Deep Learning Method for Hyperspectral Remote Sensing Images with Small Samples

doi:10.16182/j.issn1004731x.joss.201807039

Abstract

Abstract: In order to solve the problem of large information dimension and fewer labeled training samples of hyperspectral remote sensing images, this paper proposes a hyperspectral remote sensing image classification framework HSI-CNN, which reduces the number of model parameters while maintaining the depth of neural network. Image pattern invariance and spectral channel contribution rate are analyzed, and the spectral redundancy information is reduced by principal component analysis. A full convolution neural network structure suitable for small sample hyperspectral remote sensing images is designed and the amount of network parameters is effectively reduced. Three kinds of HSI-CNN structures are proposed and compared. The experimental results on Pavia University and Salinas hyperspectral remote sensing data sets show that HSI-CNN can extract the spectral feature information only by using a small amount of training samples effectively.

Key words: hyperspectral remote sensing image classification, deep learning, convolution neural network, spectral redundancy, principal component analysis

CLC Number:

TP391.4

Shi Xiangbin, Zhong Jian, Liu Cuiwei, Liu Fang, Zhang Deyuan. Deep Learning Method for Hyperspectral Remote Sensing Images with Small Samples[J]. Journal of System Simulation, 2018, 30(7): 2744-2752.

References

[1] Camps-Valls G, Bruzzone L.Kernel-based methods for hyperspectral image classification[J]. IEEE Transactions on Geoscience & Remote Sensing (S0196-2892), 2005, 43(6): 1351-1362.
[2] Cao X, Xu L, Meng D, et al.Integration of 3-dimensional discrete wavelet transform and Markov random field for hyperspectral image classification[J]. Neurocomputing (S0925-2312), 2017, 226(C): 90-100.
[3] Wang Q, Lin J, Yuan Y.Salient band selection for hyperspectral image classification via manifold ranking[J]. IEEE transactions on neural networks and learning systems (S2162-237X), 2016, 27(6): 1279-1289.
[4] Ghamisi P, Benediktsson J A, Sveinsson J R.Automatic Spectral-Spatial Classification Framework Based on Attribute Profiles and Supervised Feature Extraction[J]. IEEE Transactions on Geoscience & Remote Sensing (S0196-2892), 2014, 52(9): 5771-5782.
[5] Xia J, Chanussot J, Du P, et al.Rotation-based support vector machine ensemble in classification of hyperspectral data with limited training samples[J]. IEEE Transactions on Geoscience and Remote Sensing (S0196-2892), 2016, 54(3): 1519-1531.
[6] Wang W, Shen J, Shao L.Video Salient Object Detection via Fully Convolutional Networks[J]. IEEE Transactions on Image Processing, 2017, 27(1): 38-49.
[7] W Wang, J Shen.Deep cropping via attention box prediction and aesthetics assessment[C]//The IEEE International Conference on Computer Vision(ICCV), Piscataway: NJ, 2017: 2186-2194.
[8] He K, Zhang X, Ren S, et al.Deep Residual Learning for Image Recognition[J]. 2015:770-778.
[9] Krizhevsky A, Sutskever I, Hinton G E.ImageNet classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems. Curran Associates Inc. 2012: 1097-1105.
[10] Hu W, Huang Y, Wei L, et al.Deep Convolutional Neural Networks for Hyperspectral Image Classification[J]. Journal of Sensors (S1530-437X), 2015, 2015(2): 1-12.
[11] Chen Y, Lin Z, Zhao X, et al.Deep Learning-Based Classification of Hyperspectral Data[J]. IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing (S1939-1404), 2014, 7(6): 2094-2107.
[12] Ma X, Geng J, Wang H.Hyperspectral image classification via contextual deep learning[J]. Eurasip Journal on Image & Video Processing (S1687-5176), 2015, 2015(1): 20.
[13] 霍雷刚, 冯象初. 基于主成分分析和字典学习的高光谱遥感图像去噪方法[J]. 电子与信息学报, 2014, 36(11): 2723-2729.
Huo Leigang, Feng Xiangchu.Denoising of Hyperspectral Remote Sensing Image Based on Principal Component Analysis and Dictionary Learning[J]. Journal of Electronics and Information Technology, 2014, 36(11): 2723-2729.
[14] Ioffe S, Szegedy C.Batch normalization: Accelerating deep network training by reducing internal covariate shift[C]//International Conference on Machine Learning, 2015: 448-456.
[15] Clevert D A, Unterthiner T, Hochreiter S. Fast and accurate deep network learning by exponential linear units (elus)[J]. arXiv preprint arXiv:1511.07289, 2015.
[16] Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[J]. arXiv preprint arXiv:1409.1556, 2014.