基于背景计算的运动目标视频检测方法研究

《基于背景计算的运动目标视频检测方法研究》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、中国农业大学硕士学位论文基于背景计算的运动目标视频检测方法研究姓名：陈望申请学位级别：硕士专业：机械电子工程指导教师：陈兵旗20060501AbstractMovingtargetdetectionisadifficultyinvisionprocessingsystemusedinopeningenvironments．Complicatedmethodswhichslowertheprocessinghavebeendevelopedtosolvetheproblemcausedbytheinterferentialfactors．Anewfastal

2、gorithmforcalculatingbackgroundwasdevelopedinthisdissertation，andthemovingtargetswereextractedusingbackgrounddifferencingmethod．Thebackgroundwasupdatedbasedontheresultsofmovingtargetdetectionbyreplacingthebackupbackground．ThealgorithmcanbeappliedtotheIntelligentTransportSystem．Aft

3、erdetectingthemovingvehides，theirshadowwasremovedbygraycharacteristics．Thenaccordingtotexturecharacteristics，thevehicleswereseparatedindriveways．Thevehiclesamountwasacquiredbymatchingimagesoquenees．Experimentalresultsdemonstratedthatthealgorithmcancountthevehiclesaccuratelybothins

4、traightandroundroadway,andtheaccuracyratewasabove97％．ThealgorithmcallalsobeappliedtotheIntelligentSurveillanceSystem．Throughidentifyingthemovingtargetsinasecretlocale，thevideowassavedintelligently．Thecarry—overinainvolutelocalewasdetectedusingthespacestability．Theintruderinaappoin

5、tedregionwasdetectedandtrackedusingtheregulationchangeofthecenterofitsmass．Experimentsshowedthatthealgorithmcancorrectlydealwiththesethreemonitoringsurveillancemodds．ThewholeprocessingmodulewasdevelopedonWindowsXPsystemwiththeapplicationofMicrosoftVisualC++6．0．usingtheMIASfunction

6、toreadandsavethevideo．Keywords：movingtargetdetection，backgroundextraction，countingvehicles，intelligentvideosurveillanceIl图表目录图1-1计算机视觉处理过程示意图⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯1图1-2帧间差分法示意图⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯4表2—1样本视频图像信息表⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯10图2-1视频图像中样本点的取值分布规律图⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯

7、⋯⋯⋯⋯⋯⋯⋯11图2-2区间计算⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯13图2-3背景提取及更新处理流程图⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯15图2_4不同自然光照的道路背景图像⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯16图2-5运动II标提取结果⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯17表2-2背景计算结果比较表⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯18图2-6样本点取值分布圈⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯19图3-1图像采集示意图⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯

8、⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯。23图3—2区域标记示意图⋯