基于视频图像处理的交通检测算法及应用研究
详细信息    本馆镜像全文|  推荐本文 |  |   获取CNKI官网全文
摘要
交通检测主要包括交通流参数检测和车牌识别技术(LPR),它们都是智能交通系统(ITS)的重要指标,如何准确、快速、实时地检测和识别成为该领域的热门课题。本文使用近几年快速发展的图像处理技术对该课题展开研究,相比其它检测识别方法具有很大的优越性。
     本文首先介绍了图像处理的一些基本方法,如灰度化、中值滤波、边缘检测等,后面的研究都是基于这些方法进行的。提出了一种基于虚拟线圈的交通流参数检测法,该方法能有效地统计车流量、判断车型、检测车速和车头时距等参数。采用背景自适应差分法,能适应外部环境的变化,如雨雪雾天气、阴天、抖动等干扰。实验表明,本方法有效可行、算法简单、操作方便,能满足交管部门对交通流参数的提取要求。
     车牌识别技术是ITS的重要组成部分,已广泛地应用在了众多领域。车牌定位部分采用Sobel垂直边缘检测算子、二值化和水平投影相结合的方法,同时使用Hough变换对倾斜车牌进行倾斜度校正,以便能准确定位车牌。根据车牌字符垂直投影后灰度分布呈现峰、谷、峰的特点,字符的分割部分采用垂直投影法,并对字符粘连和断裂的情况进行了特殊处理,这样精确地得到了单个的车牌字符,可将非字符的杂质去除。字符的识别部分采用一种基于模板匹配的识别法,建立标准模板库,利用归一化后的字符与之匹配,并对相似字符再识别,具有较高的识别率。
Traffic detection mainly includes traffic stream parameter detection and Licence Plate Recognition (LPR), both of them are important parts of Intelligent Transportation System (ITS), how to detect and recognize them exactly, quickly and timely has became a hot subject in this field. With the development of digital image processing technology (DIPT) by these years, which is always be used in this field, has more advantages compared with other detection and recognition methods.
     This article introduces some primitive method of DIPT firstly, like gray change, median filter and edge detection, all later research is based on these methods. We proposed a traffic stream parameter detection method based on virtual coil, this method can statistic traffic flow, judge motorcycle type, detect speed and motor head interval efficiently. By using self-adaption background difference method, it can adapt external environment changing, like some bad weather and camera shaking. The experiment shows that the algorism is simple and operation is easy, so this method can achieve transport department requirements.
     LPR is an important part of ITS, it has been widely used in many fields. By combining with Sobel vertical edge detection operator, binarization and horizontal project method to locate plate position, and using Hough transform to correct lean plate. According to gray level distribution of plate character appears peak, valley and peak feature after vertical projection, the plate division is used vertical projection method, and we do some special handling to adhesion and fracture character. In that case, we get single plate character accurately and wipe off impurity. One recognized method based on template matching is used in character recognition, after establish standard template library, we use normalized character to matching library and matching similar character again to get high recognition rate.
引文
[1]安实,王健,徐亚国等.城市智能交通管理技术与应用[M].北京:科学出版社,2005
    [2]冈萨雷斯,伍兹.数字图象处理[M].阮秋琦译.北京:电子工业出版社,2007
    [3]龚声蓉,刘纯平,王强等.数字图象处理与分析[M].北京:清华大学出版社,2006
    [4]何东健,耿楠,张义宽等.数字图象处理[M].西安:西安电子科技大学出版社,2006
    [5]勒中鑫.数字图象信息处理[M].北京:国防工业出版社,2003
    [6]杨淑莹.VC++图象处理程序设计[M].北京:清华大学出版社,2003
    [7]王润生.图像理解[M].长沙:国防科技大学出版社,1993
    [8]周道炳,朱卫纲.几种边缘检测算子的评估[J].指挥技术学院学报,2000,11(1):59-63
    [9]《中华人民共和国道路交通安全法实施条例》国务院令第405号
    [10]何云,许建龙,孙树森等.一种改进的视频监控背景更新算法[J].浙江理工大学学报,2010, 27(4):585-589
