An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection
Change detection (CD) in hyperspectral images has become a research hotspot in the field of remote sensing due to the extremely wide spectral range of hyperspectral images compared to traditional remote sensing images. It is challenging to effectively extract features from redundant high-dimensional...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-09-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/18/4646 |
| _version_ | 1797482736227713024 |
|---|---|
| author | Hongyu Zhao Kaiyuan Feng Yue Wu Maoguo Gong |
| author_facet | Hongyu Zhao Kaiyuan Feng Yue Wu Maoguo Gong |
| author_sort | Hongyu Zhao |
| collection | DOAJ |
| description | Change detection (CD) in hyperspectral images has become a research hotspot in the field of remote sensing due to the extremely wide spectral range of hyperspectral images compared to traditional remote sensing images. It is challenging to effectively extract features from redundant high-dimensional data for hyperspectral change detection tasks due to the fact that hyperspectral data contain abundant spectral information. In this paper, a novel feature extraction network is proposed, which uses a Recurrent Neural Network (RNN) to mine the spectral information of the input image and combines this with a Convolutional Neural Network (CNN) to fuse the spatial information of hyperspectral data. Finally, the feature extraction structure of hybrid RNN and CNN is used as a building block to complete the change detection task. In addition, we use an unsupervised sample generation strategy to produce high-quality samples for network training. The experimental results demonstrate that the proposed method yields reliable detection results. Moreover, the proposed method has fewer noise regions than the pixel-based method. |
| first_indexed | 2024-03-09T22:37:45Z |
| format | Article |
| id | doaj.art-670ce0b9ec1d476784286dfae348617a |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T22:37:45Z |
| publishDate | 2022-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-670ce0b9ec1d476784286dfae348617a2023-11-23T18:46:10ZengMDPI AGRemote Sensing2072-42922022-09-011418464610.3390/rs14184646An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change DetectionHongyu Zhao0Kaiyuan Feng1Yue Wu2Maoguo Gong3School of Electronic Engineering, Xidian University, Xi’an 710071, ChinaSchool of Electronic Engineering, Xidian University, Xi’an 710071, ChinaSchool of Computer Science and Technology, Xidian University, Xi’an 710071, ChinaSchool of Electronic Engineering, Xidian University, Xi’an 710071, ChinaChange detection (CD) in hyperspectral images has become a research hotspot in the field of remote sensing due to the extremely wide spectral range of hyperspectral images compared to traditional remote sensing images. It is challenging to effectively extract features from redundant high-dimensional data for hyperspectral change detection tasks due to the fact that hyperspectral data contain abundant spectral information. In this paper, a novel feature extraction network is proposed, which uses a Recurrent Neural Network (RNN) to mine the spectral information of the input image and combines this with a Convolutional Neural Network (CNN) to fuse the spatial information of hyperspectral data. Finally, the feature extraction structure of hybrid RNN and CNN is used as a building block to complete the change detection task. In addition, we use an unsupervised sample generation strategy to produce high-quality samples for network training. The experimental results demonstrate that the proposed method yields reliable detection results. Moreover, the proposed method has fewer noise regions than the pixel-based method.https://www.mdpi.com/2072-4292/14/18/4646change detection (CD)recurrent neural network (RNN)convolutional neural network (CNN)hyperspectral image (HSI) |
| spellingShingle | Hongyu Zhao Kaiyuan Feng Yue Wu Maoguo Gong An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection Remote Sensing change detection (CD) recurrent neural network (RNN) convolutional neural network (CNN) hyperspectral image (HSI) |
| title | An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection |
| title_full | An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection |
| title_fullStr | An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection |
| title_full_unstemmed | An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection |
| title_short | An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection |
| title_sort | efficient feature extraction network for unsupervised hyperspectral change detection |
| topic | change detection (CD) recurrent neural network (RNN) convolutional neural network (CNN) hyperspectral image (HSI) |
| url | https://www.mdpi.com/2072-4292/14/18/4646 |
| work_keys_str_mv | AT hongyuzhao anefficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection AT kaiyuanfeng anefficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection AT yuewu anefficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection AT maoguogong anefficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection AT hongyuzhao efficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection AT kaiyuanfeng efficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection AT yuewu efficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection AT maoguogong efficientfeatureextractionnetworkforunsupervisedhyperspectralchangedetection |