基于深度特征提取残差网络的高光谱图像分类

Hyperspectral image classification based on deep feature extraction residual network

深度学习由于其模块化设计和强大的特征提取能力,已成为高光谱图像分类的重要手段之一。然而,如何有效地提取更深层次的特征以及同时提高分析空间和光谱联合特征的能力仍是亟待解决的问题。针对这些问题,本文提出了一种深度特征提取的残差网络,该网络由两个关键部分组成:多级传递融合残差网络和空间-光谱多分辨率融合注意力残差网络。多级传递融合残差网络可以有效促进特征信息之间的相互作用,获得更深层次的特征。接着利用空间-光谱多分辨率融合注意力残差网络可以确保从高光谱数据中全面提取空间-光谱联合特征和多分辨率特征。为了验证其有效性,本文在Indian Pines,Pavia University和Salinas Valley三个高光谱数据集上对所提出方法的性能进行了评估,分类精度分别达到了98.10%,99.81%和99.94%。实验结果表明,与其他方法相比,该网络具有更好的泛化能力和分类性能。

Deep learning has become one of the important tools for hyperspectral image classification due to its modular design and powerful feature extraction capability.However,effectively extracting deeper features and simultaneously improving the analysis of spatial and spectral joint features remains an urgent challenge.In response to these issues,a deep feature extraction residual network is proposed in this paper,composed of two key components:a multi-level transfer fusion residual network and a spatial-spectral multi-resolution fusion attention residual network.The multi-level transfer fusion residual network effectively promotes interaction between feature information to obtain deeper-level features.Subsequently,the spatial-spectral multi-resolution fusion attention residual network ensures comprehensive extraction of spatial-spectral joint features and multi-resolution features from hyperspectral data.To validate its effectiveness,the performance of the proposed method was evaluated on three hyperspectral datasets,Indian Pines,Pavia University,and Salinas Valley,achieving classification accuracies of 98.10%,99.81%,and 99.94%respectively.Experimental results demonstrate that,compared to other methods,this network exhibits better generalization capability and classification performance.