基于視頻圖像的車輛檢測和匹配跟蹤方法研究
發(fā)布時間:2018-04-05 00:38
本文選題:智能交通 切入點:車輛檢測 出處:《長安大學(xué)》2014年碩士論文
【摘要】:車輛的檢測和跟蹤是智能交通系統(tǒng)中必不可少的重要組成部分,它為后續(xù)的交通控制管理提供重要的數(shù)據(jù)依據(jù)。基于視頻的車輛檢測和跟蹤較傳統(tǒng)的檢測方法,具有直觀性好、抗干擾強、檢測范圍廣和性價比高等顯著優(yōu)點。隨著信息技術(shù)的快速發(fā)展,基于視頻的車輛檢測和跟蹤方法的研究吸引了眾多研究者的關(guān)注,成為當(dāng)今的研究熱點。 論文介紹了視頻車輛檢測和跟蹤技術(shù)的發(fā)展現(xiàn)狀,給出了整個檢測系統(tǒng)的結(jié)構(gòu)框架,并結(jié)合圖像處理的相關(guān)知識對視頻序列進(jìn)行處理,對系統(tǒng)中的關(guān)鍵技術(shù)進(jìn)行了研究,具體研究內(nèi)容如下: 在車輛檢測方面,研究了常用的背景提取方法(如,統(tǒng)計法、均值法和中值法)和車輛提取方法(如,光流法、幀間差分法和背景差分法)。因為利用以上方法提取到的車輛含有大量陰影信息,故又對車輛陰影的去除做了大量試驗研究,最后采用一種基于陰影特征的方法去除陰影,大體步驟如下:首先確定陰影所在的方向。其中,,針對深色車輛的特殊情況,提出一種通過對比差分圖像各邊界灰度值方差的方法,來確定陰影的方向;然后計算出車輛陰影灰度值的分布區(qū)間;最后根據(jù)分布區(qū)間,來達(dá)到去除陰影的目的。去除陰影后,利用形態(tài)學(xué)、連通區(qū)域和圖像歸并的相關(guān)知識和操作,最后確定出運動目標(biāo)車輛的輪廓。 在車輛跟蹤方面,通過研究了幾種常見的跟蹤方法,最后采用了一種基于區(qū)域灰度值的方法來進(jìn)行匹配跟蹤,并對這種方法進(jìn)行了改善,利用提取到的車輛區(qū)域來確定匹配區(qū)域,從而縮小了匹配范圍。結(jié)果顯示,基于區(qū)域灰度值的車輛匹配跟蹤效果良好。
[Abstract]:Vehicle detection and tracking is an essential part of intelligent transportation system, which provides important data basis for the subsequent traffic control management.Vehicle detection and tracking based on video has many advantages, such as good visualization, strong anti-jamming, wide detection range and high performance-to-price ratio.With the rapid development of information technology, the research of vehicle detection and tracking based on video has attracted the attention of many researchers.This paper introduces the development of video vehicle detection and tracking technology, gives the framework of the whole detection system, processes the video sequence with the relevant knowledge of image processing, and studies the key technologies of the system.The specific contents of the study are as follows:In vehicle detection, common background extraction methods (such as statistical method, mean method and median method) and vehicle extraction methods (such as optical flow method, inter-frame difference method and background difference method) are studied.Because the vehicles extracted by the above methods contain a lot of shadow information, a lot of experiments have been done on the shadow removal of vehicles. Finally, a shadow removal method based on shadow features is adopted.The general steps are as follows: first, determine the direction of the shadow.According to the special situation of dark vehicle, a method is proposed to determine the direction of shadow by comparing the variance of gray values of each boundary of the difference image. Then, the distribution interval of gray value of vehicle shadow is calculated. Finally, according to the distribution interval,To remove shadows.After the shading is removed, the vehicle contour of the moving target is determined by using the relevant knowledge and operation of morphology, connected region and image merging.In the aspect of vehicle tracking, several common tracking methods are studied. Finally, a method based on region gray value is used to match and track, and this method is improved.By using the extracted vehicle area to determine the matching area, the matching range is reduced.The results show that the vehicle matching and tracking effect based on region gray value is good.
【學(xué)位授予單位】:長安大學(xué)
【學(xué)位級別】:碩士
【學(xué)位授予年份】:2014
【分類號】:U495
【參考文獻(xiàn)】
相關(guān)期刊論文 前10條
1 王紅梅;李言俊;張科;;一種改進(jìn)型椒鹽噪聲濾波算法[J];光電子.激光;2007年01期
2 張文溥;;視頻車輛檢測技術(shù)及發(fā)展趨勢[J];中國人民公安大學(xué)學(xué)報(自然科學(xué)版);2010年01期
3 劉國宏;郭文明;;改進(jìn)的中值濾波去噪算法應(yīng)用分析[J];計算機工程與應(yīng)用;2010年10期
4 周金和;彭福堂;;一種有選擇的圖像灰度化方法[J];計算機工程;2006年20期
5 祖仲林;李勃;陳啟美;;基于局部紋理特性的運動車輛陰影消除[J];計算機工程;2009年16期
6 王文豪;周泓;嚴(yán)云洋;;一種基于連通區(qū)域的輪廓提取方法[J];計算機工程與科學(xué);2011年06期
7 吳思,林守勛,張勇東;基于動態(tài)背景構(gòu)造的視頻運動對象自動分割[J];計算機學(xué)報;2005年08期
8 肖又發(fā);基于環(huán)形線圈車檢器的車輛分類研究[J];交通部上海船舶運輸科學(xué)研究所學(xué)報;2004年02期
9 賈小軍;喻擎蒼;;基于開源計算機視覺庫OpenCV的圖像處理[J];計算機應(yīng)用與軟件;2008年04期
10 儲s
本文編號:1712446
本文鏈接:http://sikaile.net/kejilunwen/jiaotonggongchenglunwen/1712446.html
最近更新
教材專著