3D密集全卷积的高光谱地物分类算法研究

doi:10.3778/j.issn.1002-8331.2107-0464

摘要/Abstract

摘要： 针对高光谱遥感图像训练样本较少、光谱维度较高、空间特征与频谱特征存在差异性而导致高光谱地物分类的特征提取不合理、分类精度不稳定和训练时间长等问题，提出了基于3D密集全卷积（3D-DSFCN）的高光谱图像（HSI）分类算法。算法通过密集模块中的3D卷积核分别提取光谱特征和空间特征，采用特征映射模块替换传统网络中的池化层和全连接层，最后通过softmax分类器进行分类。实验结果表明，基于3D-DSFCN的HSI分类方法提高了地物分类的准确率、增强了低频标签的分类稳定性。

关键词: 地物分类, 遥感图像, 特征提取, 高光谱图像分类

Abstract: In view of the problems that hyperspectral remote sensing images have few training samples, high spectral dimension, and differences between spatial and spectral features, which lead to unreasonable feature extraction, unstable classification accuracy and long training time, a hyperspectral image（HSI） classification algorithm based on 3D dense full convolution（3D-DSFCN） is proposed. The algorithm extracts spectral features and spatial features through 3D convolution kernel in dense module, replaces pool layer and full connection layer in traditional network with feature mapping module, and finally classifies by softmax classifier. Experimental results show that the HSI classification method based on 3D-DSFCN improves the accuracy of ground object classification and enhances the classification stability of low frequency tags.

Key words: ground object classification, remote sensing image, feature extraction, hyperspectral image classification

厍向阳, 景任杰, 董立红. 3D密集全卷积的高光谱地物分类算法研究[J]. 计算机工程与应用, 2023, 59(3): 112-117.

SHE Xiangyang, JING Renjie, DONG Lihong. Research on Hyperspectral Ground Object Classification Algorithm Based on 3D Dense Full Convolution[J]. Computer Engineering and Applications, 2023, 59(3): 112-117.

参考文献

[1] LI W，DOU Z G，CUI L J，et al.Suitability of hyperspectral data for monitoring nitrogen and phosphorus content in constructed wetlands[J].Remote Sensing Letters，2020，11（5）：495-504.
[2] 王君光，关松.高光谱成像卫星发展研究[J].光电技术应用，2020，35（3）：1-7.
WANG J G，GUAN S.Research on development of hyper-spectral imaging satellites[J].Application of Optoelectronic Technology，2020，35（3）：1-7.
[3] ZHANG Y S，WU L，REN H Z，et al.Mapping water quality parameters in urban rivers from hyperspectral images using a new self-adapting selection of multiple artificial neural networks[J].Remote Sensing，2020，12（2）：336.
[4] 兰玉彬，朱梓豪，邓小玲，等.基于无人机高光谱遥感的柑橘黄龙病植株的监测与分类[J].农业工程学报，2019，35（3）：92-100.
LAN Y B，ZHU Z H，DENG X L，et al.Monitoring and classification of citrus Huanglongbing based on UAV hyperspectral remote sensing[J].Journal of Agricultural Engineering，2019，35（3）：92-100.
[5] LI S T，SONG W W，FANG L Y，et al.Deep learning for hyperspectral image classification：an overview[J].IEEE Transactions on Geoscience and Remote Sensing，2019，57（9）：6690-6709.
[6] 高小梅，李燕丽，卢碧林，等.基于高光谱和数字图像特征指数的受渍冬小麦SPAD估算[J].应用生态学报，2021，32（3）：959-966.
GAO X M，LI Y L，LU B L，et al.Estimation of SPAD value in waterlogged winter wheat based on characteristic indices of hyperspectral and digital image[J].Chinese Journal of Applied Ecology，2021，32（3）：959-966.
[7] 万亚玲，钟锡武，刘慧，等.卷积神经网络在高光谱图像分类中的应用综述[J].计算机工程与应用，2021，57（4）：1-10.
WAN Y L，ZHONG X W，LIU H，et al.Survey of application of convolutional neural network in classification of hyperspectral images[J].Computer Engineering and Applications，2021，57（4）：1-10.
[8] MELGANI F，BRUZZONE L.Classification of hyperspectral remote sensing images with support vector machines[J].IEEE Transactions on Geoscience and Remote Sensing，2004，42（8）：1778-1790.
[9] DEMIR，B，ERTIIRK S.Hyperspectral image classification using relevance vector machines[J].IEEE Geoscience and Remote Sensing Letters，2007，4（4）：586-590.
[10] 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 and Remote Sensing，2014，7（6）：2094-2107.
[11] YANG J，ZHAO Y，CHAN C W.Hyperspectral image classification using two-channel deep convolutional neural network[C]//2016 IEEE International Geoscience and Remote Sensing Symposium，2016：5079-5082.
[12] LI Y，ZHANG H，SHEN Q.Spectral-spatial classification of hyperspectral imagery with 3D convolutional neural network[J].Remote Sensing，2017，9（1）：67-85.
[13] ZHONG Z L，LI J，LUO Z M，et al.Spectral-spatial residual network for hyperspectral image classification：a 3-D deep learning framework[J].IEEE Transactions on Geoscience and Remote Sensing，2018，56（2）：847-858.
[14] WANG W，DOU S，JIANG Z，et al.A fast dense spectral-spatial convolution network framework for hyperspectral images classification[J].Remote Sensing，2018，10（7）：1068.
[15] HAN M X，CONG R M，LI X Y，et al.Joint spatial-spectral hyperspectral image classification based on convolutional neural network[J].Pattern Recognition Letters，2020，130：38-45.
[16] ZHANG B，ZHAO L，ZHANG X L.Three-dimensional convolutional neural network model for tree species classification using airborne hyperspectral images[J].Remote Sensing of Environment，2020，247：111938.
[17] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[18] KONG F Q，ZHOU Y B，SHEN Q，et al.End-to-end multispectral image compression using convolutional neural network[J].Chinese Journal of Lasers，2019，46（10）：1009001.
[19] KRIZHEVSKY A，SUTSKEVER I，HINTON G E.Image net classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems，2012．
[20] 张雄山，赵艮平，程良伦.基于3D深度胶囊网络的高光谱和LiDAR数据融合分类[J].计算机应用研究，2021，38（8）：2526-2529.
ZHANG X S，ZHAO G P，CHENG L L.3D-deep capsule network classification for fused hyperspectral and LiDAR data[J].Application Research of Computers，2021，38（8）：2526-2529.