遥感影像纹理特征提取及其定量分析

《遥感影像纹理特征提取及其定量分析》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、上海交通大学硕士学位论文遥感影像纹理特征提取及其定量分析姓名：李武跃申请学位级别：硕士专业：模式识别与智能系统指导教师：方涛20080101硕士学位论文方法进行介绍和分析，并完成了“译陆”平台中特征提取模块的设计和实现。结合分割和分类的遥感影像处理应用背景设计高空间分辨率的遥感图像处理模块、较为深入的讨论和分析了各种特征提取方法是本文的主要贡献。全面支持面向遥感影像分类的特征定量分析与特征自动选择，多种分类器分类结果融合等自动处理功能，提升遥感处理平台自动化、智能化处理水平是以后的主要工作方向。关键词：纹理分析，光谱信息，特征提取，遥感技术，图像处

2、理，遥感影像分类硕士学位论文REMOTESENSINGIMAGETEXTUREFEATUREEXTRACTIONANDQUANTITATIVEANALYSISABSTRACTWithChina'ssustainedrapideconomicdevelopmentandtheprocessofurbanizationcontinuestoaccelerate,dynamicdetectingandautomaticupdatingtechnologyinlandchangesbecomeincreasinglyimportant.Remotesens

3、ingtechnologyisintergritedwithcomputertechnology,satellitetechnology,andearthknowledgetocarryoutobservationandmonitoringoftheearth.Thedevelopmentofcomputertechnology,highspatialresolutionimagesinremotesensingmakedetectingdynamicchangesinlandandautomaticupdatespossible.Traditio

4、nalpixel-basedremotesensingimageprocessingmethodsisnotsuitableforrichlydetailedhigh-resolutionremotesensingimageprocessingtechnology.Thereforeanobject-orientedremotesensingimageprocessingmethodhasbeencarriedout.Intheobject-orientedapproach,fromhigh-resolutionremotesensingimage

5、sofspaceobjectsfromtheanalysisoftexturefeatureisakeyfactorofintelligenceofremotesensingprocessingtechnology.RemotesensingresearchGroupatShanghaiJiaotongUniversitydesignedandImplementationedanobject-orientedremotesensingimageprocessingsystem-"TranslationlandingELU",andupdatedto

6、version2.0.Thesoftwarecanprocesshigh硕士学位论文spatialresolutionremotesensingimagesegmentation,featureextraction,classification,monitoringchangesinlandswell,buidinganewbreakthroughsengageinremotesensingimageprocessing.Inthispaper,inconjunctionwiththeproject,featureextractionmethods

7、andanalysisinhigh-resolutionremotesensingimagesareintroduced,anddesigenedandimplementatedFeatureExtractionModuleinthe"translationlandingELU"platform.Combiningsegmentationandclassificationofremotesensingimageprocessingapplicationsbackground,thedesignofthehighspatialresolutionre

8、motesensingimagefeatureextractionmodule,amorein-depthdiscussi