Wavefusion: Wavelet Assistant Fusion Model for Pan-Sharpening

Xing, Yinghui; Zhang, Yan; Zhang, Yanning

doi:10.1109/igarss46834.2022.9884867

Cited by 3 publications

(8 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditional pansharpening methods can be categorized into three primary approaches: component substitution (CS)-based methods [3][4][5], multi-resolution analysis (MRA)-based methods [6][7][8], and variational optimization (VO)-based methods [9][10][11]. CS-based methods involve transforming the multispectral (MS) image into a different space and substituting its spatial components with those of the panchromatic (PAN) image.…”

Section: Related Work Summarymentioning

confidence: 99%

“…In addition, it is generally believed that the high-frequency and low-frequency sub-bands retain more spatial information and spectral information, respectively [8,10]. Therefore, we design an edge-preserving preprocess to obtain the high-frequency sub-bands and low-frequency sub-bands.…”

Section: Preprocessingmentioning

confidence: 99%

“…In preliminary studies, various traditional methods were proposed to solve pansharpening problems, and most of these methods can be roughly classified into three categories: component substitute (CS)-based methods [3][4][5], multi-resolution analysis (MRA)-based methods [6][7][8], and variational optimization (VO)-based methods [9][10][11]. CS-based methods normally perform well in terms of preserving spatial information, but they may induce spectral distortion.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Triple‐loss driven generative adversarial network for pansharpening

Huang,

Li,

Zhang

2023

IET Image Processing

View full text Add to dashboard Cite

Pansharpening aims at fusing a panchromatic (PAN) image and a low‐resolution multispectral (LRMS) image into a high‐resolution multispectral (HRMS) image. In recent years, GAN‐based pansharpening methods have achieved excellent results, but they suffer from inadequate feature preservation and unstable training. To address these issues, a novel GAN‐based model named TriLossGAN is proposed. This method constructs three loss components with the help of the generator and the dual‐discriminator, which are calculated in both the original spatial domain and the transform domain to better preserve high‐frequency and low‐frequency information in the fused image. Additionally, a new training strategy is designed to stabilize the training process. In extensive experiments, the proposed method achieved satisfactory results on three datasets with QNR values of 0.9584 on GaoFen‐2, 0.9601 on QuickBird, and 0.9138 on WorldView‐3. Qualitative and quantitative comparisons demonstrate that TriLossGAN outperforms other state‐of‐the‐art methods.

show abstract

Section: Related Work Summarymentioning

confidence: 99%

Section: Preprocessingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Triple‐loss driven generative adversarial network for pansharpening

Huang,

Li,

Zhang

2023

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…The traditional pansharpening methods can be categorized into three primary approaches: component substitution (CS)-based methods [5][6][7], multi-resolution analysis (MRA)based methods [8][9][10], and variational optimization (VO)-based methods [11][12][13]. CS-based methods involve transforming the multi-spectral (MS) image into a different space and substituting its spatial components with those of the panchromatic (PAN) image.…”

Section: Introductionmentioning

confidence: 99%

“…Additionally, CNN concern more on local features rather than global features, which consequently leads to difficulties in extracting and fusing significant features from source images, as well as in preserving high-frequency details in the pansharpening results [23]. Another challenge is the presence of artefacts in the fused images, which can significantly lower the quality of the fused images [8]. These artefacts can cause distortions and other anomalies that ultimately undermine the performance of downstream tasks.…”

Section: Introductionmentioning

confidence: 99%

SWPanGAN: A hybrid generative adversarial network for pansharpening

Huang,

Li,

Zhang

2024

IET Image Processing

View full text Add to dashboard Cite

Pansharpening is a vital technique in remote sensing that combines a low‐resolution multi‐spectral image with its corresponding panchromatic image to obtain a high‐resolution multi‐spectral image. Despite its potential benefits, the challenge lies in extracting features from the source images and eliminating artefacts in the fused images. In response to the challenge, a hybrid generative adversarial network‐based model, termed SWPanGAN, is proposed. For better feature extraction, the conventional convolution neural network is replaced with a Swin transformer in the generator, which provides the generator with the ability to model long‐range dependencies. Additionally, to suppress artefacts, a wavelet‐based discriminator is proposed for effectively distinguishing the frequency discrepancy. With these modifications, both the generator and discriminator networks of SWPanGAN are enhanced. Extensive experiments illustrate that our SWPanGAN can generate high‐quality pansharpening images and surpass other state‐of‐the‐art methods.

show abstract