本文關鍵詞：汽車牌照識別技術的研究　出處：《華南理工大學》2014年碩士論文　論文類型：學位論文

  更多相關文章： 汽車牌照識別 車牌定位 字符分割 字符識別 投影法 BP神經(jīng)網(wǎng)絡 模板匹配

【摘要】：汽車牌照識別技術不僅可以應用在交通管理領域，而且在收費站、停車場等車輛管理領域以及其它公共場所等民用領域也都有廣泛的應用；目前隨著圖象識別技術的快速發(fā)展，汽車牌照識別技術的應用也已經(jīng)滲透到了當今社會的各個領域，并且為社會管理和人民生活帶來了極大的便利。因此，進一步開展汽車牌照識別技術研究能夠極大地促進我國車輛管理公共事業(yè)的發(fā)展，對提高我國交通領域的現(xiàn)代化管理水平具有重要的學術研究意義和應用價值。 本文在理解并掌握汽車牌照識別相關關鍵技術的基礎上，分別從圖像預處理、汽車牌照圖像定位、圖像分割和圖象識別等方面展開了較深入的研究。論文作者的主要研究工作包括如下內(nèi)容： （一）在圖像預處理技術研究中，分別對圖像的灰度化、降噪、圖像增強以及二值化技術進行了較深入的探討，仿真效果表明，，通過圖像預處理后，汽車牌照圖像的質量得到了明顯的改善； （二）在車牌定位研究中，研究了基于數(shù)學形態(tài)學和紋理特征的粗定位和精確定位方法，比較了兩種定位方法的仿真效果，證明了精確定位方法的有效性和準確性； （三）在車牌字符分割技術研究中，分別研究了基于聯(lián)通區(qū)域和基于投影法的字符分割方法，闡述了兩種方法各自的適用性，并對其分割效果進行了對比，結果表明，在本文的研究中，投影法的定位效果較好； （四）在汽車牌照字符識別技術研究中，綜合了圖像預處理、定位和分割各種算法的基礎上，提出了基于模板匹配和基于BP神經(jīng)網(wǎng)絡的字符識別方法，并通過對33幅隨機汽車牌照圖像采用MATLAB仿真進行了兩種識別效果的對比，結果表明，基于BP神經(jīng)網(wǎng)絡的汽車牌照字符識別方法的識別效果較為準確和可靠。
[Abstract]:The vehicle license plate recognition technology can be used not only in the field of traffic management, but also in the toll station, parking lot and other civilian areas of vehicle management areas and other public places also have a wide range of applications; at present, with the rapid development of image recognition technology, the application of license plate recognition technology has penetrated into all areas of society, and brings great convenience to people's life and social management. Therefore, the further study of vehicle license plate recognition can greatly promote the development of China's vehicle management of public utilities, it has important academic significance and application value to improve the traffic field of our country's modern management level.
On the basis of understanding and grasping the key technologies of license plate recognition, this paper launches a deep research from the aspects of image preprocessing, license plate location, image segmentation and image recognition.
(1) in the research of image preprocessing technology, we discussed the grayscale, noise reduction, image enhancement and two value technology of image respectively. The simulation results show that the quality of license plate images has been significantly improved through image pre-processing.
(two) in the research of license plate location, we studied coarse positioning and precise location method based on mathematical morphology and texture features, compared the simulation results of two positioning methods, and proved the accuracy and validity of the precise location method.
(three) research on license plate segmentation technique, were studied and Unicom region segmentation method based on projection characters, expounds the applicability of the two methods, and the segmentation results were compared. The results show that, in this study, better positioning effect projection method;
(four) in the study of vehicle license plate character recognition technology, integrated image preprocessing, positioning and segmentation algorithms based on template matching is proposed. And the method of character recognition based on BP neural network based on 33 images, and through random car license plate using MATLAB simulation were compared, two recognition results the results show that the vehicle license plate character recognition method based on BP neural network recognition effect is more accurate and reliable.

【學位授予單位】：華南理工大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：U495;TP391.41

【參考文獻】

相關期刊論文 前10條

1 李小平,任江興,楊德剛;車牌識別系統(tǒng)中若干問題的探討[J];北京理工大學學報;2001年01期

2 袁寶民,于萬波,魏小鵬;汽車牌照定位研究綜述[J];大連大學學報;2002年02期

3 馮國進,顧國華,鄭瑞紅;基于自適應投影方法的快速車牌定位[J];紅外與激光工程;2003年03期

4 汪志兵,崔慧娟;一種基于紋理特征抽取的車牌定位預處理方法[J];計算機應用研究;2004年01期

5 牛欣,沈蘭蓀;基于特征的車輛牌照定位算法[J];交通與計算機;2000年01期

6 李志強;李永斌;;車牌識別技術的發(fā)展及研究現(xiàn)狀[J];科技信息;2012年05期

7 葉晨洲,楊杰,宣國榮;車輛牌照字符識別[J];上海交通大學學報;2000年05期

8 曹仁帥,董慧穎,于霞;汽車牌照快速搜索算法研究[J];沈陽工業(yè)大學學報;2002年06期

9 吳留生;陳家新;黎蔚;;基于Niblack的手掌靜脈圖像二值化技術研究[J];通信技術;2010年01期

10 廖翔云,許錦標,龔仕偉;車牌識別技術研究[J];微機發(fā)展;2003年S2期

本文編號：1397542