視覺輔助著陸系統(tǒng)中跑道識(shí)別算法研究與模擬驗(yàn)證.pdf

1、南京航空航天大學(xué)碩士學(xué)位論文視覺輔助著陸系統(tǒng)中跑道識(shí)別算法研究與模擬驗(yàn)證姓名：李航申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：導(dǎo)航、制導(dǎo)與控制指導(dǎo)教師：曹云峰201012視覺輔助著陸系統(tǒng)中跑道識(shí)別算法研究與模擬驗(yàn)證iiABSTRACTThispaperfocusedontherunwayidentificationtechnologiesduringtheprocessofUAVautonomousling.Besidesasetofgroundsimul

2、ationverificationplatfmwasdesignedwhichcouldhelpustoverifycrespondingidentificationmethodsthefineresultsweregot.Firstlythepapergenerallyintroducesapresentoverviewofthestatusofautonomouslingtechniquesathomeabroad.Byanalyz

3、ingtheacteristicsofdigitalimagesfromfwardlookingcamerawhichreflecttheairptenvironmentthevisionprocessingobjectrecognitionprocessweredetermined.Accdingtosalientfeaturesofreflectedinthedifferentperiodoflinganadaptivethresh

4、oldingmethodbasedonhistogramlocalminimumwasproposedwhichcouldeffectivelysegmentrunwaytargetareawithlongrange.MeanwhilewiththeadvantageofanimprovedCannyedgedetectionamethodthatcouldreliablyextractedgeinfmationofrunwaytarg

5、etinthecloserdistancewasproposed.Secondlybasedontheareaedgeinfmationoftherunwayacompletesetofrunwayobjectrecognitionsolutionwasdeterminedaccdingtothedifferentapplicationscopeoffmertwoimageprocessingmethodsthewholelingpro

6、cesswasdefinitelydividedintofarstageclosestage.Ffardistancestageatemplatematchingmethodfrunwaytargetrecognitioncouldbeused.fclosestagewithamethodbasedonlineacteristicsextractionthevisionsystemcanalsorecognizerunwaytarget

7、areareliably.Throughthetransitionconditionsdeterminedarecognitionmethodwhichcouldswitchautomaticallyduringtheprocessfromfarnearareawasrealized.Againindertoverifytesttheimageprocessingalgithmthewholerecognitionmethodanemb

8、eddedsystemsimulationverificationplatfmmainlybasedonDM642wasdesigned.ItcansimulatepartsofscenariosofUAVlingprocessonthegroundprovidetherunwayidentificationexperiments.Finallythecrespondingexperimentalresultsshowthatthede

9、signedrunwayidentificationsolutionisvalidduringthewholelingprocess.Intheadditionthecontentoftheresearchisalsosummarizedprospectsinthispaper.Keywds:AutonomousLingAdaptiveThresholdingEdgeDetectionRunwayIdentificationDM642.