基于云模型的高光谱遥感图像的分类研究

Classification of hyperspectral remote sensing image based on the cloud models

现有的分类方法没有考虑或者没有彻底考虑高光谱遥感图像数据的不确定性,因此提出了一种基于云模型的高光谱遥感图像分类方法。云分类时,首先根据训练样本集,由逆向云发生器生成每类地物的多维云模型,然后利用X条件云发生器计算出各测试像素对每类地物的隶属度,最终采用极大判定法实现对每个测试样本的分类。仿真结果表明,该方法简单、计算量小,可以取得高于传统方法的分类精度,具有很好的发展前景。

The existing methods of classification didn’t consider or thoroughly consider the uncertainty of the hyperspectral remote sensing image data.Therefore propose a classification of hyperspectral remote sensing image based on the cloud models.Firstly,given the training samples,produce the multidimensional cloud models of each cluster by backward cloud generator.Secondly,compute the certainty degree of each test pixel using X condition cloud generator.At last,determine which cluster each test pixel belongs to with the maximum likelihood principle.Simulation results show that this method has the characteristics of simpler,less calculation.It can get higher accuracy than traditional algorithms.In a word,this method has a bright future.