基于遥感影像光谱信息的二叉决策分类树自动生成方法研究

使用决策树方法进行遥感影像分类时需要人工辅助构建决策树,这种方法存在着一旦数据变化就要人工参与重新建立决策树的缺点。针对这一问题,文中提出了依据地物类的样本训练结果,自动构建二叉决策分类树的方法。从阈值的完整性、有效性以及自动选择满足用户需求最佳的建立决策树标准等方面保证决策树的精度以及合理性等。分类结果表明,与最大似然法相比,文中提出算法分类效果良好。

Research on automatically generation of binary decision classification tree based on spectral information

Manual decision tree is needed when using decision tree to classify the remote sensing images at present. However, one of the disadvantages is that decision tree must be rebuilded once data changes. To solve the problem, the paper proposes a method of automatically creating binary decision tree based on spectral statistical information of feature class. The accuracy and rationality of this decision tree is guaranteed by the integrality and validity of threshold, automatic selection of optimal decision tree according to user's requirement. And classification results show that the method is better than maximum likelihood.