#### DMCA

## Bayesian Fusion of Multi-Band Images

Citations: | 2 - 1 self |

### Citations

2408 | Nonlinear dimensionality reduction by locally linear embedding
- Roweis, Saul
(Show Context)
Citation Context ...1) Subspace learning: Learning the matrix V in (5) is a preprocessing step, which can be solved by different strategies. A lot of DR methods might be exploited, such as locally linear embedding (LLE) =-=[42]-=-, independent component analysis (ICA) [43], hyperspectral signal subspace identification by minimum error (HySime) [32], minimum change rate deviation (MCRD) [44] and so on. In this work, we propose ... |

1489 | Monte Carlo Statistical Methods
- Robert, Casella
- 2004
(Show Context)
Citation Context ...rd model and the a priori modeling is defined in a high dimensional space, which makes difficult the use of any conventional MCMC algorithm, e.g., the Gibbs sampler or the Metropolis-Hastings sampler =-=[25]-=-. To overcome this difficulty, a particular MCMC scheme, called Hamiltonian Monte Carlo (HMC) algorithm, is derived [26], [27]. It differs from the standard Metropolis-Hastings algorithm by exploiting... |

732 | Probabilistic inference using Markov chain Monte Carlo methods
- Neal
- 1993
(Show Context)
Citation Context ...), which is considered to generate vectors u directly. More precisely, we consider the HMC algorithm initially proposed by Duane et al. for simulating the lattice field theory in [26]. As detailed in =-=[39]-=-, this technique allows mixing property of the sampler to be improved, especially in a highdimensional problem. It exploits the gradient of the distribution to be sampled by introducing auxiliary “mom... |

435 |
A universal image quality index
- Wang, Bovik
- 2002
(Show Context)
Citation Context ...ssed in radians and thus belongs to [−pi2 , pi2 ]. The smaller the absolute value of SAM, the less important the spectral distortion. c) UIQI: The universal image quality index (UIQI) was proposed in =-=[48]-=- for evaluating the similarity between two single band images. It is related to the correlation, luminance distortion and contrast distortion of the estimated image to the reference image. The UIQI be... |

218 | Joint MAP registration and highresolution image estimation using a sequence of undersampled images
- Hardie, Barnard, et al.
- 1997
(Show Context)
Citation Context ... Gaussian prior for the vectors ui is also motivated by the fact that this kind of prior has been used successfully in several works related to the fusion of multiple degraded images, including [20], =-=[34]-=-, [35]. Note finally that the Gaussian prior has the interest of being a conjugate distribution relative to the statistical model (4). As it will be shown in Section IV, coupling this Gaussian prior d... |

174 |
Hyperspectral imaging: Techniques for spectral detection and classification
- Chang
- 2003
(Show Context)
Citation Context ...ch consists of acquiring a same scene in several hundreds of contiguous spectral bands, has opened a new range of relevant applications, such as target detection, classification and spectral unmixing =-=[5]-=-. The visualization of HS images is also interesting to be explored [6]. Naturally, to take advantage of the newest benefits offered by HS images, the problem of fusing HS and PAN images has been expl... |

160 | The Bayesian Choice: from decision-theoretic foundations to computational implementation - Robert - 2001 |

143 |
A Bayesian approach to image expansion for improved definition,”
- Schultz, Stevenson
- 1994
(Show Context)
Citation Context ...o a single-band image (i.e., nλ,p = mλ = 1) with a decimation factor d in both spatial dimensions, it is easy to show that Fp is an nx,pny,p ×mxmy block diagonal matrix with mx = dnx,p and my = dny,p =-=[31]-=-. Another example of degradation frequently encountered in the signal and image processing literature is spatial blurring [19], where Fp (·) usually represents a 2-dimensional convolution by a kernel ... |

