论文部分内容阅读
在遥感数据处理研究中,高维高光谱数据的冗余信息和噪声严重影响高光谱数据的分类精度,针对此问题提出基于互信息波段选择和经验模态分解的高精度高光谱数据分类算法(M-IEMD-SVM)。分别采用基于互信息波段选择方法和经验模态分解实现对高光谱数据的冗余信息处理和特征提取,并获得处理后的高光谱数据X″。采用支持向量机分类算法对处理后的高光谱数据X″进行分类实验。仿真实验结果证实MI-EMD-SVM算法不仅提高高光谱数据分类精度,同时还减少支持向量数目,提高高光谱数据分类速度。
In the research of remote sensing data processing, the redundant information and noise of high-dimensional hyperspectral data seriously affect the classification accuracy of hyperspectral data. A high-precision hyperspectral data classification algorithm based on mutual information band selection and empirical mode decomposition M-IEMD-SVM). The redundant information processing and feature extraction of hyperspectral data are realized respectively based on the mutual information band selection method and the empirical mode decomposition, and the processed hyperspectral data X ’’ is obtained. The processed hyperspectral Data X "classification experiment. Simulation results show that the MI-EMD-SVM algorithm can not only improve the classification accuracy of hyperspectral data, but also reduce the number of support vectors and improve the classification speed of hyperspectral data.