DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal Fusion

Remote sensing image spatiotemporal fusion (STF) aims to generate composite images with high-temporal and spatial resolutions by combining remote sensing images captured at different times and with different spatial resolutions (DTDS). Among the existing fusion algorithms, deep learning-based fusion...

Full description

Saved in:
Bibliographic Details
Main Authors: Yan Zhang, Rongbo Fan, PeiPei Duan, Jinfang Dong, Zhiyong Lei
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10707182/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850150717758636032
author Yan Zhang
Rongbo Fan
PeiPei Duan
Jinfang Dong
Zhiyong Lei
author_facet Yan Zhang
Rongbo Fan
PeiPei Duan
Jinfang Dong
Zhiyong Lei
author_sort Yan Zhang
collection DOAJ
description Remote sensing image spatiotemporal fusion (STF) aims to generate composite images with high-temporal and spatial resolutions by combining remote sensing images captured at different times and with different spatial resolutions (DTDS). Among the existing fusion algorithms, deep learning-based fusion models have demonstrated outstanding performance. These models treat STF as an image super-resolution problem based on multiple reference images. However, compared to traditional image super-resolution tasks, remote sensing image STF involves merging a larger amount of multitemporal data with greater resolution difference. To enhance the robust matching performance of spatiotemporal transformations between multiple sets of remote sensing images captured at DTDS and to generate super-resolution composite images, we propose a feature fusion network called the multiscale deformable convolution distillation generative adversarial network (DCDGAN-STF). Specifically, to address the differences in multitemporal data, we introduce a pyramid cascading deformable encoder to identify disparities in multitemporal images. In addition, to address the differences in spatial resolution, we propose a teacher–student correlation distillation method. This method uses the texture details' disparities between high-resolution multitemporal images to guide the extraction of disparities in blurred low-resolution multitemporal images. We comprehensively compared the proposed DCDGAN-STF with some state-of-the-art algorithms on two landsat and moderate-resolution imaging spectroradiometer datasets. Ablation experiments were also conducted to test the effectiveness of different submodules within DCDGAN-STF. The experimental results and ablation analysis demonstrate that our algorithm achieves superior performance compared to other algorithms.
format Article
id doaj-art-045a5400aa3a49f4b75de61a79ccd30b
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-045a5400aa3a49f4b75de61a79ccd30b2025-08-20T02:26:28ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352024-01-0117194361945010.1109/JSTARS.2024.347615310707182DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal FusionYan Zhang0https://orcid.org/0000-0002-4636-9386Rongbo Fan1https://orcid.org/0000-0003-3284-9685PeiPei Duan2Jinfang Dong3Zhiyong Lei4https://orcid.org/0009-0008-9805-5136School of Mechatronic Engineering, Xi'an Technological University, Xi'an, ChinaSchool of Automation, Northwestern Polytechnical University, Xi'an, ChinaSchool of Computer Science, Xi'an Shiyou University, Xi'an, ChinaShaanxi Meteorological Service Center of Agricultural Remote Sensing and Economic Crops, Xi'an, ChinaSchool of Electronic and Information Engineering, Xi'an Technological University, Xi'an, ChinaRemote sensing image spatiotemporal fusion (STF) aims to generate composite images with high-temporal and spatial resolutions by combining remote sensing images captured at different times and with different spatial resolutions (DTDS). Among the existing fusion algorithms, deep learning-based fusion models have demonstrated outstanding performance. These models treat STF as an image super-resolution problem based on multiple reference images. However, compared to traditional image super-resolution tasks, remote sensing image STF involves merging a larger amount of multitemporal data with greater resolution difference. To enhance the robust matching performance of spatiotemporal transformations between multiple sets of remote sensing images captured at DTDS and to generate super-resolution composite images, we propose a feature fusion network called the multiscale deformable convolution distillation generative adversarial network (DCDGAN-STF). Specifically, to address the differences in multitemporal data, we introduce a pyramid cascading deformable encoder to identify disparities in multitemporal images. In addition, to address the differences in spatial resolution, we propose a teacher–student correlation distillation method. This method uses the texture details' disparities between high-resolution multitemporal images to guide the extraction of disparities in blurred low-resolution multitemporal images. We comprehensively compared the proposed DCDGAN-STF with some state-of-the-art algorithms on two landsat and moderate-resolution imaging spectroradiometer datasets. Ablation experiments were also conducted to test the effectiveness of different submodules within DCDGAN-STF. The experimental results and ablation analysis demonstrate that our algorithm achieves superior performance compared to other algorithms.https://ieeexplore.ieee.org/document/10707182/Deformable convolutiongenerative adversarial network (GAN)remote sensing image spatiotemporal fusion (STF)teacher–student correlation distillation
spellingShingle Yan Zhang
Rongbo Fan
PeiPei Duan
Jinfang Dong
Zhiyong Lei
DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal Fusion
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Deformable convolution
generative adversarial network (GAN)
remote sensing image spatiotemporal fusion (STF)
teacher–student correlation distillation
title DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal Fusion
title_full DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal Fusion
title_fullStr DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal Fusion
title_full_unstemmed DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal Fusion
title_short DCDGAN-STF: A Multiscale Deformable Convolution Distillation GAN for Remote Sensing Image Spatiotemporal Fusion
title_sort dcdgan stf a multiscale deformable convolution distillation gan for remote sensing image spatiotemporal fusion
topic Deformable convolution
generative adversarial network (GAN)
remote sensing image spatiotemporal fusion (STF)
teacher–student correlation distillation
url https://ieeexplore.ieee.org/document/10707182/
work_keys_str_mv AT yanzhang dcdganstfamultiscaledeformableconvolutiondistillationganforremotesensingimagespatiotemporalfusion
AT rongbofan dcdganstfamultiscaledeformableconvolutiondistillationganforremotesensingimagespatiotemporalfusion
AT peipeiduan dcdganstfamultiscaledeformableconvolutiondistillationganforremotesensingimagespatiotemporalfusion
AT jinfangdong dcdganstfamultiscaledeformableconvolutiondistillationganforremotesensingimagespatiotemporalfusion
AT zhiyonglei dcdganstfamultiscaledeformableconvolutiondistillationganforremotesensingimagespatiotemporalfusion