Hyperspectral Data Classification Based on EMAP and Multi-Scale Cloud Model

With the development of hyperspectral remote sensing classification technology, how to obtain higher classification accuracy under the condition of small samples has become a difficult and hot spot in the field of hyperspectral remote sensing technology.In recent years, Cloud model has been used in hyperspectral data analysis.In this paper, a method of hyperspectral remote sensing image classification based on EMAP (Extended Multi-Attribute Profile) and cloud model is proposed.Firstly, EMAP is a kind of texture feature which is obtained by a series of attribute filters filtering hyperspectral image, and is fused with spectral feature.The fused feature is processed by LDA to reduce the dimension, and the data of dimension reduction is taken as the final classification data.Finally, according to the training sample set, the inverse cloud generator is used to generate the multi dimension cloud model for each object, and different scales are added into different cloud models to form multi-scale cloud model.Then, the membership degree of each pixel is calculated by multi-scale cloud model.Finally, we use the maximum decision method to classify each test sample.The experimental results show that the method is simple and efficient, and the classification accuracy is improved effectively.