1.The National Joint Engineering Laboratory of Internet Applied Technology of Mines,China University of Mining and Technology,Xuzhou 221008,China
2.School of Information and Control Engineering,China University of Mining and Technology,Xuzhou 221008,China. Corresponding author: ZHAO Xiao-hu,E-mail: xiaohuzhaozxh@163. com
To address the problems of low resolution, blurred details, and high reconstruction difficulty in remote sensing images, a super-resolution reconstruction algorithm based on spatial adaptive feature modulation, named SARGAN, was proposed. Firstly, regarding the generator, a spatial adaptive multi-scale feature extraction module was proposed, and a multi-level residual nested structure was designed to modulate the extracted global and local features layer by layer. A feature optimization module was introduced to effectively utilize and integrate multi-scale information. Secondly, regarding the discriminator, SARGAN optimized the discriminator structure of the traditional visual geometry group (VGG) network, further improving the stability of network training and the quality of generated images. Finally, experimental results show that compared with other algorithms, SARGAN not only has improvements in peak signal-to-noise ratio and structural similarity but also exhibits good generalization ability. The clarity and structural integrity of the reconstructed images are improved.
DongC, LoyC C, HeK M, et al. Image super-resolution using deep convolutional networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(2): 295-307.
[2]
LimB, SonS, KimH, et al. Enhanced deep residual networks for single image super-resolution[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Honolulu, 2017: 1132-1140.
[3]
KimJ, LeeJ K, LeeK M. Accurate image super-resolution using very deep convolutional networks[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, 2016: 1646-1654.
[4]
ZhangY L, LiK P, LiK, et al. Image super-resolution using very deep residual channel attention networks[C]//Computer Vision-ECCV 2018. Cham: Springer, 2018: 294-310.
[5]
LedigC, TheisL, HuszárF, et al. Photo-realistic single image super-resolution using a generative adversarial network[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, 2017: 105-114.
ZhangW L, LiuY H, DongC, et al. RankSRGAN: super resolution generative adversarial networks with learning to rank[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(10): 7149-7166.
[8]
LiangJ Y, CaoJ Z, SunG L, et al. SwinIR: image restoration using swin transformer[C]//2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). Montreal, 2021: 1833-1844.
[9]
XiaoY, YuanQ Q, JiangK, et al. TTST: a top-k token selective transformer for remote sensing image super-resolution[J]. IEEE Transactions on Image Processing, 2024, 33: 738-752.
[10]
MehriA, BehjatiP, CarpioD, et al. SRFormer: efficient yet powerful transformer network for single image super resolution[J]. IEEE Access, 2023, 11: 121457-121469.
[11]
LeiS, ShiZ W, MoW J. Transformer-based multistage enhancement for remote sensing image super-resolution[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 5615611.
[12]
ZhangH K, LianJ, YiZ T, et al. HAU-Net: Hybrid CNN-Transformer for breast ultrasound image segmentation[J]. Biomedical Signal Processing and Control, 2024, 87: 105427.
[13]
LeiS, ShiZ W. Hybrid-scale self-similarity exploitation for remote sensing image super-resolution[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 5401410.
[14]
SunL, DongJ X, TangJ H, et al. Spatially-adaptive feature modulation for efficient image super-resolution[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV). Paris, 2024: 13144-13153.
[15]
SunL, PanJ S, TangJ H. ShuffleMixer: an efficient ConvNet for image super-resolution[EB/OL]. (2022-05-30)[2024-10-01]2022: arXiv: 2205.15175.
[16]
JinX, XieY P, WeiX S, et al. Delving deep into spatial pooling for squeeze-and-excitation networks[J]. Pattern Recognition, 2022, 121: 108159.
[17]
LiangL H, ZhangY, ZhangS Q, et al. Fast hyperspectral image classification combining transformers and SimAM-based CNNs[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 5522219.
[18]
WangZ Y, LiL L, XueY, et al. FeNet: feature enhancement network for lightweight remote-sensing image super-resolution[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 5622112.
[19]
HarisM, ShakhnarovichG, UkitaN. Deep back-projecti networks for single image super-resolution[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(12): 4323-4337.
[20]
WangX Y, WuY D, MingY, et al. Remote sensing imagery super resolution based on adaptive multi-scale feature fusion network[J]. Sensors, 2020, 20(4): 1142.