论文部分内容阅读
以森林资源遥感分类为切入点,以黄土高原丘陵沟壑区陕西黄龙县境内的水土保持林作为对象,针对ETM+遥感影像在森林信息提取中存在的大量混合像元的问题,引入一种基于针叶林-阔叶林-灌草(C-B-G)模型的混合像元分解方法,通过这种方法,分别得到研究区针叶林、阔叶林、灌草的覆盖图像,并提取出了针阔混交林的分布情况。采用ER-DAS 9.1对分类结果进行精度评价,结果表明针阔混交林的分类精度相对于通用分类方法——监督分类的精度提高了20%,说明该方法可以改善植被信息提取的效果。
Based on remote sensing classification of forest resources, taking the soil and water conservation forests in Huanglong County of Shaanxi Province in the hilly and gully regions of the Loess Plateau as an example, aiming at the problem that a large number of mixed pixels exist in the forest information extraction of ETM + remote sensing images, Forest-broadleaved forest-shrubby grass (CBG) model. By this method, the coverage images of coniferous forest, broad-leaved forest and shrub and grass were obtained, and the mixed coniferous and broad-leaved forest The distribution of The accuracy of the classification results was evaluated by ER-DAS 9.1. The results showed that the classification accuracy of the coniferous-broad-leaved mixed forest was improved by 20% compared with the general classification method-supervised classification, which indicated that this method could improve the effect of vegetation information extraction.