基于雙目視覺(jué)的稠密立體匹配算法研究

《基于雙目視覺(jué)的稠密立體匹配算法研究》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。

1、分類號(hào)UDC密級(jí)基學(xué)號(hào)1307010094于雙目視覺(jué)的稠密立體碩士學(xué)位論文匹配算法研究基于雙目視覺(jué)的稠密立體匹配算法研究柯柯俊山俊山學(xué)科門(mén)類：理學(xué)學(xué)科名稱：數(shù)學(xué)西指導(dǎo)教師：戴芳教授安理工申請(qǐng)日期：2016年3月大學(xué)萬(wàn)方數(shù)據(jù)萬(wàn)方數(shù)據(jù)萬(wàn)方數(shù)據(jù)西安理工大學(xué)碩士學(xué)位論文算法平均排名為112.7。最后，針對(duì)被動(dòng)測(cè)量技術(shù)中，三角測(cè)量方法不能充分利用攝像機(jī)內(nèi)參數(shù)的缺陷對(duì)其進(jìn)行了改進(jìn)。我們應(yīng)用（2）中的S-ELAS視差計(jì)算方法，使用KITTI視覺(jué)基準(zhǔn)數(shù)據(jù)庫(kù)中的交通場(chǎng)景圖像對(duì)改進(jìn)的三角測(cè)量方法進(jìn)行了測(cè)試。測(cè)試結(jié)果

2、表明，改進(jìn)的三角測(cè)量方法的平均坐標(biāo)準(zhǔn)確率是89.79%，平均距離準(zhǔn)確率是96.60%，可見(jiàn)修正的距離計(jì)算方法非常精確。關(guān)鍵詞：雙目視覺(jué)；立體匹配；SIFT；SLIC；ELASII萬(wàn)方數(shù)據(jù)AbstractTitle：STUDYONDENSESTEREOCORRENPONDENCEALGORITHMFORBINOCULARVISIONMajor：MathematicsName：JunshanKeSignature:Supervisor：Prof.FangDaiSignature:AbstractBa

3、sedontheunderstandingofbinocularstereovisionsystem,wehavedonesomeresearchesonfeature-basedandwindow-basedstereocorrenpondencealgorithms,andmainlydiscussedhowtoimprovethedisparityresultofbinocularstereomatchingalgorithmandhowtodevelopdistancecalculati

4、onmethod.Thestereomatchingalgorithmoftenmakeitdifficulttoobtainaccuratedisparitysearchrange,expensivecostofcomputationandlowmatchingrate,thisthesispresentstwoimprovedalgorithms:(1)TakingtheadvantagesofSIFTfeatureextractionalgorithm,EMDalgorithmandwin

5、dow-basedmatchingtechniquesintoconsideration,aimproveddensestereocorrespondencealgorithmisproposed.First,weproposealocalfeaturedetectionalgorithmcombinedSIFTwithEMD.BecauseSIFTisaveryrobustlocalfeatureextractiontechniquewhichisgoodforcapturingthemost

6、importantfeaturesofimage,andtheEMDmethodisself-adaptiveandhighlyefficientinanalyzingnonlinearandnon-stationarysignalandthismethodisabletodecomposethesignalintothesumofsimplesignalsindifferentfrequency.Afterthefeatureextractionofimage,thematchingpoint

7、pairswithhigheraccuracyareobtainedbyfeaturematchingmethodandepipolargeometricconstraint,andthenthesematchingpointsareusedtoestimatetheinitialdisparitysearchrange.Second,onthebasisofnewblockmatchingstrategyandmodifiedmatchingcostfunction,thedensedispa

8、ritymapisobtainedbytheimproveddisparityreliabilityfunctionorweightenergyfunction.TheresultsshowthatthefeaturedetectionalgorithmcombinedSIFTwithEMDismoreaccuratethanthetraditionalSIFTapproach,anditcouldclosertheactualrequirementdisparitysearchrange.An