基于SPOT5遥感影像的珲春林业局森林分类研究  

作　　者：张黎明[1] 孙亚峰[1] 李娟 

机构地区：[1]吉林省林业调查规划院,吉林长春130022 [2]河北省林业调查规划设计院

出　　处：《山东林业科技》2010年第6期18-20,共3页Journal of Shandong Forestry Science and Technology

摘　　要：利用SPOT5卫星图像对研究区森林类型进行计算机分类,总体精度普遍不高,并且许多森林地类或林分类型的生产者精度和用户精度均比较低,林分类型与检验样本都存在较大偏差。在四种分类方法中,马氏分类法精度最高,总体精度达到57.13%,Kappa系数0.5208。It studies the area forest type using the SPOT5 remote sensing phantom to carry on the computer classification,the overall precision is not generally high.Also,many woodlands class of the Judeich type＇s producer precision and the user precision are quite low,with the Judeich type and the examination sample having the biggest deviation.In four taxonomic approaches,the Mahalanobis classification precision is highest,and the overall precision achieves the 57.13%,Kappa coefficient 0.5208.

关 键 词：SPOT5卫星图像影像分类最大似然法马氏分类法 

分 类 号：S771.8[农业科学—森林工程]