基于组合核函数相关向量机的高光谱图像分类

Hyperspectral Image Classification Based on Composite Kernel Relevance Vector Machine

本文提出了一种基于组合核函数相关向量机的高光谱图像分类方法。基于核函数的特性构造出三种形式的组合核函数。使用多尺度数学形态学方法从主成分变换后的图像上提取空间特征,通过组合核函数融合图像的光谱特征与空间特征,采用组合核函数相关向量机进行分类。使用AVIRIS高光谱图像对算法进行了验证。实验结果表明,与传统的基于光谱特征的相关向量机分类器相比,组合核函数相关向量机方法的总体精度和Kappa系数均有明显提升。同时,组合核函数相关向量机能够用较少的训练样本获得较高的分类精度,在高光谱图像分类中具有实用价值。

This paper presents a composite kernel Relevance Vector Machine (RVM) algorithm for hyperspec-tral image classification. This paper constructs three forms of composite kernels based on the properties of kernels. The spatial feature is extracted using multi-scale morphological method from the image after principal components transform. The final classification is achieved by our composite kernel RVM algorithm. The proposed approach is tested in experiments on AVIRIS data. Com-pared with spectral kernel RVM, the overall accuracy and Kappa coefficient of the composite kernel RVM algorithm increased obviously. However, the training time dose not increase. Meanwhile, composite kernel RVM has ability to get high accuracy with relative small raining set. The proposed method has practical use in hyperspectral imagery classification.