基于Grey-Sigmoid核函数支持向量机高光谱遥感图像分类方法研究

摘要

针对传统支持向量机方法中存在的野值噪声敏感问题,提出了一种基于紧密度的Grey-Sigmoid核函数支持向量机,不仅考虑样本与所属类中心之间的关系,还考虑了各个样本之间的距离.通过样本之间的紧密度来描述各个样本之间的关系,利用包围同一类样本的最小超球半径来衡量样本间的紧密度,样本灰度依据样本在球中的位置确定.通过对田间小麦全蚀病的遥感图像分类的实验验证,证明Grey-Sigmoid核函数和传统的Sigmoid核函数相比,计算速度更快,且精度没有明显损失.%Since SVM is sensitive to the noises and outliers in the training set,a new SVM algorithm based on affinity Grey-Sigmoid kernel is proposed in the paper.The cluster membership is defined by the distance from the cluster center,but also defined by the affinity among samples.The remote sensing image classification experiments conducted in the field of wheat prove that,compared with the Grey-Sigmoid kernel function and Sigmoid kernel function of the traditional,it is faster in computational speed and there is no obvious loss in accuracy.