基于計(jì)算機(jī)視覺(jué)檢測(cè)的車(chē)牌自動(dòng)識(shí)別系統(tǒng)

摘要
隨著我國(guó)經(jīng)濟(jì)的高速發(fā)展，汽車(chē)的數(shù)量在急劇增加，國(guó)內(nèi)高速公路、城市道路、停車(chē)場(chǎng)等設(shè)施的建設(shè)也越來(lái)越多，因此對(duì)交通控制、安全管理的要求也日益提高 。智能交通系統(tǒng)已成為當(dāng)前交通管理發(fā)展的重要方向，而車(chē)牌自動(dòng)識(shí)別就是是智能交通系統(tǒng)的重要組成部分。
本文論述了計(jì)算機(jī)視覺(jué)技術(shù)的檢測(cè)原理、系統(tǒng)組成，研究了在車(chē)牌識(shí)別系統(tǒng)中的幾項(xiàng)關(guān)鍵技術(shù)——圖像預(yù)處理，定位及字符分割技術(shù)。
首先，本文對(duì)我國(guó)汽車(chē)牌照的特點(diǎn)進(jìn)行了分析，在進(jìn)行圖像預(yù)處理時(shí)采用了適合字符紋理特征的垂直邊緣檢測(cè)算法，基于形態(tài)學(xué)的車(chē)牌區(qū)域檢測(cè)，使用連通域分析的思想提取車(chē)牌區(qū)域。
而后，在進(jìn)行車(chē)牌定位時(shí)采用了垂直邊緣檢測(cè)與形態(tài)學(xué)運(yùn)算相結(jié)合的方法，這種方法定位準(zhǔn)確，且對(duì)有一定角度傾斜的車(chē)牌有較好的適應(yīng)性。
最后，在進(jìn)行字符分割時(shí)采用字符縱向切割的方法，能夠很好地去除車(chē)牌邊框，分割出各個(gè)字符，再運(yùn)用模板匹配法進(jìn)行字符識(shí)別。
本文的算法以車(chē)牌區(qū)域的紋理分析、數(shù)學(xué)形態(tài)學(xué)為基礎(chǔ)，有效提高了定位的準(zhǔn)確性。實(shí)驗(yàn)結(jié)果表明該算法能夠比較快速、準(zhǔn)確、定位分割出車(chē)輛牌照。
關(guān)鍵詞：計(jì)算機(jī)視覺(jué)技術(shù)；圖像處理；車(chē)牌定位；字符分割；字符識(shí)別。

ABSTRACT
With the rapid development of China’s economy,the number of cars is growing sharply, domestic highway, urban roads and parking lot are becoming more and more too. So the requirements to the traffic control, safety management are also improving day by day. Intelligent transportation system has become the important direction of the current traffic management development, and automatic license plate recognition is an important part of it.
This article has discussesde the detection principle of Automated Visual Inspection and its system composition, studied several key technologies domain in the automatic license plate recognition ： car license image pretreatment, the car license localization and the character division technology .
First, this article carefully analyzed our country license plate characteristic, the image pretreatment part has used has suited the character texture characteristic the vertical edge examination algorithm, based on the morphology car license region examination, the thought extraction car license region which the use connected domain analyzed .
 Next , has used the method in the car license localization part which the vertical edge examination and the morphology operation unifies, this method localization is accurate, also to has certain angle incline car license to have the good compatibility .
 Finally, after the character division part uses first the crosswise longitudinal cutting method, can remove the car license frame well, divides each character .
 This article algorithm take the car license region texture analysis as a foundation, enhanced the localization accuracy effectively .The experimental result indicated this algorithm can quite fast, accurate, the localization divide the vehicles license plate.
 Key word: Automated Visual Inspection, Imagery processing, automatic license plate localization, Character division，Character recognition.

目錄
基于計(jì)算機(jī)視覺(jué)檢測(cè)的車(chē)牌自動(dòng)識(shí)別系統(tǒng) 1
摘要 1
ABSTRACT 2
目錄 3
第一章 緒 論 4
1.1 選題研究意義 4
1.2 車(chē)牌識(shí)別系統(tǒng)的研究現(xiàn)狀 4
1.3 本文的框架結(jié)構(gòu) 5
第二章 車(chē)牌自動(dòng)識(shí)別系統(tǒng)簡(jiǎn)介 6
2.1 車(chē)牌識(shí)別系統(tǒng)基本工作原理 6
2.2 車(chē)牌自動(dòng)識(shí)別系統(tǒng)中的問(wèn)題與難點(diǎn) 7
第三章 圖像預(yù)處理 9
3.1圖像濾波 9
3.2 圖像增強(qiáng) 9
第四章 車(chē)牌定位 12
4.1邊緣檢測(cè) 12
4.2 基于形態(tài)學(xué)的車(chē)牌區(qū)域檢測(cè) 15
4.3 連通域標(biāo)記與車(chē)牌區(qū)域提取 17
第五章 字符分割 20
5.1 車(chē)牌區(qū)域二值化 20
5.2 車(chē)牌水平邊框去除 22
5.3 字符縱向分割 23
第六章 字符識(shí)別 25
實(shí) 驗(yàn) 結(jié) 果 27
總 結(jié) 30
1.本文主要研究的內(nèi)容 30
2.主要?jiǎng)?chuàng)新點(diǎn) 30
3.存在的不足 30
參考文獻(xiàn) 31