资源描述:
《基于视频的行人检测研究论》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、基于视频的行人检测研究摘要随着中国经济的快速发展,公路交通日益发达,更加便利,但交通事故频繁发生,并且有着进一步恶化的趋势,这个问题是亟待解决的。为了保证行人的人生安全,保证交通的有序通畅,保证中国社会的和谐安定,视频的行人检测技术研发势在必行。本论文主要是针对于基于视频的行人检测进行了研究,主要内容包括:(1)对视频文件进行预处理,将视频一帧一帧地分离开来,以图片的形式进行保存,同时对其进行灰度化和去噪,加强视频图像的对比度,便于之后的目标识别和提取。(2)目标图像的提取方法主要有背景差分法和帧间差分法,这两种方
2、法各自有自己的优点和缺点,可以通过将两者相互结合,提取出完整的目标人物。(3)当然二值化后的目标很可能产生空洞和斑点,致使图像走形,我们可以通过形态学处理,填充目标的空洞,从而呈现于实体相一致的二值图像。关键词:行人检测;背景差分法;帧间差分法;MATLAB;视频ThePedestrianDetectionBasedVideoAbstractWiththerapiddevelopmentofChina'seconomy,qualityoflifecontinuestoimprove,roadtransportinc
3、reasinglydeveloped,moreconvenient,butaccidentsoccurfrequently,andhasafurtherdeterioration,theproblemissolved.Inordertoensurethesafetyofpedestriansinlife,toensuresmoothandorderlytraffic,toensureharmonyandstabilityofChinesesociety,videopedestriandetectiontechnol
4、ogydevelopmentimperative.Inaddition,insomelargesupermarkets,shoppingmalls,cinemas,especiallybanks,arelikelytoexisttheft,robberyevil,inordertocrackdownoncrimeandhelppolicecrackcasesmoreeffectively,weneedontheseoccasionsinstallationofpedestriandetectionsystem.Th
5、ispaperismainlyaimedatvideo-basedpedestriandetectionhasbeenstudied,themaincontentsinclude:(1)Pre-processingofthevideofile,thevideoframebyframeseparated,intheformofpicturestosave,whileitsgray,strengthenthecontrastofvideoimages,objectrecognitionandextractiontofa
6、cilitatelater.(2)Extractionmethodsmainlytargetimagebackgroundsubtractionandframedifferencemethod,thesetwomethodseachhavetheirownadvantagesanddisadvantages,youcancombinethetwotoeachother,toextractthecompletetargetfigure.(3)Ofcourse,afterthetargetbinaryislikelyt
7、ohavevoidsandspots,causingtheimagetotakeshape,wecanmorphologicalprocessing,fillingtheemptygoal,therebyrenderingtheentityconsistentbinaryimage.Keywords:PedestrianDetection;BackgroundSubtraction;FrameDifferenceMethod;MATLAB;Video目录1绪论............................
8、............................................................................................................11.1研究背景及意义.............................................................