PWNet: An Adaptive Weight Network for the Fusion of Panchromatic and Multispectral Images

Liu, Junmin; Feng, Yunqiao; Zhou, Changsheng; Zhang, Chunxia

doi:10.3390/rs12172804

Cited by 14 publications

(5 citation statements)

References 42 publications

(99 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, inspired by the idea of MRA algorithms, MIPSM [16] designs a spatial detail extraction network for the PAN images and injects the details into the LRMS images. Liu et al propose an adaptive weight network for integrating the advantages of different classic methods [14]. It overcomes the shortcomings of the CS and MRA algorithms, and outperforms some SOTA deep learning based methods.…”

Section: Deep Learning Based Methodsmentioning

confidence: 99%

Deep Gradient Projection Networks for Pan-sharpening

Xu¹,

Zhang²,

Zhao³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Pan-sharpening is an important technique for remote sensing imaging systems to obtain high resolution multispectral images. Recently, deep learning has become the most popular tool for pan-sharpening. This paper develops a model-based deep pan-sharpening approach. Specifically, two optimization problems regularized by the deep prior are formulated, and they are separately responsible for the generative models for panchromatic images and low resolution multispectral images. Then, the two problems are solved by a gradient projection algorithm, and the iterative steps are generalized into two network blocks. By alternatively stacking the two blocks, a novel network, called gradient projection based pan-sharpening neural network, is constructed. The experimental results on different kinds of satellite datasets demonstrate that the new network outperforms state-of-the-art methods both visually and quantitatively. The codes are available at https://github. com/xsxjtu/GPPNN .

show abstract

Section: Deep Learning Based Methodsmentioning

confidence: 99%

Deep Gradient Projection Networks for Pan-sharpening

Xu¹,

Zhang²,

Zhao³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Zhou et al [ 36 ] proposed an unsupervised pansharpening network based on perceptual loss and an automatic encoder. Liu et al [ 37 ] based on the fusion results of different adaptive ground averaging methods, combined with the complementary properties of CS and MRA methods, proposed a generalized sharpening weighted network. Li et al [ 38 ] proposed a multi-scale perceptual dense coding convolutional neural network to generate high-quality pansharpened images.…”

Section: Related Workmentioning

confidence: 99%

Pansharpening Model of Transferable Remote Sensing Images Based on Feature Fusion and Attention Modules

Liu

Deng

Dou

et al. 2023

Sensors

View full text Add to dashboard Cite

The purpose of the panchromatic sharpening of remote sensing images is to generate high-resolution multispectral images through software technology without increasing economic expenditure. The specific method is to fuse the spatial information of a high-resolution panchromatic image and the spectral information of a low-resolution multispectral image. This work proposes a novel model for generating high-quality multispectral images. This model uses the feature domain of the convolution neural network to fuse multispectral and panchromatic images so that the fused images can generate new features so that the final fused features can restore clear images. Because of the unique feature extraction ability of convolution neural networks, we use the core idea of convolution neural networks to extract global features. To extract the complementary features of the input image at a deeper level, we first designed two subnetworks with the same structure but different weights, and then used single-channel attention to optimize the fused features to improve the final fusion performance. We select the public data set widely used in this field to verify the validity of the model. The experimental results on the GaoFen-2 and SPOT6 data sets show that this method has a better effect in fusing multi-spectral and panchromatic images. Compared with the classical and the latest methods in this field, our model fusion obtained panchromatic sharpened images from both quantitative and qualitative analysis has achieved better results. In addition, to verify the transferability and generalization of our proposed model, we directly apply it to multispectral image sharpening, such as hyperspectral image sharpening. Experiments and tests have been carried out on Pavia Center and Botswana public hyperspectral data sets, and the results show that the model has also achieved good performance in hyperspectral data sets.

show abstract

“…Here, deep network, namely MS/HS Fusion Net was devised for learning proximal operators and model parameters by CNN. Liu, J et al [64] devised pansharpening weight network (PWNet) for adaptively fusing MSI and HSI. The PWNet was adapted for learning adaptive weight maps using CS-based and MRA-based technique using trainable neural network (NN).…”

Section: Other Deep Learning Modelsmentioning

confidence: 99%

“…From Table 1, it can be evaluated that the MATLAB is frequently used software tool for performing HSI-MSI fusion. [1], [2], [11], [12], [14], [15], [16], [17], [19], [20], [22], [23], [26], [28], [31], [33], [32], [34], [36], [ 39], [42], [43], [44], [45], [46], [47], [55], [56], [64], [66], [70], [71], [74], [77], [78] Python [18], [27], [61], [62], [76] Tensorflow [3], [4], [29], [53], [60], [67], [68] 4…”

Section: Analysis On the Basis Of Implementation Toolmentioning

confidence: 99%

Brain Tumor Classification using the Probabilistic-Deformable Fuzzy system and RideNN

Srishailam¹

2022

View full text Add to dashboard Cite

The fusion of multispectral (MSI) and hyperspectral images (HSI) has been acknowledged as a promising method for performing HSI-MSI fusion, which is also termed to be an essential part for precise recognition and cataloging of the underlying materials. The goal of HSI-MSI fusion is to fuse the low-resolution HSI with the high-resolution MSI to produce high-resolution images. The HSI-MSI fusion is employed as an imperative part for addressing the problems of image processing. However, the HSI-MSI fusion needs high resolution image to perform precise analysis and decision making, which is the major issue due to deficiencies of current satellite sensors that seems impossible to attain quality images. Numerous techniques are devised in the prior works that employed image fusion using HSI and MSI. This paper presents complete survey of 80 papers using HSI-MSI fusion methodologies, which involves several techniques like Pansharpening methods, Subspace-based methods, Artificial intelligence-based methods, Deep learning methods and Hybrid models. In addition, thorough investigation are performed based on the year of publication, adapted methodology, datasets used, implementation tool, evaluation metrics, and values of evaluation metrics. Finally, the issues of existing methods and the research gaps considering conventional HSI-MSI fusion schemes are elaborated to obtain improved contribution for devising significant HSI-MSI fusion techniques.

show abstract

PWNet: An Adaptive Weight Network for the Fusion of Panchromatic and Multispectral Images

Cited by 14 publications

References 42 publications

Deep Gradient Projection Networks for Pan-sharpening

Deep Gradient Projection Networks for Pan-sharpening

Pansharpening Model of Transferable Remote Sensing Images Based on Feature Fusion and Attention Modules

Brain Tumor Classification using the Probabilistic-Deformable Fuzzy system and RideNN

Contact Info

Product

Resources

About