Multispectral Image Super-Resolution via RGB Image Fusion and Radiometric Calibration

Pan, Zhiwei; Shen, Hui-Liang

doi:10.1109/tip.2018.2881911

Cited by 39 publications

(13 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, most of these methods all assume that the input low-resolution hyperspectral image and the high-resolution auxiliary image are well co-registered. In practical applications, obtaining such well co-registered auxiliary images would be difficult, if not impossible [7], [8], [9].…”

Section: Introductionmentioning

confidence: 99%

Learning Spatial-Spectral Prior for Super-Resolution of Hyperspectral Imagery

Jiang

Sun

Liu

et al. 2020

IEEE Trans. Comput. Imaging

190

134

View full text Add to dashboard Cite

Recently, single gray/RGB image super-resolution reconstruction task has been extensively studied and made significant progress by leveraging the advanced machine learning techniques based on deep convolutional neural networks (DC-NNs). However, there has been limited technical development focusing on single hyperspectral image super-resolution due to the high-dimensional and complex spectral patterns in hyperspectral image. In this paper, we make a step forward by investigating how to adapt state-of-the-art residual learning based single gray/RGB image super-resolution approaches for computationally efficient single hyperspectral image super-resolution, referred as SSPSR. Specifically, we introduce a spatial-spectral prior network (SSPN) to fully exploit the spatial information and the correlation between the spectra of the hyperspectral data. Considering that the hyperspectral training samples are scarce and the spectral dimension of hyperspectral image data is very high, it is nontrivial to train a stable and effective deep network. Therefore, a group convolution (with shared network parameters) and progressive upsampling framework is proposed. This will not only alleviate the difficulty in feature extraction due to high dimension of the hyperspectral data, but also make the training process more stable. To exploit the spatial and spectral prior, we design a spatial-spectral block (SSB), which consists of a spatial residual module and a spectral attention residual module. Experimental results on some hyperspectral images demonstrate that the proposed SSPSR method enhances the details of the recovered high-resolution hyperspectral images, and outperforms state-of-the-arts. The source code is available at https://github.com/junjun-jiang/SSPSR.

show abstract

Section: Introductionmentioning

confidence: 99%

Learning Spatial-Spectral Prior for Super-Resolution of Hyperspectral Imagery

Jiang

Sun

Liu

et al. 2020

IEEE Trans. Comput. Imaging

190

134

View full text Add to dashboard Cite

show abstract

“…The better fusion effect can be obtained; this conclusion can be seen from the aspects of error rate, information entropy, mean value, gradient, similarity, etc. Along with the value of i continues to grow, the quality of fusion image shows a certain decline in the above indicators, especially i takes in Liu et al (2019) , Pan and Shen (2019) , this change is even more obvious. This shows that the accuracy of the calculation has decreased with the change of value i .…”

Section: Experimental and Analysismentioning

confidence: 98%

“…With the continuous recommendation of medical image processing research in recent years, image fusion is an effective solution that automatically detects the information in different images and integrates them to produce one composite image in which all objects of interest are clear. Image fusion ( Zhang et al, 2018 ; Liu et al, 2019 ; Ma et al, 2019 ; Pan and Shen, 2019 ) is a specific algorithm to combine two or more images into a new image. Because of its wide application value, multimodal medical image fusion is an important branch in the field of image fusion.…”

Section: Introductionmentioning

confidence: 99%

Multimodal Medical Supervised Image Fusion Method by CNN

Zhao

et al. 2021

Front. Neurosci.

View full text Add to dashboard Cite

This article proposes a multimode medical image fusion with CNN and supervised learning, in order to solve the problem of practical medical diagnosis. It can implement different types of multimodal medical image fusion problems in batch processing mode and can effectively overcome the problem that traditional fusion problems that can only be solved by single and single image fusion. To a certain extent, it greatly improves the fusion effect, image detail clarity, and time efficiency in a new method. The experimental results indicate that the proposed method exhibits state-of-the-art fusion performance in terms of visual quality and a variety of quantitative evaluation criteria. Its medical diagnostic background is wide.

show abstract

“…Within the spectral remote sensing domain, SR techniques have mainly been used to increase the spatial resolution of the spectral images. Two of the main categories of superresolution techniques that have been used include: single image SR leveraging dictionarybased learning, which learns correlations between low resolution and high resolution data [Liebel and Körner, 2016], and sensor fusion between a low spatial but high spectral resolution imager and a high spatial but low spectral resolution imager [Abd El-Samie et al, 2012;Pan and Shen, 2019;Fang et al, 2018;Ma et al, 2014]. In the dictionary-based learning technique, correlations between the low spatial resolution data and the high spatial resolution data are learned from a prior dataset.…”

Section: State-of-the-art Spectral Super-resolution Techniquesmentioning

confidence: 99%

“…In the sensor fusion technique, correlations between the low spectral resolution and high spectral resolution data are made, assuming sparsity, to achieve a high spectral resolution and spatial resolution output [Abd El-Samie et al, 2012]. In one example, the spectral information from a multispectral camera was fused with a high spatial resolution RGB camera, to achieve a high spatial and spectral resolution image [Pan and Shen, 2019]. While increasing the spatial resolution is useful in heterogeneous regions, these techniques would not provide meaningful information in sub-meter sampled ocean color scenarios where the spatial resolution is only required to separate distinct regions.…”

Section: State-of-the-art Spectral Super-resolution Techniquesmentioning

confidence: 99%

Computational approaches for sub-meter ocean color remote sensing

O'Shea¹

2021

View full text Add to dashboard Cite

Multispectral Image Super-Resolution via RGB Image Fusion and Radiometric Calibration

Cited by 39 publications

References 41 publications

Learning Spatial-Spectral Prior for Super-Resolution of Hyperspectral Imagery

Learning Spatial-Spectral Prior for Super-Resolution of Hyperspectral Imagery

Multimodal Medical Supervised Image Fusion Method by CNN

Computational approaches for sub-meter ocean color remote sensing

Contact Info

Product

Resources

About