以长白山汪清林区为例,分析了星载激光雷达( ICESat-GLAS)数据在森林类型识别上的应用效果。采用软件Matlab和IDL对原始二进制数据进行处理,得到GLAS回波波形图;进一步提取与森林类型相关的波形特征参数,作为支持向量分类机( C-SVC)的输入量,进行森林类型识别,并采用K-折交叉验证方法对核函数选择进行评价。结果表明:C-SVC分类方法能够识别阔叶林和针叶林2种森林类型,识别精度达到85.24%。
InWangqing area of Changbai Mountains, the investigation was conducted to analyze the capability of the ICESat-GLAS (Ice, Cloud and Land Elevation Satellite-Geoscience Laser Altimeter System) data on the forest type identification. The original binary ICESat-GLAS data was processed to obtain waveforms by MATLAB and IDL.The GLAS waveform met-rics were derived as the input dataset of C-Support Vector Classification method (C-SVC) for forest type identification, and the kernel function selection was evaluated using K-fold cross-validation method.The C-SVC method is capable to classify broadleaf and coniferous forests with the accuracy of 85 .24%.