112 | Mcmc using hamiltonian dynamics.
- Neal
- 2011
(Show Context)
Citation Context ...CMC algorithm, e.g., the Gibbs sampler or the Metropolis-Hastings sampler [25]. To overcome this difficulty, a particular MCMC scheme, called Hamiltonian Monte Carlo (HMC) algorithm, is derived [26], =-=[27]-=-. It differs from the standard Metropolis-Hastings algorithm by exploiting Hamiltonian evolution dynamics to propose states with higher acceptance ratio, reducing the correlation between successive sa... |

81 | Hyperspectral subspace identification
- Bioucas-Dias, Nascimento
- 2008
(Show Context)
Citation Context ...location (with i = 1, · · · ,mxmy). Since adjacent HS bands are known to be highly correlated, the HS vector xi usually lives in a subspace whose dimension is much smaller than the number of bands mλ =-=[32]-=-, i.e., xi = V Tui (5) where ui is the projection of the vector xi onto the subspace spanned by the columns of VT ∈ Rmλ×m̃λ . Note that VT is possibly known a priori from the scene or can be learned f... |

76 |
Coupling and ergodicity of adaptive Markov chain Monte Carlo algorithms
- Roberts, Rosenthal
(Show Context)
Citation Context ...r experiment, the counting window at time t contains the vectors x̃(t−NW+1), x̃(t−NW), · · · , x̃(t) with NW = 50) and Na,t is the number of accepted samples in this window at time t. As explained in =-=[45]-=-, the adaptive tuning should adapt less and less as the algorithm proceeds to guarantee that the generated samples form a stationary Markov chain. In the proposed implementation, the parameter ε is ad... |

75 | Bayesian curve fitting using MCMC with applications to signal segmentation
- Punskaya, Andrieu, et al.
- 2002
(Show Context)
Citation Context ...ts two parameters. For simplicity, we propose to fix the hyperparameter ν whereas the hyperparameter γ will be estimated from the data. This strategy is very classical for scale parameters (e.g., see =-=[36]-=-). Note that the inverse-gamma distribution (7) is conjugate for the statistical model (4), which will allow closed-form expressions to be obtained for the conditional distributions f ( s2p,i|z ) of t... |

68 |
A wavelet transform method to merge landsat tm and spot panchromatic data,
- Zhou, Civco, et al.
- 1998
(Show Context)
Citation Context ...ctivities have already been conducted for this practical multi-band fusion problem [15]. Noticeably, a lot of pansharpening methods, such as component substitution [2], relative spectral contribution =-=[16]-=- and high-frequency injection [17] are inapplicable or inefficient for the HS+MS fusion problem. To address the challenge raised by the high dimensionality of the data to be fused, innovative methods ... |

66 | Joint Bayesian endmember extraction and linear unmixing for hyperspectral imagery
- Dobigeon, Moussaoui, et al.
(Show Context)
Citation Context ...ered [22]. Moreover, to define the prior distribution assigned to this image, we resort to geometrical considerations well admitted in the HS imaging literature devoted to the linear unmixing problem =-=[23]-=-. In particular, the high spatial resolution HS image to be estimated is assumed to live in a lower dimensional subspace, which is a suitable hypothesis when the observed scene is composed of a finite... |

43 | Introduction of sensor spectral response into image fusion methods. Application to wavelet-based methods,”
- Otazu, Gonzalez-Audicana, et al.
- 2005
(Show Context)
Citation Context ...ent images. More precisely, the sensor specifications (i.e., spectral or spatial responses) are exploited to properly design the spatial or spectral degradations suffered by the image to be recovered =-=[22]-=-. Moreover, to define the prior distribution assigned to this image, we resort to geometrical considerations well admitted in the HS imaging literature devoted to the linear unmixing problem [23]. In ... |

42 | SVM- and MRF-based method for accurate classification of hyperspectral images
- Tarabalka, Fauvel, et al.
- 2010
(Show Context)
Citation Context ...t have been reduced to 93 bands after removing the water vapor absorption bands (with spectral range from 0.43 to 0.86 µm). This image has received a lot of attention in the remote sensing literature =-=[50]-=-. The HS blurring kernel is the same as in paragraph V-A whereas the PAN image was obtained by averaging all the high resolution HS bands. The SNR of the PAN image is 30dB. Apart from [18], [19], we a... |

