色度空間上基于子塊區(qū)域生長(zhǎng)的彩色圖像分割方法.pdf

1、華中師范大學(xué)碩士學(xué)位論文色度空間上基于子塊區(qū)域生長(zhǎng)的彩色圖像分割方法姓名：蔣蘇蓉申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：馮剛2003.1.1⑩碩士學(xué)位論文MASTER’S”EsISAbstractImagesegmentationisthebasicandimportanttechnologyinImageEngineeringItisafirsttaskofanyimageanalysisprocessbecauseallth

2、esubsequenttasks(featureextractionandobjectrecognitionetc)relyheavilyonthequalityofthesegmentationInrecentyearswiththedevelopmentoftechnologies，suchasmachinevision，pattemrecognitionandcontent—basedimageretrieveetc，andthe

3、wideuseofcolorimages，imagesegmentation，especiallycolorimagesegmentation，hasplayedamoreandmoreimportantroleManyexistingmethodsforimagesegmentationhavebeenstudiedandcomparedinthisthesis，andthenanimprovedregiongrowingalgori

4、thmisproposedcolorimagesegmentationusingchromaspacebasedonblockregiongrowingHSVcolorspacethatisconsistentwithhumaneyesisselected；achromaticinformationisusedastheweighonthehuevalueforstrangepixelsNotonlythepixelproperties

5、butalsothepixelgrouppropertiesareconsideredFirstlyanimageisdividedintomostsuitableblocksForeachblock，themeanandvariancevaluesCanbeseenasthepixelgrouppropertiesTheimagesegmentationisdonewiththeapproachofregiongrowingbased

6、onblockmeanandvarianceConsideringthesizeofextractedobjectsandtheirrelativepositioninformation，getridofthe“noise”objectsandmergethemeaningfulfragmentaryregionsintotheircorrespondingbiggeronesFinallyallpixelsinnon—objectbl