基于空间预处理联合稀疏表示高光谱图像分类

针对高光谱图像分类时光谱信息和空间信息利用不充分、分类精度低的情况, 提出一种结合空间预处理的联合稀疏表示分类方法。一方面能够弥补联合稀疏表示固定窗口模式中空间信息利用不充分的问题, 另一方面也避免了像元多次参与联合稀疏模型的构建过程。考虑每个像元对联合稀疏模型的贡献不同, 通过赋予邻域像元相应权重以提高稀疏重构精度。最后, 充分利用训练样本的已知信息修正分类结果, 在Pavia University和AVIRIS Salinas两个数据集上进行实验验证。实验结果表明, 所提方法能够有效地提高高光谱图像分类精度。

Joint sparse representation hyperspectral image classification based on spatial preprocessing

Aiming at the situation of insufficient utilization of spectral information or spatial information and low classification accuracy in hyperspectral image classification, a joint sparse representation classification method combined with spatial preprocessing is proposed. On the one hand, it can make up for the problem of insufficient utilization of spatial information in the fixed window mode of the joint sparse representation; on the other hand, it also avoids the pixels from participating in the construction process of the joint sparse model multiple times. Considering that each pixel contributes differently to the joint sparse model, the neighboring pixels are given corresponding weights to improve the accuracy of sparse reconstruction. Finally, make full use of the known information of the training samples to correct the classification results. Experiments were conducted on Pavia University and AVIRIS Salinas. The experimental results show that the method proposed in this paper can effectively improve the classification accuracy of hyperspectral images.