Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification
Convolutional neural network (CNN) has been successfully introduced to hyperspectral image (HSI) classification and achieved effective performance. With the depth of the CNN increases, it may cause the gradient to become zero, and the structure lacks the utilization of the correlated spatial feature...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9809814/ |
| _version_ | 1818195017308569600 |
|---|---|
| author | Lianhui Liang Shaoquan Zhang Jun Li |
| author_facet | Lianhui Liang Shaoquan Zhang Jun Li |
| author_sort | Lianhui Liang |
| collection | DOAJ |
| description | Convolutional neural network (CNN) has been successfully introduced to hyperspectral image (HSI) classification and achieved effective performance. With the depth of the CNN increases, it may cause the gradient to become zero, and the structure lacks the utilization of the correlated spatial feature information between different convolutional layers. At the same time, this single-scale convolution kernel is insufficient in expressing the complex spatial structure information of HSI. In addition, the CNN-based methods treat the HSIs spectral band data as a disordered vector in the process of feature extraction, which abandons the exploitation of its internal spectral correlations. To address these issues, we propose a novel spectral–spatial network classification framework based on multiscale dense connected convolutional network (DenseNet) and bidirection recurrent neural network (Bi-RNN) with attention mechanism network (MDRN). For the proposed MDRN, in terms of spatial feature extraction, a multiscale DenseNet is exploited to combine shallow and deep convolution features to extract the multiscale and complex spatial structure features at each layer. In the aspects of spectral feature extraction, Bi-RNN with attention mechanism is used to capture the inner spectral correlations within a continuous spectrum. Three standard real hyperspectral datasets were used to verify the effectiveness of the proposed MDRN approach. Experimental results indicate that the proposed MDRN method can make full use of the spectral and spatial information of the image, and it has better performance than some advanced algorithms in HSI classification. Finally, in the application of hyperspectral data captured by Gaofen-5 satellite, the practicability of the proposed MDRN method is also superior to other methods. |
| first_indexed | 2024-12-12T01:11:30Z |
| format | Article |
| id | doaj.art-89b0be65ca9d4ed7a7b9a2d06710e907 |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-12-12T01:11:30Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-89b0be65ca9d4ed7a7b9a2d06710e9072022-12-22T00:43:28ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352022-01-01155401541510.1109/JSTARS.2022.31870099809814Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image ClassificationLianhui Liang0https://orcid.org/0000-0001-6958-0443Shaoquan Zhang1https://orcid.org/0000-0002-1454-9665Jun Li2https://orcid.org/0000-0003-1613-9448College of Electrical and Information Engineering, Hunan University, Changsha, ChinaJiangxi Province Key Laboratory of Water Information Cooperative Sensing and Intelligent Processing, School of Information Engineering, Nanchang Institute of Technology, Nanchang, ChinaHubei Key Laboratory of Intelligent Geo-Information Processing, School of Computer Science, China University of Geosciences, Wuhan, ChinaConvolutional neural network (CNN) has been successfully introduced to hyperspectral image (HSI) classification and achieved effective performance. With the depth of the CNN increases, it may cause the gradient to become zero, and the structure lacks the utilization of the correlated spatial feature information between different convolutional layers. At the same time, this single-scale convolution kernel is insufficient in expressing the complex spatial structure information of HSI. In addition, the CNN-based methods treat the HSIs spectral band data as a disordered vector in the process of feature extraction, which abandons the exploitation of its internal spectral correlations. To address these issues, we propose a novel spectral–spatial network classification framework based on multiscale dense connected convolutional network (DenseNet) and bidirection recurrent neural network (Bi-RNN) with attention mechanism network (MDRN). For the proposed MDRN, in terms of spatial feature extraction, a multiscale DenseNet is exploited to combine shallow and deep convolution features to extract the multiscale and complex spatial structure features at each layer. In the aspects of spectral feature extraction, Bi-RNN with attention mechanism is used to capture the inner spectral correlations within a continuous spectrum. Three standard real hyperspectral datasets were used to verify the effectiveness of the proposed MDRN approach. Experimental results indicate that the proposed MDRN method can make full use of the spectral and spatial information of the image, and it has better performance than some advanced algorithms in HSI classification. Finally, in the application of hyperspectral data captured by Gaofen-5 satellite, the practicability of the proposed MDRN method is also superior to other methods.https://ieeexplore.ieee.org/document/9809814/Attention mechanismbidirection recurrent neural network (Bi-RNN)convolutional neural network (CNN)dense connected convolutional network (DenseNet)hyperspectral image (HSI) classification |
| spellingShingle | Lianhui Liang Shaoquan Zhang Jun Li Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Attention mechanism bidirection recurrent neural network (Bi-RNN) convolutional neural network (CNN) dense connected convolutional network (DenseNet) hyperspectral image (HSI) classification |
| title | Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification |
| title_full | Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification |
| title_fullStr | Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification |
| title_full_unstemmed | Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification |
| title_short | Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification |
| title_sort | multiscale densenet meets with bi rnn for hyperspectral image classification |
| topic | Attention mechanism bidirection recurrent neural network (Bi-RNN) convolutional neural network (CNN) dense connected convolutional network (DenseNet) hyperspectral image (HSI) classification |
| url | https://ieeexplore.ieee.org/document/9809814/ |
| work_keys_str_mv | AT lianhuiliang multiscaledensenetmeetswithbirnnforhyperspectralimageclassification AT shaoquanzhang multiscaledensenetmeetswithbirnnforhyperspectralimageclassification AT junli multiscaledensenetmeetswithbirnnforhyperspectralimageclassification |