機(jī)場(chǎng)目標(biāo)分割與識(shí)別方法研究.pdf

1、西北工業(yè)大學(xué)碩士學(xué)位論文機(jī)場(chǎng)目標(biāo)分割與識(shí)別方法研究姓名：謝建春申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：趙榮椿20040301西北工業(yè)人學(xué)碩1:學(xué)位論文AbstractThethesisisfocusedonairporttargetdetectionsegmentationandrecognitionfortheremotesensingimage.Asawidelyusedtechnologyobjectrecognition

2、hasfollowingcharacteristics:pertinenceanddependence.IntheactualapplicationfacingthedifferentobjectwealwaysproposedifferentsupposesandselectdiferentmethodsEvenforthesameobjectdifferentmethodsareselectedinallusiontodiferen

3、tapplications.Basedonpredecessorworkwehopetomakesomeimprovementstotheirresearchresults.Intheaspectofobjectdetectionandsegmentationalgorithmsareclassifiedtofourpartsbydiscontinuityandcomparabilityofthepixels.Webeginwithpa

4、ralleltechnologyofedgeandareasegmentation.Thensometypicalalgorithmsareintroduced.Inallusiontotheweaknessofsinglealgorithmandcharacteristicsoftheairportweproposeamethodcomposedofmultisegmentationalgorithms.Theexperimentsr

5、esultshowthatitworkswell.Intheaspectofrecognitionsomequestionsondesignrecognitionsystemarediscussed.Weintroducethefactorsofrecognitioncomplicationtheexpressionofairportmodelshapebasedfeatureextractionmethods.Afterwardbas

6、edoncharacteristicsofairportarecognitionalgorithmusingHoughtransformationisintroducedandtheexperimentsshowitworkswell.Intheendrecognitionondeeperresearchingairportobjectsdetectiondiscusstheparallelprogramdesign.Basedonse