基于星载大光斑lidar数据反演森林冠层高度与应用的的研究

《基于星载大光斑lidar数据反演森林冠层高度与应用的的研究》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、摘要5)利用GIS软件提取空间分辨率为30m的森林冠层高度和森林地上生物量数据，结果表明：研究区的森林平均冠层高度估测值为18．7m。77个检验样地的森林平均冠层高度估测值为19．2m，预估精度为90．14％。剔除农阳、建筑用地、公路、河流等非林业用地，研究区的森林平均冠层高度估测值为23．1m。研究区森林平均地上生物量估测值为91．017t／ha，去除非林业用地后的森林平均地上生物量估测值为104．561t／ha，研究区森林地上生物量总量为5，747，996t。以上研究结果可为ICESat—GLAS波形数据处理和基于大光斑激光雷达波形数据对林分平均

2、冠层高度和生物量的估测应用研究提供方法基础。关键词ICESat-GLAS；平均冠层高度；反演；小波变换；激光雷达‘AbStract=目_____目E

3、目E_l_l

4、目=g目===审鼍_III=j皇目目自自j}!目=目_目墨_-EE!g===目晕自目E=!=目目-E目目E==目目AbstractLiDAR(LightDetectionandRanging)isoneoftheresearchfocusesonremotesensing，Earthobservation，forestryinvestigationinrecentyears，andfore

5、stsurveyandmappinginChinaandabroadhaveawiderangeofapplications．Therecentstudiesconcentrateontheestimationofforestcanopyheightandforestrybiomass．Therefore，thepapertakeJilinprovince’SWangqingForestryBureauasthestudyarea,whereislocatedinthelowmountainintheChangbaiMountains．Thefirs

6、tchapterreviewsofthestatusandtrendsonLiDARtechnologyapplicationsinforestryathomeandabroad，andanalysisofitsresearchproblems，maincontentsandexpectedobjectivesinthisstudy；Then，thepaperhasanalysisofICESat—GLAS(theIce，Cloud，andlandElevation—GeoscienceLaserAltimeterSystem)returnwavef

7、ormspreprocessingandextractionparametermethod，andalsostudiesonfactors(plotslope，etc．)thateffectonforestcanopyheightestimation．Finally，toverifytherelationshipbetweenICESat-GLASreturnwaveformparametersandthemeanforestcanopyheightandforestabovegroundbiomass，theresearchcombineswith

8、fieldsurveydatatoconstructthemeanforestcanopyheightandforestabovegroundbiomassretrievalmodelbasedonICESat。GLASreturnwavefotinparameters．Meanwhile，theresearchalsousestheGISspatialanalysistechniquestomakefurtherimprovementonthepredictedaccuracyonforestmeancanopyheightandabovegrou

9、ndbiomass．Themainresultsareasfollows：11RespectivelyusingWaveletTransformmethodandGaussianFiltertoobtaintheICESat．GLASreturnwaveformsdenoising，thepaperextractsthedenoisedreturnwaveformparametersforestimationoftheforestmeancanopyheightandforestbiomass．Itsresultsshowthatcomparedwi

10、ththeGaussianFilter，WaveletTransformmakestheRootMeanSq