蜂群智能算法的遥感影像分类方法.pdf

《蜂群智能算法的遥感影像分类方法.pdf》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、第42卷第5期测绘学报Vo1．42，No．520I3年lO月ActaGeodaeticaetCartographicaS『nica0ct．，2O13CAOMin，SHIZhaoliang，YANGJianyi．AnInnovativeMethodtoClassifyRemoteSensingImagesUsingArtificialBeeColonyAlgorithm[J]．ActaGeodaeticaetCartographicaSinica，2013，42(5)745751．(曹敏，史照良，阳建逸．蜂群智能算法的遥感影像分类方法EJ]．测绘学报，2o13，42(5)：74575

2、1．)蜂群智能算法的遥感影像分类方法曹敏，史照良～，阳建逸1．南京师范大学虚拟地理环境教育部重点实验室，江苏南京210023；2．国土环境与灾害监测国家测绘地理信息局重点实验室，江苏徐州221116；3．江苏省地理信息技术重点实验室，江苏南京210093AnInnovativeMethodtoCIassifyRemoteSensingImagesUsingArtificiaIBeeCoIonyAIaorithmCAOMjn，SHJZhaoljang，YANGJJanYj1．KeyLaboratoryofVirtualGeographicEnvironmentofMinistryof

3、Education，NanJingNormaIUniversity，Nanjing210023，China；2KeyLaboratoryofLandEnvironmentandDisasterMonitoringofSBSM，ChinaUniversityofM}ningandTechnology，Xuzhou221116，China；3．JiangsuKeyLaboratoryofGeographicInformationTechnology，NanJ{ngUniversity，NanJing210093，China．Abstract：Bythecombinationoftheb

4、io—inspiredcomputingmethods，anintelIigentmethodtoclassifyremote—sens—ingimagesusingartificiaIbeecolony(ABC)aIgorithmisproposed．ThisABC—basedcIassificationmethodsearchesfortheoptimaIsplitpointsoneachbandofremote-sensingimageautomaticaIIY，wheretheoptima}solutionofsplitpointsemergefromintelIigent

5、behaviourofhoneybeeswarms．Themethoddefinesbeessearchpathastheconnec—tionbetweenoptimaIsplitpointsandfeaturecategorynode，andconstructsctassificationrulesintheformof‘IF—THEN’．ThissmartapproachcanexpressthecornplexcIassificationmorecIearIYwilhoutmathematicaIformuIas．BvtakinganexampleofALOSimagein

6、thenorthshoreoftheYangtzeRiverestuary，theproposedcIassificationmethodbasedonABCaIgorithmwereimpiementedandtestedagainstSee5．0decision—treemethod．TheoveraIIclassJfJcatJonaccuracyandKappacoefficientofABCJntelJigentmethodarehigherthantheSee5．0decision—treeone．ItisshownthattheABCintelIigentmethodh

7、asbettercIassificationresuIts．Keywords：artificiaIbeecolony；bio—inspiredcornputing；remotesensing；cIassification摘要：结合仿生智能计算方法，提出一种基于人工蜂群算法的智能式遥感分类方法。采用模仿蜜蜂采蜜行为的蜂群智能优化算法，自动搜索遥感影像各波段的最优分割点，定义各波段最优分割点和地物类别节点的连线为蜂群搜索路径，构造以IF-THEN形式表达的遥感分类规则。该