Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing Images
Previously, generative adversarial networks (GAN) have been widely applied on super resolution reconstruction (SRR) methods, which turn low-resolution (LR) images into high-resolution (HR) ones. However, as these methods recover high frequency information with what they observed from the other image...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-03-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/6/1104 |
| _version_ | 1797541433238880256 |
|---|---|
| author | Yuanfu Gong Puyun Liao Xiaodong Zhang Lifei Zhang Guanzhou Chen Kun Zhu Xiaoliang Tan Zhiyong Lv |
| author_facet | Yuanfu Gong Puyun Liao Xiaodong Zhang Lifei Zhang Guanzhou Chen Kun Zhu Xiaoliang Tan Zhiyong Lv |
| author_sort | Yuanfu Gong |
| collection | DOAJ |
| description | Previously, generative adversarial networks (GAN) have been widely applied on super resolution reconstruction (SRR) methods, which turn low-resolution (LR) images into high-resolution (HR) ones. However, as these methods recover high frequency information with what they observed from the other images, they tend to produce artifacts when processing unfamiliar images. Optical satellite remote sensing images are of a far more complicated scene than natural images. Therefore, applying the previous networks on remote sensing images, especially mid-resolution ones, leads to unstable convergence and thus unpleasing artifacts. In this paper, we propose Enlighten-GAN for SRR tasks on large-size optical mid-resolution remote sensing images. Specifically, we design the enlighten blocks to induce network converging to a reliable point, and bring the Self-Supervised Hierarchical Perceptual Loss to attain performance improvement overpassing the other loss functions. Furthermore, limited by memory, large-scale images need to be cropped into patches to get through the network separately. To merge the reconstructed patches into a whole, we employ the internal inconsistency loss and cropping-and-clipping strategy, to avoid the seam line. Experiment results certify that Enlighten-GAN outperforms the state-of-the-art methods in terms of gradient similarity metric (GSM) on mid-resolution Sentinel-2 remote sensing images. |
| first_indexed | 2024-03-10T13:14:59Z |
| format | Article |
| id | doaj.art-e8f3dca42717435b9955313367ff051d |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T13:14:59Z |
| publishDate | 2021-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-e8f3dca42717435b9955313367ff051d2023-11-21T10:29:06ZengMDPI AGRemote Sensing2072-42922021-03-01136110410.3390/rs13061104Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing ImagesYuanfu Gong0Puyun Liao1Xiaodong Zhang2Lifei Zhang3Guanzhou Chen4Kun Zhu5Xiaoliang Tan6Zhiyong Lv7State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), No.129 Luoyu Road, Wuhan 430079, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), No.129 Luoyu Road, Wuhan 430079, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), No.129 Luoyu Road, Wuhan 430079, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), No.129 Luoyu Road, Wuhan 430079, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), No.129 Luoyu Road, Wuhan 430079, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), No.129 Luoyu Road, Wuhan 430079, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), No.129 Luoyu Road, Wuhan 430079, ChinaSchool of Computer and Engineering, Xi’An University of Technology, No.5 Jin Hua South Road, Xi’an 710048, ChinaPreviously, generative adversarial networks (GAN) have been widely applied on super resolution reconstruction (SRR) methods, which turn low-resolution (LR) images into high-resolution (HR) ones. However, as these methods recover high frequency information with what they observed from the other images, they tend to produce artifacts when processing unfamiliar images. Optical satellite remote sensing images are of a far more complicated scene than natural images. Therefore, applying the previous networks on remote sensing images, especially mid-resolution ones, leads to unstable convergence and thus unpleasing artifacts. In this paper, we propose Enlighten-GAN for SRR tasks on large-size optical mid-resolution remote sensing images. Specifically, we design the enlighten blocks to induce network converging to a reliable point, and bring the Self-Supervised Hierarchical Perceptual Loss to attain performance improvement overpassing the other loss functions. Furthermore, limited by memory, large-scale images need to be cropped into patches to get through the network separately. To merge the reconstructed patches into a whole, we employ the internal inconsistency loss and cropping-and-clipping strategy, to avoid the seam line. Experiment results certify that Enlighten-GAN outperforms the state-of-the-art methods in terms of gradient similarity metric (GSM) on mid-resolution Sentinel-2 remote sensing images.https://www.mdpi.com/2072-4292/13/6/1104super resolution reconstructionmid-resolution remote sensing imagesgenerative adversarial network |
| spellingShingle | Yuanfu Gong Puyun Liao Xiaodong Zhang Lifei Zhang Guanzhou Chen Kun Zhu Xiaoliang Tan Zhiyong Lv Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing Images Remote Sensing super resolution reconstruction mid-resolution remote sensing images generative adversarial network |
| title | Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing Images |
| title_full | Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing Images |
| title_fullStr | Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing Images |
| title_full_unstemmed | Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing Images |
| title_short | Enlighten-GAN for Super Resolution Reconstruction in Mid-Resolution Remote Sensing Images |
| title_sort | enlighten gan for super resolution reconstruction in mid resolution remote sensing images |
| topic | super resolution reconstruction mid-resolution remote sensing images generative adversarial network |
| url | https://www.mdpi.com/2072-4292/13/6/1104 |
| work_keys_str_mv | AT yuanfugong enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages AT puyunliao enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages AT xiaodongzhang enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages AT lifeizhang enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages AT guanzhouchen enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages AT kunzhu enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages AT xiaoliangtan enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages AT zhiyonglv enlightenganforsuperresolutionreconstructioninmidresolutionremotesensingimages |