文献类型：期刊文献

中文题名：基于AISA Eagle Ⅱ机载高光谱数据的森林可燃物类型识别方法

英文题名：An Identification Method on Forest Fuel Types based on AISA Eagle Ⅱ Hyperspectral Data

作者：李晓彤[1] 覃先林[1] 刘倩[1] 赵俊鹏[1] 王崇阳[1]

第一作者：李晓彤

机构：[1]中国林业科学研究院资源信息研究所,国家林业局林业遥感与信息技术实验室,北京100091

年份：2021

卷号：36

期号：3

起止页码：544-551

中文期刊名：遥感技术与应用

外文期刊名：Remote Sensing Technology and Application

收录：CSTPCD;;北大核心:【北大核心2020】；CSCD:【CSCD2021_2022】；

基金：中央级公益性科研院所基本科研业务费专项资金课题“基于多源空间数据的森林火灾综合监测技术与应用示范”(2017JYZJ20);国防科工局重大专项“高分森林灾害监测应用示范子系统(二期)”(21-Y30B02-9001-19/22/6)。

语种：中文

中文关键词：森林火灾;森林可燃物类型;AISA EagleⅡ;随机森林

外文关键词：Forest fire;Fuel types;AISA EagleⅡ;Random forest

分类号：TP79

摘要：为形成林场级森林可燃物类型遥感精细识别方法,以内蒙古大兴安岭根河林业局潮查林场为试验区,结合地面调查和森林资源调查等资料,建立了该区域的森林可燃物类型机载高光谱影像数据分类体系;通过对各类型的原始光谱曲线、一阶微分曲线、二阶微分曲线和包络线消除曲线进行分析,得到了适用于各类型识别的特征波段;并基于生成的特征波段的主成分分析分量(信息量95%以上)及其纹理特征影像,采用随机森林方法对该区域的森林可燃物类型进行了识别。利用机载高光谱遥感数据与其他数据相结合,研究形成了林场级森林可燃物类型随机森林识别方法,识别总体精度达86.31%,Kappa系数0.836;兴安落叶松和白桦的制图精度分别达到95.58%、94.34%,表明该方法适宜用于乔木可燃物的细分,可为林场级森林可燃物更新管理、森林火灾的科学预防及扑救提供技术支撑。
In order to provide a feasible scheme for precise identification of fuel types at forest farm level,the Chaocha Forest Farm,Genhe Forestry Bureau,Inner Mongolia,has been selected as the study area. The fuel classification system has been developed based on field survey and forestry resource survey data. The suitable feature bands which were used for identification of forest fuel types have been analyzed and evaluated by different process methods,including original spectral band selection,first derivative reflectance,second derivative reflectance and continuum removed methods. The images generated from the feature bands by using the Principal Component Analysis(PCA)components and their texture features were used to identify the fuel types by Random Forest(RF)method. Combined with the hyperspectral remote sensing data and other data,an identification method for fuel types in forest farms was formed. The results showed that the overall accuracy of the fuel type identification was 86.31% and the Kappa coefficient was 0.836. The producer accuracy for Larix gmelinii and Betula platyphylla was 95.58% and 94.34%,respectively. This method is suitable for the identification of tree fuel types. In addition,it can also provide technical support for updating the fuel parameters and scientifically preventing and extinguishing the forest fires at forest farm levels.