资源描述：
《基于几何区域的灰度共生矩阵和region+map的sar图像分割方法》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。
1、万方数据AbstractSARimagesegmentationisoneofthebasicandcrucialtechnologyinSARimageprocessingandinterpreting.Thesegmentationresultshavegreatinfluenceonthesubsequentimageprocessing.Becauseoftheuniqueimagingmechanism,theregionsgatheredinSARimagesalwayshavetheregularshadegrayscalecharacteristicandthis
2、leadstotheprobleminexistingSARimagesegmentationmethodsthatthesameareamaybeclassedintodifferentareas.Theprimarysketchmodelisthesparserepresentationoftheimage,withitssketchlinesdepictingthegraychangesoftheimage.SARimagesegmentationmethodbasedonPrimarySketchMapandsemanticinformation,proposedbyLi
3、uFangandYuanjialin,analyzedtheaggregationofneighborsketchlines,dividedsketchlinestoclusteredlinesandnon-clusteredlinesaccordingtostatisticaldistributionoflines’aggregation.Thendifferentareaextractingoperatorweredesignedtoextractmoreabstractareamaponthesketchmapandintermediatesemanticlayerbase
4、donareaswasconstructedandsegmentationresultwasgottencombinedwiththetraditionalsegmentationmethod..BasedonthementionedmethodaswellasthesketchlinesandtheGLCM,thispaperproposesaSARimagesegmentationmethodbasedonGLCMandRegionMapofthegeometryarea,Themainworkofthepaperisasfollows:First,fortheproblem
5、sexistingintheSARimagesegmentation,weanalyzetheprimarysketchmodeltheoryandtheGLCMtheory.WiththegeometryinformationoftheSARimagesketchlinesgottenfromtheprimarysketchmodel,wedefinethegeometryregionofthesketchlines.WecountuptheGLCMofthegeometryregionswiththesketchlinestoshowthedifferentsemantici
6、nformationofthesketchlines.Wealsodividethesketchlinesintotwoclasseswiththefirstclassdepictingthebordersandthegraychanges,andtheseconddepictingtheshadegraychanges.Basedontheclassificationandtheclusteringfeaturesofthelines,weextracttheregionmapofthegatheringregion,borderregion,non-lineregion.Se
7、cond,basedontheRegionMapextractedfromthesketch,thispaperpresentsaSARimagesegmentationmethodbasedonRegionMap.WiththeRegionMap,themethodmapstheoriginalSARimageintogatheringregion,borderregion,non-lineregion.WesegmenttheSARimageintomanysuper-pix
