基于圖像處理的自動對焦理論和技術(shù)研究.pdf

1、武漢理工大學(xué)碩士學(xué)位論文基于圖像處理的自動對焦理論和技術(shù)研究姓名：蔣婷申請學(xué)位級別：碩士專業(yè)：機(jī)械制造及其自動化指導(dǎo)教師：譚躍剛20080401武漢理工大學(xué)碩士學(xué)位論文ABSTRACTAutofocusingtechnologyisakeytechnologyintherobotvisionandimagingsystemandvariousopticalprecisioninstrumentsWiththerapiddevelopme

2、ntofscienceandtechnologytheautofocusingproblemhasbeenmuchconcernedbypeople，anditsapplicationhasbeenmoreandmoreimportantItcanimprovetheaccuracyandspeedofmeasurementautomaticlevelandreducethelaboringintensityetcFromthelate

3、yearsof80’S，domesticandforeignscholarshavemademanyresearchesandachievementsintheautofocusingfieldHoweveratpresenttherearestillmanyshortcomingsinautofocusingtechniquesusedindigitalimagingsystemsbecauseofthecomplexityofima

4、gingobjectsandconditionsSuchashowtoselectfocusfunctionsandfocuswindows，howtoperformfeedbackcontrol，andhowtoimprovetheautofocusingvelocityetc，arefarfromcompletelysolvedFirstlymuchworkofpredecessorsissummarized，andthedevel

5、opmentoftheautofocusingtechniqueathomeandabroadiSintroduced雛wellasthebaSicmethodsofautofocusingBasedonFourieropticaltheoryPSF(PointSpreadFunction)andOTF(OpticalTransferFunction)areintroducedThenthemechanismwhydigitalimag

6、eisfocusedinopticsimagingsystemisalsoanalyzedThen，thethreekeytechnologiesinautofocusingareanalyzedT1lefocusingwindowdirectlyaffectsthespeedandaccuracyofautofocusingBasedontheimportanceofthefocusingwindowthenecessityofwin

7、dowsselectioniSverifiedthroughanalyzingasimpleandfunctionalopticalimagingmodel，atlast，thegeneralstrategyofwindowsselectionindifferentimagingsituationarebroughtforwardAimingattheprecisionandantinoiseabilityofthefunction，b

8、asedonthecomparisonsandresearchesofsomewellknownautofocusingfunctionsandedgedetectionalgorithms，anewclarityevaluationfunctionwasadvancedInthisfunction，animprovedSobelfunctionwasadopted，tosomeextent，theproblemofthelimited

9、orientationsinotherclarityevaluationfunctionsWaSsolvedeffectivelyMeanwhile，standarddeviationasathresholdparameterwasintroducedteducetheinfluenceofnoiseTheresultsshowthatthisfunctionhasthecharacteristicsofpowerfulsingleap

10、ex，highsensitiveandstrongantinosieabilityPeaksearchingalgorithmrealizesthesearchingandorientationofthefocalpoint，thrOughthecomparisonoftheevaluationfunctionvaluestodecidethemovingstepsanddirections，andgetthefeedbacktocon

11、troluntilgetstothebestimagingpositionSeveralcommonlyusedpeaksearchingalgorithmsareintroduced，includingFibonacciSearching，F(xiàn)unctionApproximationMethod，andMountainClimbingSearchingAimingatthelimitsoftraditionalclimbingsearc

12、hing，animprovedMountainClimbingSearchingiSproposedInthelastchapterhardwareandsoftwarearchitecturearealsobrieflydiscussedAlso，theconclusionofmyworkandfutureexpectationareproposedKeywords：Autofocusing，F(xiàn)ocusingWindowSelecti