中文摘要：

基于MODIS 影像的多种特征信息,以吉林省汪清林业局为例,结合该地区的地形分布,在空间维上分析影像的纹理特征信息和地形指标信息,时间维上分析植被的生长规律和地表温度信息,统计分析植被的特征信息差异,利用决策树方法对该林区的森林类型进行分类,并对分类结果进行精度评价。结果表明,总体分类精度为87.80%,卡帕系数为0.851 0,其中,针叶林、阔叶林、混交林、和其他用地的分类精度均达到80%以上。在分类决策中,时间和空间上的多种数据特征信息的加入,可有效地提高植被类型的分类精度。该结果有助于更好的了解植被在时间和空间上的分布情况,为大区域尺度的森林动态信息监测提供更好的依据。

英文摘要：

Based on multi-feature information of MODIS images, taking the forest areas administrated by Linqing Forestry Bureau, Jilin Province, the paper classified the forest types in this area by decision tree method, in which the accuracy of the classification results was evaluated. The image texture feature information and topographic index information in the spatial dimension were analyzed, while in time dimension, vegetation growth rule and surface temperature information were extracted to analyze statistically the difference in distinct information of the vegetation. The results showed that the overall classification accuracy reached 86.89%, the Kappa coefficient was 0.821 6, and the classification accuracy of Needle leaves-forests, broad leaved-forests, mixed forests, and other land use classification reached more than 80%. Derived by classification, the addition of a variety of the data feature information of time and space in classification decisions could improve effectively the classification accuracy of the vegetation types. Overall, this study could promote the better understanding of the distribution of vegetation in time and space, and provide a better reference for forest dynamic monitoring in large regional scale.