基于经验模态分解的SVM-KNN高光谱图像分类方法

Hyperspectral Image Classification Based on SVM-KNN Method Of Empirical Mode Decomposition

为了提高高光谱图像分类的分类精度,考虑在已知分类器SVM-KNN的基础上,结合经验模态分解,提出了一种EMD-SVM-KNN的新的分类方法,并将其应用到AVIRIS数据92AV3C,仿真结果表明该算法不仅提高了高光谱图像分类精度,而且可减少支持向量数目,以提高高光谱图像分类速度。

In order to improve the classification accuracy of hyperspectral image classification, basisd on the known classifier SVM-KNN, is improued a new classification method of EMD-SVM-KNN is propos, by It is then applied to AVIRIS data 92AV3C. Simulation results show that the algorithm not only improves the hyperspectral image classification accuracy, but also reduces the number of support vectors, so that to improve the speed of hyperspectral image classification.