    [11]杨志奇.车辆检测系统中背景提取与更新算法研究[J].计算机与现代化,2010,179(7):111-114
    [12]夏伟才,曾致远.一种基于卡尔曼滤波的背景更新算法[J].计算机技术与发展,2007,17(10):134-143
    [13]汤晖.基于视频运动检测的多车道交通流信息采集技术研究:[硕士学位论文].江西赣州:江西理工大学,2008
    [14]潘秦华.车辆目标检测与交通流检测系统的研究:[硕士学位论文].西安:西安电子科技大学,2005
    [15]崔凤英.基于视频的车流量检测算法研究:[硕士学位论文].青岛:青岛科技大学,2007
    [16]张玲,陈丽敏,何伟等.基于视频的改进帧差法在车流量检测中的应用[J].重庆大学学报,2004,27(5):31-73
    [17]何最红.基于视频的交通流参数检测方法研究:[硕士学位论文].广州:广东工业大学,2006
    [18]朱矿军.基于视频虚拟检测线特征的交通流参数检测:[硕士学位论文].大连:大连理工大学,2005
    [19]李树广,刘允才.智能交通的发展与研究[J].计算机应用,2005, 21(6):1-6
    [20]许伦辉,陈衍平,修科鼎.基于图像处理的静态车牌识别技术[J].江西理工大学学报,2011, 32(1):47-50
    [21]戴青云,余英林.一种基于小波与形态学的车牌图象分割方法[J].中国图像图形学报,2005, 5(1):412-415
    [22] Guo Jie,Shi PengFei. Color and texture analysis based vehicle lit cense plate location. Journal of Image and Graphics.2002,7(5) :472-476
    [23] Shi Xifan, Zhao Weizhong, Shen. Automatic license plate recognition system based on color image processing. Yonghang Source: Lecture Notes in Computer Science, v 3483,n IV,Computational Science and Its Applications-ICCSA 2005: International Conference,Proceedings: 2005, 11(5) :59-1168
    [24] Park, S .H. Kim, K .I. Jung, K .Kim, H .J. ,Locating vehicle license plates usingNeural networks, Electronics Letters 2002, 35(17):1475-1477
    [25] Liu Ji Lin, Song Jia Tao. Vehicle License Plate Recognition System with High Performance.Acta Automatica Sinica.2003,4(3):77-80
    [26] Paolo Comelli, Paolo Ferragina, Mario Notturno Granieri, and Flavio Stabile.Optical recognition of motor vehicle license plates. IEEE Transactions on Vehicular technology. Nov2003, vol. 44(4):790-799
    [27]朱光忠,黄云龙.边缘检测算子在汽车牌照区域检测中的应用[J].计算机技术与发展,2006,16(3):161-162
    [28] Starck J L, Candes E J, Donoho D L .The curvelet transform for image denoising. IEEE Trans.Image Processing.2002, 6(3):670-684
    [29] Yasuharu Yanamura, Masahiro Goto, Daisuke Nishiyama. Extraction and Tracking of the License Plate Using Hough Transform and Voted Block Matching. IEEEI V2003 Intelligent Vehicles Symposium Conference, 2003,21(7) :236-240
    [30]路小波,凌小静,黄卫.一种基于模板匹配的汽车牌照识别方法[J].交通与计算机,2004, 5(6):79-82
    [31]安勇,张高伟.基于灰度图像的车牌识别系统.计算机工程与科学[J].2006.28(2):1420 - 1422
    [32] YuntaoCui, QianHuang. Extracting characters of license plates from video Sequences.Machine Vision and Applications. 2004, vo1.10 (5-6): 308-320
    [33]陈振学,汪国有,刘成云.一种新的车牌图像字符分割与识别算法[J].微电子学与计算机,2007, 11(2):101-103
    [34]《中华人民共和国机动车号牌》GA36-1992行业标准
    [35] Park, S .H. Kim, K .I. Jung, K .Kim, H .J. ,Locating vehicle license plates using Neural networks, Electronics Letters 2002, 35(17):1475-1477
    [36] R.Parist,et al .Car Plate Recognition by Neural Networks and Image Processing.IEEE International Symposium on Circuits and Systems. 2000,5(31) :212-220
    [37]王春,刘波,周新志.采用BP神经网络的车牌字符识别方法研究[J].中国测试技术,2005, 7(1):26-28
    [38]胡晓燕,蒋先刚,刘海峰.基于神经网络的车牌字符识别算法实验及程序校验[J].华东交通大学学报,2005,11(2):71-75
    [39]史忠科,曹力.交通图像检测与分析[M].北京:科学出版社,2007

© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号

地址:北京市海淀区学院路29号 邮编:100083

电话:办公室:(+86 10)66554848;文献借阅、咨询服务、科技查新:66554700