Hyperspectral and Multispectral Image Fusion Based on Local Low Rank and Coupled Spectral Unmixing

Zhou, Yuan; Feng, Liyang; Hou, Chunping; Kung, Sun‐Yuan

doi:10.1109/tgrs.2017.2718728

Cited by 57 publications

(17 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The future availability of systematic global archives of both hyperspectral and short revisit time multispectral observations is expected to offer significant potential for studies that leverage both the spectral resolution of hyperspectral sensors and the temporal resolution of multispectral sensors. A growing number of data fusion techniques attempt to explicitly merge hyperspectral and multispectral data (e.g., [ 9 , 10 ]). More generally, hyperspectral imagery can be used to better understand fundamental properties of multispectral data.…”

Section: Introductionmentioning

confidence: 99%

Multisensor Analysis of Spectral Dimensionality and Soil Diversity in the Great Central Valley of California

Sousa

Small

2018

Sensors

View full text Add to dashboard Cite

Planned hyperspectral satellite missions and the decreased revisit time of multispectral imaging offer the potential for data fusion to leverage both the spectral resolution of hyperspectral sensors and the temporal resolution of multispectral constellations. Hyperspectral imagery can also be used to better understand fundamental properties of multispectral data. In this analysis, we use five flight lines from the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) archive with coincident Landsat 8 acquisitions over a spectrally diverse region of California to address the following questions: (1) How much of the spectral dimensionality of hyperspectral data is captured in multispectral data?; (2) Is the characteristic pyramidal structure of the multispectral feature space also present in the low order dimensions of the hyperspectral feature space at comparable spatial scales?; (3) How much variability in rock and soil substrate endmembers (EMs) present in hyperspectral data is captured by multispectral sensors? We find nearly identical partitions of variance, low-order feature space topologies, and EM spectra for hyperspectral and multispectral image composites. The resulting feature spaces and EMs are also very similar to those from previous global multispectral analyses, implying that the fundamental structure of the global feature space is present in our relatively small spatial subset of California. Finally, we find that the multispectral dataset well represents the substrate EM variability present in the study area -despite its inability to resolve narrow band absorptions. We observe a tentative but consistent physical relationship between the gradation of substrate reflectance in the feature space and the gradation of sand versus clay content in the soil classification system.

show abstract

Section: Introductionmentioning

confidence: 99%

Multisensor Analysis of Spectral Dimensionality and Soil Diversity in the Great Central Valley of California

Sousa

Small

2018

Sensors

View full text Add to dashboard Cite

show abstract

“…where λ is the balancing parameter to control the tradeoff between L rec and L cos . Second, according to the observation model, the LR MSI patchŶ L is the spectral degradation of the LR HSI patchX L [14], which can be formulated aŝ…”

Section: Loss Functionmentioning

confidence: 99%

“…Over decades, many methods [8,9] have been proposed to reconstruct the desired HR HSI by fusing HR MSIs and LR HSIs, including sparse representation-based methods [10,11], Bayesian-based methods [12,13], spectral unmixing-based methods [1,14], and tensor factorization-based methods [15,16]. Sparse representation-based, Bayesianbased, and spectral unmixing-based methods usually first learn spectral bases (or endmembers) from the LR HSI [9,10].…”

Section: Introductionmentioning

confidence: 99%

Hyperspectral Image Super-Resolution with Self-Supervised Spectral-Spatial Residual Network

Chen

Zheng

2021

Remote Sensing

View full text Add to dashboard Cite

Recently, many convolutional networks have been built to fuse a low spatial resolution (LR) hyperspectral image (HSI) and a high spatial resolution (HR) multispectral image (MSI) to obtain HR HSIs. However, most deep learning-based methods are supervised methods, which require sufficient HR HSIs for supervised training. Collecting plenty of HR HSIs is laborious and time-consuming. In this paper, a self-supervised spectral-spatial residual network (SSRN) is proposed to alleviate dependence on a mass of HR HSIs. In SSRN, the fusion of HR MSIs and LR HSIs is considered a pixel-wise spectral mapping problem. Firstly, this paper assumes that the spectral mapping between HR MSIs and HR HSIs can be approximated by the spectral mapping between LR MSIs (derived from HR MSIs) and LR HSIs. Secondly, the spectral mapping between LR MSIs and LR HSIs is explored by SSRN. Finally, a self-supervised fine-tuning strategy is proposed to transfer the learned spectral mapping to generate HR HSIs. SSRN does not require HR HSIs as the supervised information in training. Simulated and real hyperspectral databases are utilized to verify the performance of SSRN.

show abstract

“…Multiple sensors can reflect more complete characteristics and information. The hyperspectral (HS) remote sensing sensors provide the HS imagery (HSI) which has abundant spectral information [1]. The Panchromatic (PAN) remote sensing sensors are capable of providing the PAN image (PANI) that possesses high spatial resolution (HSR).…”

Section: Introductionmentioning

confidence: 99%

Hyperspectral and Panchromatic Image Fusion via Adaptive Tensor and Multi-Scale Retinex Algorithm

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Fusion of panchromatic image (PANI) and hyperspectral image (HSI) to obtain an output image with high spatial and spectral resolutions has received increasing interests recently. We propose a new image fusion method for HSI and PANI by combining adaptive tensor with a multi-scale Retinex algorithm in this paper. In the proposed method, an adaptive tensor based method is presented to effectively extract the structure information of HSI, and multi-scale Retinex algorithm is introduced to obtain the spatial and structure details of PANI. To integrate spatial structure information, a gradient-based weighted fusion strategy is proposed to combine spatial details of HSI and PANI. The integrated structure details are injected to generate the fused HSI. Experiments using both simulated and real remote sensing data sets demonstrated that the proposed fusion algorithm performs better than the state-of-the-art algorithms in visual inspection and objective assessment. INDEX TERMS Image fusion, hyperspectral image, panchromatic image, adaptive tensor, multi-scale Retinex algorithm.

show abstract

Hyperspectral and Multispectral Image Fusion Based on Local Low Rank and Coupled Spectral Unmixing

Cited by 57 publications

References 34 publications

Multisensor Analysis of Spectral Dimensionality and Soil Diversity in the Great Central Valley of California

Multisensor Analysis of Spectral Dimensionality and Soil Diversity in the Great Central Valley of California

Hyperspectral Image Super-Resolution with Self-Supervised Spectral-Spatial Residual Network

Hyperspectral and Panchromatic Image Fusion via Adaptive Tensor and Multi-Scale Retinex Algorithm

Contact Info

Product

Resources

About