杉木人工林平均树高遥感反演模型研究  

Research on Average Tree Height Inversion Model of Cunninghamia lanceolata Plantation

作　　者：陈利[1] 林辉[1] 孙华[1] 严恩萍[1] 

机构地区：[1]中南林业科技大学林业遥感信息工程研究中心,湖南长沙410004

出　　处：《西南林学院学报》2012年第2期53-56,61,F0003,共6页Journal of Southwest Forestry College

基　　金：国家自然科学基金项目(30871962)资助;国家林业局林业公益项目(201104028)资助;湖南省高等学校科学研究项目(11C1313)资助

摘　　要：以湖南攸县黄丰桥国有林场杉木人工林为研究对象,以高分辨率SPOT5影像及1∶10 000地形图为数据源,提取海拔、坡度、坡向、郁闭度,B1(反射率)、B2(反射率)、B3(反射率)、B4(反射率),B1/B4、B2/B4、B3/B1,EVI、NDVI、RVI等14个因子,运用主成分分析法以及岭迹估计法剔除与平均树高相关性小的变量因子,确定影响平均树高估测的主要因子为:B2(反射率)、B4(反射率)、坡向、郁闭度、NDVI。基于最小二乘法建立遥感反演关系模型,用实地调查数据进行模型检验,平均树高估测回归模型的相关系数、决定系数、调整相关系数以及标准估计的误差分别为0.891 0、0.793 0、0.774 0、0.842 2,树高估测模型达到较好的拟合效果,得到杉木人工林的平均树高模型。Taking the high resolution SPOT5 image and 1：10 000 topographic map as data sources, 14 factors including the elevation, slope gradient, slope aspect, canopy density, the reflectivity of B1 （1st band）, B2 （2nd band）, B3（3rd band）,B4（4th band）, B1/B4, B2/B4, B3/B1, EVI, NDVI and RVI that affected on the tree height estimation of Ctmninghamia lanceolata plantation at Huangfengqiao State-owned Forest Farm, Youxian Coun- ty, Hunan Province were extracted, and finally B2, B4, slope aspect, canopy density and NDVI were determined as the 5 principal factors influencing on the tree height estimation by means of the Principal Component Analysis method and the Ridge Estimation method to eliminate the low-correlation variable factors. The inversion model was built based on the Least Squares method, and the model was testified with field survey data, the correlation coeffi- cient, coefficient of determination, adjustment correlation coefficient and the error of standard assessment of the re- gression model were obtained respectively as 0.891 0, 0. 793 0, 0. 774 0 and 0. 842 2. The results showed that the imitation effect of the tree height estimation model for C. lanceolata plantation was pretty good, and the average tree height estimation model was set up.

关 键 词：杉木人工林 森林结构 最小二乘法 SPOT 5 

分 类 号：S757.2[农业科学—森林经理学]