中文摘要：

以吉林省汪清林业局为例,基于Landsat5-TM影像,充分利用遥感影像光谱信息,分别采用动态聚类法和组合监督分类法对该林区的森林类型进行分类,并对分类结果的精度进行比较分析.研究结果表明,利用组合监督分类的精度比动态聚类法分类的精度要高,总体分类精度高11％,其中针叶林、阔叶林、混交林和其他用地的分类精度分别高8％、11％、17％和10％.

英文摘要：

Based on spectral information of Landsat5 - TM remote sensing images, the paper employed the ISODATA classifica- tion method and the combined supervised classification method to respectively classify the forest types in Wangqing Forestry Bureau in Jilin Province. The accuracy of the classification results was analyzed and compared. The results showed that the accuracy of com- bined supervised classification was higher than that of ISODATA classification, with 11% higher in overall classification accuracy and 8% , 11% , 17% , and 10% higher for needle leaves-forest, broad leaved-forest, mixed forest, and other land use classification, respectively.