基于序列圖像的多目標(biāo)認(rèn)同技術(shù)研究.pdf

1、中北大學(xué)碩士學(xué)位論文基于序列圖像的多目標(biāo)認(rèn)同技術(shù)研究姓名：郭紅玉申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：信號(hào)與信息處理指導(dǎo)教師：韓焱20080428中北大學(xué)學(xué)位論文ResearchintoMulti—objectIdentificationTechnologyBasedonSequenceImagesAbstractMultiimagesensorintersectionmeasurementisavitalresearchpartincomputer

2、visionInmulti—imagesensorintersectionmeasurementitusuallyneededcorrelatedinformationofthesameobjectwhichWasindifferentimagesensorsSpeciallywhenthereweremanyobjectsthatmovingwithrandominspace，duetothetimeandspacelocationo

3、feveryobjects’correlatedinformationchangingineverytime，atthistimeusingdifferentimagesensorswhichmeasuredbyintersectionmeasurementforanyobjects，itwasverydifficultAimingtothispuzzlethepaperanalysesandsummarizestheobjectide

4、ntificationtechnologyinimagematchingfirstlyanalysestheresultofidentificationandindicatestheshortageofapplicationinmultiimagesensorLatertothedrawbackofmultiobjectidentificationinimagematchingitresearchesthecorrelatedident

5、ificationtechnologywhichisbetweenthelocationinformationofmulti—imagesensor謝tllthecorrelatedinformationofobjectimageByanalyzingtheCalTleraimagingmodelinspaceitUSesthemethodofepipolarconstrainteducingwrongmatchingofcorresp

6、ondingpointsOnthebaseoftheresearchabovementionedprincipallystudiestheestimatingmethodsofepipolargeometryandfundamentalmatrixbasedonwhichakindmethodthatthewholeleastsquareestimationofRANSACfundamentalmatrixestimationwaspr

7、oposedFinallythemethodappliesinthispaperisvalidatedandanalyzedTheresultindicatesthattheimproved劬d鋤a叫霜matrixestimationcanreducewrongmatchingeffectivelyandcanavoidtherandomofobjectidentificationItrealizesmultiobjectidentif