate positioning in the license plate image and the license plate character recognition more difficult. This requires the license plate location and recognition to have high anti-interference and robustness.IV. CONCLUSIONVehicle license plate intelligence recognition system as the core of traffic identification system will play an important role in the future traffic control. This paper studied the vehicle license plate recognition system based on image processing in the intelligent traffic system, proposed a character recognition solution based on neural network perceptron to solve the license plate recognition problem in the real-time traffic flow, and also had some research on the vehicle license plate character recognition algorithm, the test results showed that the system had high anti-interference and robustness.REFERENCES(2). Dai Yan, Ma Hongqing, Liu Jilin. High Performance License PlateRecognition System Based on the Web Technique, IEEE Intelligent Transportation Systems Conference, August 25-29, 2001, 325-334. (3). Taleb A, Hamad A, Tilmant D. Vehicle license Plate recognition inmarketing application [C]. IEEE Transaction on Intelligent Vehicles Symposium, 2003, 90-94.(4). Kawaguchi H. Application system using license Plate recognitiontechnology [Jl. NEC Technical Journal, 2005, 54(7): 19-22.(5). Tsang-Hong Wang, Feng-Chou Ni, Keh-Tsong Li et al. RobustLicense Plate Recognition based on Dynamic Projection Warping. In: Sensing & Control, Taipei, Taiwan, 2004, 784-788.(6). Wenjing Jia, Huaifeng Zhang, Xiangjian He et al. Mean Shift forAccurate License Plate Localization. In: Intelligent Transportation Systems, Vienna, Austria, 2005, 566-570.(7). Hyun-Chul Kim, Shaoning Pang, Hong-Mo Je, Daijin Kim, SungYang Bang. Pattern Classification Using Support Vector Machine Ensemble, In: Pattern Recognition, Proceedings, South Korea, 2002, 160-163.。