光学学报, 2009, 29 (9): 2607, 网络出版: 2009-10-09
基于非线性相关系数核方法的超谱数据分类
Nonlinear Correlation Coefficient Based Kernel Method for Hyperspectral Data Classification
遥感 超谱数据分类 核方法 非线性相关系数 支持向量机 径向基函数 remote sensing hyperspectral data classification kernel method nonlinear correlation coefficient support vector machine radial basis function
摘要
针对一对一策略的支持向量机算法进行了加权改造,提出一种新的基于非线性相关系数的核方法,在没有地物真实参考图的情况下进一步提高了超谱数据的分类精度。该方法考虑到遥感超谱数据信息依波段分布不均匀的特性,采用非线性相关系数对各波段数据在核函数内部进行加权,使得与参考图相关信息多的波段在分类器中发挥更为显著的作用。同时还提出一种基于非线性相关系数的参考图估计算法,解决了实际应用中真实参考图难以获取的问题。实验对比了采用径向基函数核的支持向量机分类器,结果显示在内部参数为典型值时,所提方法可在无需地物真实参考图的情况下将多分类平均精度和总体精度提高2.90%和3.11%,且运算耗时无明显增加。
Abstract
Under the framework of support vector machines using one against one strategy, a novel kernel method based on nonlinear correlation coefficient is proposed to raise the classification accuracy under the most conditions of no ground truth reference map. This method takes into account the non-uniform information distribution of remote sensing hyperspectral data, and assigns nonlinear correlation coefficients as weights for the corresponding bands to make the band with greater correlation information play a more important role during the process of classification. Meanwhile a new estimated reference map based on nonlinear correlation coefficient is proposed to solve the realistic problem that the real one is usually unavailable. The experimental results show that for the support vector machines based on radial basis function, after adopting the proposed kernels, the average accuracy and the overall accuracy in multi-classification are increased by 2.90% and 3.11% with typical parameter configuration and no ground truth reference map, besides the computational time increment is unobvious.
张淼, 沈毅, 王强. 基于非线性相关系数核方法的超谱数据分类[J]. 光学学报, 2009, 29(9): 2607. Zhang Miao, Shen Yi, Wang Qiang. Nonlinear Correlation Coefficient Based Kernel Method for Hyperspectral Data Classification[J]. Acta Optica Sinica, 2009, 29(9): 2607.