41 | Fusion of multispectral and panchromatic images using improved IHS and PCA mergers based on wavelet decomposition,”
- Gonzalez-Audıcana, Saleta, et al.
- 2004
(Show Context)
Citation Context ...ted for this practical multi-band fusion problem [15]. Noticeably, a lot of pansharpening methods, such as component substitution [2], relative spectral contribution [16] and high-frequency injection =-=[17]-=- are inapplicable or inefficient for the HS+MS fusion problem. To address the challenge raised by the high dimensionality of the data to be fused, innovative methods need to be developed, which is the... |

39 | Independent component analysis-based dimensionality reduction with applications in hyperspectral image analysis.
- Wang, Chang
- 2006
(Show Context)
Citation Context ... in (5) is a preprocessing step, which can be solved by different strategies. A lot of DR methods might be exploited, such as locally linear embedding (LLE) [42], independent component analysis (ICA) =-=[43]-=-, hyperspectral signal subspace identification by minimum error (HySime) [32], minimum change rate deviation (MCRD) [44] and so on. In this work, we propose to use the principal component analysis (PC... |

35 | Stochastic methods for joint registration, restoration, and interpolation of multiple undersampled images.
- Woods, Galatsanos, et al.
- 2006
(Show Context)
Citation Context ...ian prior for the vectors ui is also motivated by the fact that this kind of prior has been used successfully in several works related to the fusion of multiple degraded images, including [20], [34], =-=[35]-=-. Note finally that the Gaussian prior has the interest of being a conjugate distribution relative to the statistical model (4). As it will be shown in Section IV, coupling this Gaussian prior distrib... |

29 | Introduction to Remote - Campbell - 1996 |

24 |
A survey of classical methods and new trends in pansharpening of multispectral images,”
- Amro, Mateos, et al.
- 2011
(Show Context)
Citation Context ... high spatial and low spectral res-olution image with an auxiliary image of higher spectral but lower spatial resolution, also known as multi-resolution image fusion, has been explored for many years =-=[2]-=-. When considering remotely sensed images, an archetypal fusion task is the pansharpening, which generally consists of fusing a high spatial resolution panchromatic (PAN) image and low spatial resolut... |

19 | MAP estimation for hyperspectral image resolution enhancement using an auxiliary sensor,
- Hardie, Eismann, et al.
- 2004
(Show Context)
Citation Context ...ion problem. To address the challenge raised by the high dimensionality of the data to be fused, innovative methods need to be developed, which is the main objective of this paper. As demonstrated in =-=[18]-=-, [19], the fusion of HS and MS images can be conveniently formulated within a Bayesian inference framework. Bayesian fusion allows an intuitive interpretation of the fusion process via the posterior ... |

16 | Application of the stochastic mixing model to hyperspectral resolution enhancement - Eismann, Hardie - 2004 |

14 | Quality of high resolution synthesised images: Is there a simple criterion?” in
- Wald
- 2000
(Show Context)
Citation Context ... image, the UIQI is obtained band-by-band and averaged over all bands. d) ERGAS: The relative dimensionless global error in synthesis (ERGAS) calculates the amount of spectral distortion in the image =-=[49]-=-. This measure of fusion quality is defined as ERGAS = 100× 1d2 √ 1 mλ ∑mλ i=1 ( RMSE(i) µi ) , where 1/d2 is the ratio between the pixel sizes of the MS and HS images, µi is the mean of the ith band ... |

13 |
Coupled nonnegative matrix factorization unmixing for hyperspectral and multispectral data fusion,”
- Yokoya, Yairi, et al.
- 2012
(Show Context)
Citation Context ... below. These measures have been widely used in the HS image processing community and are appropriate for evaluating the quality of the fusion in terms of spectral and spatial resolutions [18], [46], =-=[47]-=-. a) RSNR: The reconstruction SNR (RSNR) is related to the difference between the actual and fused images RSNR(x, x̂) = 10 log10 ( ‖x‖2 ‖x−x̂‖22 ) . The larger RSNR, the better the fusion quality and ... |

