資源描述:
《改進(jìn)仿射尺度不變特征變換算法的圖像配準(zhǔn).pdf》由會員上傳分享,免費在線閱讀,更多相關(guān)內(nèi)容在應(yīng)用文檔-天天文庫。
1、JournalofComputerApplicationsISSN1001—90812014—05—10計算機(jī)應(yīng)用,2014,34(5):1449—1452CODENJYIIDUhttp://www.joca.ca文章編號:1001—9081(2014)05-1449—04doi:10.11772/j.issn.1001—9081.2014.05.1449改進(jìn)仿射尺度不變特征變換算法的圖像配準(zhǔn)范雪婷,張磊,趙朝賀(1.中國礦業(yè)大學(xué)環(huán)境與測繪學(xué)院,江蘇徐州221116;2.國土環(huán)境與災(zāi)害監(jiān)測國家測繪地理信息局重點實驗室,江蘇徐州221116;3.山
2、東中煤物探測量總公司,山東泰安271021)(}通信作者電子郵箱zhaochcumt@163.eom)摘要:為了更好地處理匹配效率、重復(fù)紋理匹配和仿射不變性匹配等問題,對完全仿射不變特征變換(ASIFT)算法進(jìn)行兩方面改進(jìn)。匹配框架中特征提取的改進(jìn)提高了ASIFT算法的匹配效率;利用優(yōu)化隨機(jī)采樣算法(ORSA)結(jié)合以單應(yīng)矩陣為幾何線性約束模型的隨機(jī)抽樣一致性(RANSAC)改進(jìn)匹配算法,提高了匹配精度和重復(fù)紋理結(jié)構(gòu)的適應(yīng)能力。實驗結(jié)果表明,提出的改進(jìn)算法能較好地匹配高度相似紋理,計算量小,計算速度快且精度高。關(guān)鍵詞:圖像配準(zhǔn);仿射尺度不變特征變換
3、算法;單應(yīng)矩陣;重復(fù)紋理匹配中圖分類號:P237文獻(xiàn)標(biāo)志碼:AImprovedASIFTalgorithmforimageregistrationFANXueting,,ZHANGLei。,ZHAOChaohe,(1.SchoolofEnvironrncntScienceandSpatiallnformatics,ChinaUniversityofMiningandTechnology,XuzhouJiangsu221116,China;2.KeyLaboratoryofLandEnvironmentandDisasterMonitoring,N
4、ationalAdministrationofSurveying,MappingandGeoinformation,XuzhouJiangsu221116,China;3.ShandongCoalGeophysicalSurveyCorporation,Tai'anShandong271021,China)Abstract:Imageregistrationisawellresearchedtopicofcomputervision.Todealwithmatchingefficiency,repetitivepatternmatchingand
5、afineinvariantmatchingbetter,twoimprovementsoverthestate-of-the—artAffine—ScaleInvariantFeatureTransform(ASIFT)algorithmwerepresented.ThefeatureextractionofmatchingframewasdevelopedtoimprovethematchingeficiencyoftheASIFTalgorithm.Thesecondincreasedtheaccuracyofmatchingandthea
6、daptivecapacityofrepetitivepatternsthroughtheuseofimprovedmatchingalgorithmbycombiningOptimizedRandomSampleConsensus(ORSA)withRandomSampleConsensus(RANSAC)algorithmbasedongeometriclinearconstraintmodelwithhomographymatrix.Theexperimentalresultsshowthattheproposedmethodisablet
7、owellmatchhighlyrepetitivepatternsandhassmallercalculation,fasterspeedandhigheraccuracyaswel1.Keywords:imageregistration;Afine—ScaleInvariantFeatureTransform(ASIVF)algorithm;homographymatrix;repetitivepatternmatching人們提出或改進(jìn)各種各樣的算法,比如主成分不變特征變換0引言(PrincipleComponentAnalysis—Sca
8、leInvariantFeature圖像配準(zhǔn)是同一場景的兩幅或多幅圖像在空間上的對Transform,PCA—SIFT)-sJ、海森仿射