基于空谱融合与协同表示的高光谱图像分类算法

Hyperspectral image classification algorithm based on spatial spectrum fusion and collaborative representation

高光谱图像光谱带间相似度高且存在大量高维非线性样本,传统的基于表示的分类方法无法对同一波段下的不同样本做出有效区分且会造成维数灾难,最终影响分类性能.提出一种空谱融合与协同表示的高光谱分类算法.通过交替学习空间和光谱特征构建具有判别性的特征字典,并用于空间感知协同表示.在分类过程中,计算特征字典与测试样本之间的相关系数,并将其与误差融合决策.在Indian Pines和Pavia University进行实验,整体精度分别为98.44%和99.22%,验证了本文算法的有效性.

Hyperspectral images have high similarity between spectral bands and a large number of high-dimensional non-linear samples.Traditional representation-based classification methods cannot effectively distinguish different samples in the same band and will cause dimensionality disasters,which ultimately affect classification performance.To this end,a hyperspectral image based on spatial spectrum fusion and collaborative representation algorithm is proposed.To construct a discriminative feature dictionary by alternately learning spatial and spectral features,and use it for spatial-aware collaborative representation.In the classification process,the correlation coefficient between the feature dictionary and the test sample is calculated,and it is combined with the error to make a decision.Experiments were conducted at Indian Pines and Pavia University,and the overall accuracy was 98.44%and 99.22%,respectively,verifying the effectiveness of the algorithm proposed in this paper.