基于線特征的規(guī)則物體影像的關(guān)系匹配研究.pdf

1、西安科技大學(xué)碩士學(xué)位論文基于線特征的規(guī)則物體影像的關(guān)系匹配研究姓名：范金健申請學(xué)位級(jí)別：碩士專業(yè)：地圖學(xué)與地理信息系統(tǒng)指導(dǎo)教師：張春森2011Subject：ResearchtoLineFeatureBasedRelationalMatchingofganizedObject’sImagerySpecialty：CartographyGeographicInfmationSystemName：FanJinjian(Signature)I

2、nstruct：ZhangChunSen(Signature)ABSTRACTRecentlywiththerapiddevelopmentofcomputervisionphotogrammetrycloserangephotogrammetryiswidelyappliedinthefieldofarchitectureindustryengineeringarcheologybiomedicaletc.Ithasadvantage

3、sofMatchingisacedifficultissueincasting3Dreconstructionbasedonimage.Mostoftheindustrialpartsusedfmeasurearemadeupoftherulesgeometricelementssuchaslinescirclesetc.butwiththelackofsignificanttexturefeaturepointfeaturematch

4、ingmethodsoftenfail.Tosolvethisprobleminthisthesislinefeaturematchingisusedinthisthesis.Firstlyweusetheimprovedleastsquaretemplatematchingtoextractimagelinefeature.Thenestablishtherelationshipbetweenimagelinefeaturesbuil

5、dtheassociationgraphdeionasabasisfsegmentationbyusingthefaceimageproposedtogenerateasurfacetopologybasedalgithm.Finallytherelationalmatchingbasedonlinearfeaturebyusingthegeneratedtopologicalrelationswhichincludetwosteps:

6、basematchmatchtransmission.Thebasematchistodetermineasetofthesamenamelines.Thedefinitionofthematchtransfmissonisthematchresults’transfmissonbyuseoftheresultofbasematchtopology.Finallyallofthesamenamepointlinepologyfeatur

7、esareobtained.Thispaperstudiestherelationalmatchmethodbasedonlinefeaturewhichgreatlyhighspeedaccurateprecisionnocontactwithtargetsadaptiontokindsofspecialcircumstancesetc.Inspectionasanessentialpartofindustrialproduction

8、yetwithcontinuousimprovementofmanufacturingprocessingisrequiredmemestrictdems.Asakeypartofqualitycontrolinspectionsystemofindustrialpartsplaysaroleinproductionqualityefficiencytosomeextent.Undertheaboveissuesthisthesisdo