|Table of Contents|

Shorten Special-spectral Parallel Bidirectional RNN for Hyperspectral Agricultural Image Classification(PDF)

南京师范大学学报（工程技术版）[ISSN:1006-6977/CN:61-1281/TN]

Issue:
2022年04期
Page:
1-8
Research Field:
计算机科学与技术
Publishing date:

Info

Title:
Shorten Special-spectral Parallel Bidirectional RNN for Hyperspectral Agricultural Image Classification
Author(s):
Wang Maofa1; Feng Shichen1; Huang Hongliang2; Gong Qizhou3; Wan Quan1; Xu Zhi1
(1.Guangxi Key Laboratory of Trusted Software,Guilin University of Electronic Technology,Guilin 541004,China)
(2.Department of Mathematics,School of Science and Technology,University of Macau,Macau 999078,China)
(3.School of Applied Science,Beijing Information Science and Technology University,Beijing 100192,China)
Keywords:
deep learning; gate recurrent unit(GRU); long short term mermory network(LSTM); recurrent neural networks(RNN); hyperspectral image classification
PACS:
TP391.4
DOI:
10.3969/j.issn.1672-1292.2022.04.001
Abstract:
A method is proposed for HSI classification using shorten spatial-spectral parallel bidirectional RNN(St-SS-pBRNN). By combining multi-convolutional layers,it not only uses the frequency information of spectrum but also considers the spatial characteristics in the images,thereby obtaining better performance and effects. Meanwhile,a combined architecture of parallel gate recurrent unit(GRU)and bidirectional RNN is adopted,which shortens the sequence length of RNN and greatly reduces the amount of calculation of the model. In the comparative experiments of agricultural hyperspectral image classification,the performance of the algorithm is stable and accurate,and its accuracy is more than 2% higher than the optimal effect of the classical shorten spatial-spectral parallel GRU(St-SS-pGRU). It is expected to be popularized and applied in a wide range of agricultural land classification in China.

References:

[1]BIOUCAS-DIAS J M,PLAZA A,CAMPS-VALLS G,et al. Hyperspectral remote sensing data analysis and future challenges[J]. IEEE Geoence and Remote Sensing Magazine,2013,1(2):6-36.
[2]舒速,杨明,赵振凯. 基于分水岭的高光谱图像分类方法[J]. 南京师大学报(自然科学版),2015,38(1):91-97.
[3]朱志宾,丁世飞. 基于TWSVM的图像分类[J]. 南京师大学报(自然科学版),2014,37(3):8-14.
[4]王岽,吴见. 农作物各类高光谱遥感识别研究[J]. 地理与地理信息科学,2015,31(2):29-33.
[5]ZHU X X,TUIA D,MOU L,et al. Deep learning in remote sensing:a comprehensive review and list of resources[J]. IEEE Geoence and Remote Sensing Magazine,2018,5(4):8-36.
[6]陈跃. 改进可拓理论的带钢表面缺陷图像分类方法[J]. 南京师范大学学报(工程技术版),2016,16(3):54-62.
[7]LEE H,KWON H. Going deeper with contextual CNN for hyperspectral image classification[J]. IEEE Transactions on Image Processing,2016,26(10):4843-4855.
[8]ZHONG Z L,LI J,MA L F,et al. Deep residual networks for hyperspectral image classification[C]//2017 IEEE International Geoscience and Remote Sensing Symposium. Fort Worth,TX,USA,2017.
[9]MOU L,GHAMISI P,ZHU X X. Deep recurrent neural networks for hyperspectral image classification[J]. IEEE Transactions on Geoence and Remote Sensing,2017,55(7):3639-3655.
[10]SUNDERMEYER M,OPARIN I,GAUVAIN J L,et al. Comparison of feedforward and recurrent neural network language models[C]//2013 IEEE International Conference on Acoustics,Speech,and Signal Processing. Vancouver,BC,Canada,2013.
[11]FELIX A G,DOUGLAS E,JÜRGEN S. Applying LSTM to time series predictable through time-window approaches[C]//Proceedings of the 12th Italian Workshop on Neural Nets. Salerno,Italy:Springer,2002:193-200.
[12]BENGIO Y,SIMARD P,FRASCONI P. Learning long-term dependencies with gradient descent is difficult[J]. IEEE Transactions on Neural Networks,1994,5(2):157-166.
[13]WILLIAMS R J,ZIPSER D. A learning algorithm for continually running fully recurrent neural networks[J]. Neural Computation,2014,1(2):270-280.
[14]LI L,ZHANG H,SHEN Q. Spectral-spatial classification of hyperspectral imagery with 3D convolutional neural network[J]. Remote Sensing,2017,9(1):67.
[15]ZHONG Z L,LI J,LUO Z M. Spectral-spatial residual network for hyperspectral image classification:a 3-D deep learning framework[J]. IEEE Transactions on Geoscience and Remote Sensing,2017,56(2):847-858.
[16]李冠东,张春菊,高飞,等. 双卷积池化结构的3D-CNN高光谱遥感影像分类方法[J]. 中国图象图形学报,2019,24(4):639-654.
[17]魏健,赵红涛,刘敦楠,等. 基于注意力机制的CNN-LSTM短期电力负荷预测方法[J]. 华北电力大学学报(自然科学版),2021,48(1):42-47.
[18]LIU Q S,ZHOU F,HANG R L,et al. Bidirectional-convolutional LSTM based spectral-spatial feature learning for hyperspectral image classification[J]. Remote Sensing,2017,9(12):1330.
[19]王文广,赵文杰. 基于GRU神经网络的燃煤电站NOx排放预测模型[J]. 华北电力大学学报(自然科学版),2020,47(1):96-103.
[20]XU Y H,DU B,ZHANG L P,et al. A band grouping based LSTM algorithm for hyperspectral image classification[C]//CCF Chinese Conference on Computer Vision. Singapore:Springer,2017:421-432.
[21]SCHUSTER M,PALIWAL K K. Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing,1997,45(11):2673-2681.
[22]LUO H W. Shorten spatial-spectral RNN with parallel-GRU for hyperspectral image classification[J]. arXiv Preprint arXiv:1810.12563,2018.

Memo

Memo:
-
Last Update: 2022-12-15