Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery Classification
In recent years, deep learning technology has been widely used in the field of hyperspectral image classification and achieved good performance. However, deep learning networks need a large amount of training samples, which conflicts with the limited labeled samples of hyperspectral images. Traditio...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/2/280 |
| _version_ | 1798031545978585088 |
|---|---|
| author | Liqin Liu Zhenwei Shi Bin Pan Ning Zhang Huanlin Luo Xianchao Lan |
| author_facet | Liqin Liu Zhenwei Shi Bin Pan Ning Zhang Huanlin Luo Xianchao Lan |
| author_sort | Liqin Liu |
| collection | DOAJ |
| description | In recent years, deep learning technology has been widely used in the field of hyperspectral image classification and achieved good performance. However, deep learning networks need a large amount of training samples, which conflicts with the limited labeled samples of hyperspectral images. Traditional deep networks usually construct each pixel as a subject, ignoring the integrity of the hyperspectral data and the methods based on feature extraction are likely to lose the edge information which plays a crucial role in the pixel-level classification. To overcome the limit of annotation samples, we propose a new three-channel image build method (virtual RGB image) by which the trained networks on natural images are used to extract the spatial features. Through the trained network, the hyperspectral data are disposed as a whole. Meanwhile, we propose a multiscale feature fusion method to combine both the detailed and semantic characteristics, thus promoting the accuracy of classification. Experiments show that the proposed method can achieve ideal results better than the state-of-art methods. In addition, the virtual RGB image can be extended to other hyperspectral processing methods that need to use three-channel images. |
| first_indexed | 2024-04-11T19:58:05Z |
| format | Article |
| id | doaj.art-68d5c0fc10954be69fc0acc444d1a225 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-04-11T19:58:05Z |
| publishDate | 2020-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-68d5c0fc10954be69fc0acc444d1a2252022-12-22T04:05:47ZengMDPI AGRemote Sensing2072-42922020-01-0112228010.3390/rs12020280rs12020280Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery ClassificationLiqin Liu0Zhenwei Shi1Bin Pan2Ning Zhang3Huanlin Luo4Xianchao Lan5Image Processing Center, School of Astronautics, Beihang University, Beijing 100191, ChinaImage Processing Center, School of Astronautics, Beihang University, Beijing 100191, ChinaSchool of Statistics and Data Science, Nankai University, Tianjin 300071, ChinaShanghai Aerospace Electronic Technology Institute, Shanghai 201109, ChinaShanghai Aerospace Electronic Technology Institute, Shanghai 201109, ChinaShanghai Aerospace Electronic Technology Institute, Shanghai 201109, ChinaIn recent years, deep learning technology has been widely used in the field of hyperspectral image classification and achieved good performance. However, deep learning networks need a large amount of training samples, which conflicts with the limited labeled samples of hyperspectral images. Traditional deep networks usually construct each pixel as a subject, ignoring the integrity of the hyperspectral data and the methods based on feature extraction are likely to lose the edge information which plays a crucial role in the pixel-level classification. To overcome the limit of annotation samples, we propose a new three-channel image build method (virtual RGB image) by which the trained networks on natural images are used to extract the spatial features. Through the trained network, the hyperspectral data are disposed as a whole. Meanwhile, we propose a multiscale feature fusion method to combine both the detailed and semantic characteristics, thus promoting the accuracy of classification. Experiments show that the proposed method can achieve ideal results better than the state-of-art methods. In addition, the virtual RGB image can be extended to other hyperspectral processing methods that need to use three-channel images.https://www.mdpi.com/2072-4292/12/2/280hyperspectral image classificationfeature extractionfully convolutional networks (fcn)virtual rgb imagemultiscale spatial feature |
| spellingShingle | Liqin Liu Zhenwei Shi Bin Pan Ning Zhang Huanlin Luo Xianchao Lan Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery Classification Remote Sensing hyperspectral image classification feature extraction fully convolutional networks (fcn) virtual rgb image multiscale spatial feature |
| title | Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery Classification |
| title_full | Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery Classification |
| title_fullStr | Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery Classification |
| title_full_unstemmed | Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery Classification |
| title_short | Multiscale Deep Spatial Feature Extraction Using Virtual RGB Image for Hyperspectral Imagery Classification |
| title_sort | multiscale deep spatial feature extraction using virtual rgb image for hyperspectral imagery classification |
| topic | hyperspectral image classification feature extraction fully convolutional networks (fcn) virtual rgb image multiscale spatial feature |
| url | https://www.mdpi.com/2072-4292/12/2/280 |
| work_keys_str_mv | AT liqinliu multiscaledeepspatialfeatureextractionusingvirtualrgbimageforhyperspectralimageryclassification AT zhenweishi multiscaledeepspatialfeatureextractionusingvirtualrgbimageforhyperspectralimageryclassification AT binpan multiscaledeepspatialfeatureextractionusingvirtualrgbimageforhyperspectralimageryclassification AT ningzhang multiscaledeepspatialfeatureextractionusingvirtualrgbimageforhyperspectralimageryclassification AT huanlinluo multiscaledeepspatialfeatureextractionusingvirtualrgbimageforhyperspectralimageryclassification AT xianchaolan multiscaledeepspatialfeatureextractionusingvirtualrgbimageforhyperspectralimageryclassification |