基于表格的影像文檔特征提取算法的研究與設(shè)計.pdf

1、五邑大學(xué)碩士學(xué)位論文基于表格的影像文檔特征提取算法的研究與設(shè)計姓名：解正梅申請學(xué)位級別：碩士專業(yè)：計算機應(yīng)用技術(shù)指導(dǎo)教師：何國輝20080415五邑大學(xué)碩士學(xué)位論文AbstractWiththerepaiddevelopmentofcomputerandmultimediatechnologytheautomaticclassification，storageandmanagementofformimaginga塢becominghot

2、spotsExtractingcharactersofimageisoneofthemostimportantresearchesindocumentimagingmanagementfieldExtractingthereasonablecharactersofimageCanimprovingtheveracityandspeedofdocumentimagingSoitisnecessaryResearchersdidlotsof

3、researchesinformimagingprocessingandmanyachivementshavebeenappliedinsomeindustries，butthosetechnologiesusuallyconcentrateonthespecialformIt’SnotimmatureintheuniversalformrecognitionandanalysisSohowtoextractfeaturesandana

4、lyzetheuniversalformisbecomingthehotspotAccordingtothesereasons，thispaperbasedonimageprocessingandpatternrecognitionItresearchedtwomethodstoextractuniversalform’SfeaturesandhasachievedgoodresultsbyMatlabThemainworksinclu

5、dedthreefacts1Inpreprocessingstage，thepaperputforwardasimpleandeffectivemethodtemoveblanksidesThemethodusedatheoryofmorphologicalItgotanimagewithblacksidesandhadsamesizeasprimaryimagebyperformingopenoperationThenwedelete

6、dblacksidesbysubtractionoperation2Inextractingfeaturesstage，thepaperresearchedafastandpracticalmethodforuniversalformcharactersFormstructure懿thecharactersTheprocessingiscitedasfollowsFirstlyweextractedtablelineSecondlywe

7、mergedordeletedrelativelinesbyalgorithmFinallyweadjustedtheformlinebasedonform’SfeaturesandgotthestructureofformExperimentalresultsshowedthatthemethodwasfeasible3Afteranalyzingthefirstmethod，weprovidedtheotheralgorithmba

8、sedonFourierTransformationWeconsideredverticalandhorizontaldirectionprojectioninformationinformfileandgotspectrumfeaturesbyusingFourierTransformationThenweadoptedBPneuralnetworkasclassifierThosefeaturescansolvethehardcla