基于bp神经网络的高分辨率遥感影像分类

《基于bp神经网络的高分辨率遥感影像分类》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、摘要本文以2014年9月获取的北京市部分地区的高分二号Pms影像作为实验数据,采用神经网络方法对研究区域进行分类处理及精度评定等研究。利用MATLAB软件进行神经网络程序开发，完成对影响进行分类，并使用误差矩阵和Kappa系数对分类结果进行评定。结果显示BP神经网络分类算法总体精度为81.6%，Kappa系数为0.765，比最大似然发监督分类总体精度提高了6.5%。主要完成了如下的任务和相应的结论：1针对高分辨率影像特征，通过地物多样性和纹理分析，选取具有代表性样区，确定分类类别数，并选取多光谱4个波段信息和纹理特征（通过计算全色波段灰度共生矩阵选取对比度纹

2、理图像）作为分类特征数据。2选取训练样本，设定BP网络结构。包括训练样本归一化处理，中间层神经元数的设定和学习率的确定。利用MATLAB软件完成BP神经网络的设计、分类后处理和精度评定。3将BP神经网络法和最大似然法监督分类结果进行对比。该网络实现总体分类精度为81.6%，Kappa系数为0.7656。最大似然法监督分类总体分类精度为75.1%，Kappa系数0.6886。关键词：BP神经网络、高分辨率影像分类、监督分类、精度评定5959AbstractThispapertaketheGF2highresolutionimageofBeijingcityas

3、basedexperimentaldata,anduseMATLABsoftwaretodeveloptheBPnetworkprogram.Weusethisprogramtoclassifytheremotesensingimages,thenweuseerrormatrixtoevaluatetheaccuracyofclassificationresult.Comparingdifferentclassificationmethods,wefoundthattheoverallclassificationresultofBPneuralnetwork

4、classificationis81.6%,whichismorethanmaximumlikelihoodmethod6.5%.Theinvestigationperformedandrelevantconclusionsareoutlinedasfollows:BasedonthecharacteristicsofGF2image,thispresentsdataprocessingandimageanalyzingofstudyarea,selectingthetypicalsampleareasandeterminingtheamountofcate

5、gory.Andthen,thefeaturedatatobeclassifiedarepreparedbycombiningthemulti-spectralbandswiththecontrast(CON)textureimagefrompanchromaticimage.ThetrainingsampleareaselectedfromthedataofthedataofstudyareaandtheBPnetworkstructureisconstructed.Toensurethestabilityanimprovetheconvergencera

6、te,someapproachesareexploredtoimprovetheBPalgorithm,suchasnormalizedpretreatmentoftrainingsamples,settingpropertrainingrateandadjustingnetworkstructureincludingthenumberofneuroninhiddenlayer.ThecomputerprogramofBPalgorithmisdesignedanddevelopedintheenvironmentofMATLABtoimplementthe

7、GF2imageclassificationandaccuracyevaluatingofstudyarea.Theexperimentsshowthatthismethodcanattainhigh-accuracy5959classificationresult.Comparedwiththoselikemaximumlikelihoodclassificationapproach,theBPhashigheroverallqualityandKappacoefficientthantheMLCwithimprovedby6.5%whichis75.1%

8、.AndtheBPKappais0.7656high