收稿日期:2020-11-04修回日期:2021-01-15第40卷第2期计算机仿真2023年2月文章编号:1006-9348(2023)02-0240-04基于NDVI时间序列数据的湿地演变RSM李婷,唐少霞,王平(海南师范大学地理与环境科学学院,海南海口571158)摘要:针对传统湿地演变遥感动态监测方法中存在的监测效率差,且精度低等问题,提出基于NDVI时间序列数据的湿地演变遥感动态监测方法。掌握NDVI曲线时序形状和数值的特征改变情况,运用兰氏距离与交叉相关光谱匹配,获取湿地演变值指数与形指数,建立湿地演变监测指标;借助Top-Hat算子剔除影像噪声,提升遥感目标与背景间的对比度;探究湿地动态变化动力机制,分析湿地变化源头及趋势,通过湿地动态改变向量,获得各时期湿地类型转移矩阵;建立分类树并区分节点中各类解释标记,将技术、精度、时间和成本作为指标,通过优化理论与多属性效用函数,构建分类优选模型,完成湿地演变遥感信息提取,获得演变趋势监测结果。仿真结果表明,采用所提方法进行湿地演变遥感动态监测的效率较快,且监测精度较高。关键词:归一化植被指数时间序列;湿地演变;遥感监测;演变值指数中图分类号:TP357文献标识码:BRemoteSensingDynamicMonitoringofWetlandEvolutionBasedonNDVITimeSeriesDataLITing,TANGShao-xia,WANGPing(SchoolofGeographyandEnvironmentalScience,HainanNormalUniversity,HaikouHainan571158,China)ABSTRACT:Indeed,thetraditionalremotesensingdynamicmonitoringmethodofwetlandevolutionhaspoormonito-ringefficiencyandlowaccuracy.Inordertoimproveitsefficiencyandaccuracy,aremotesensingdynamicmonitoringmethodofwetlandevolutionbasedonNDVItimeseriesdataisreportedinthispaper.Thetimeseriesshapeandnu-mericalcharacteristicchangesofNDVIcurvewereinvestigated.ThematchingofLAN'sdistanceandcross-correlationspectrumwasappliedtoobtainthewetlandevolutionvalueindexandshapeindexforestablishingthewet-landevolutionmonitoringindex.ImagenoisewasremovedbyapplyingtheTop-Hatoperator.Thecontrastbetweentheremotesensingtargetandbackgroundwasimproved.ThedynamicmechanismofWetlandDynamicchangewassystematicallyinvestigated,andthesourceandtrendofwetlandchangewerealsoanalyzedindetail.Accordingtothewetlanddynamicchangevector,thewetland-typetransfermatrixofeachpe...