12 |
Noise-resistant wavelet-based Bayesian fusion of multispectral and hyperspectral images,”
- Zhang, Backer, et al.
- 2009
(Show Context)
Citation Context ...is an nx,pny,p ×mxmy block diagonal matrix with mx = dnx,p and my = dny,p [31]. Another example of degradation frequently encountered in the signal and image processing literature is spatial blurring =-=[19]-=-, where Fp (·) usually represents a 2-dimensional convolution by a kernel κp. Similarly, when applied to a single-band image, Fp is an nxny×nxny Toeplitz matrix. The problem addressed in this paper co... |

11 |
Merging hyperspectral and panchromatic image data: qualitative and quantitative analysis,”
- Cetin, Musaoglu
- 2009
(Show Context)
Citation Context ... visualization of HS images is also interesting to be explored [6]. Naturally, to take advantage of the newest benefits offered by HS images, the problem of fusing HS and PAN images has been explored =-=[7]-=-–[9]. Capitalizing on decades of experience in MS pansharpening, most of the HS pansharpening approaches merely adapt existing algorithms for PAN and MS fusion [10], [11]. Other methods are specifical... |

9 |
MAP estimation for multiresolution fusion in remotely sensed images using an IGMRF prior model,”
- Joshi, Jalobeanu
- 2010
(Show Context)
Citation Context ...the model via the prior distribution assigned to the scene to be estimated. Many strategies related to HS resolution enhancement have been proposed to define this prior distribution. For instance, in =-=[3]-=-, the highly resolved image to be estimated is a priori modeled by an in-homogeneous Gaussian Markov 1http://www.satimagingcorp.com/satellite-sensors/WorldView3-DS-WV3Web.pdf 2 IEEE JOURNAL OF SELECTE... |

8 | Fusion of hyperspectral and panchromatic images using multiresolution analysis and nonlinear pca band reduction,”
- Licciardi, Khan, et al.
- 2012
(Show Context)
Citation Context ...rpening, most of the HS pansharpening approaches merely adapt existing algorithms for PAN and MS fusion [10], [11]. Other methods are specifically designed to the HS pansharpening problem (see, e.g., =-=[8]-=-, [12], [13]). Conversely, the fusion of MS and HS Copyright (c) 2014 IEEE. Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained... |

8 |
Resolution enhancement of hyperspectral data,” in
- Winter, Winter
- 2002
(Show Context)
Citation Context ...ng, most of the HS pansharpening approaches merely adapt existing algorithms for PAN and MS fusion [10], [11]. Other methods are specifically designed to the HS pansharpening problem (see, e.g., [8], =-=[12]-=-, [13]). Conversely, the fusion of MS and HS Copyright (c) 2014 IEEE. Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from ... |

8 | A model-based approach to multiresolution fusion in remotely sensed images,”
- Joshi, Bruzzone, et al.
- 2006
(Show Context)
Citation Context ...he computation of the MMSE estimator (especially in high-dimension data space), most of the Bayesian estimators have proposed to solve the HS and MS fusion problem using a MAP formulation [18], [19], =-=[24]-=-. However, optimization algorithms designed to maximize the posterior distribution may suffer from the presence of local extrema, that prevents any guarantee to converge towards the actual maximum of ... |

8 | An adaptive IHS pan-sharpening method,”
- Rahmani, Strait, et al.
- 2010
(Show Context)
Citation Context ...paragraph V-A whereas the PAN image was obtained by averaging all the high resolution HS bands. The SNR of the PAN image is 30dB. Apart from [18], [19], we also compare the results with the method of =-=[51]-=-, which proposes a popular pansharpening method. The results are displayed in Fig. 6 and the quantitative results are reported in Table III. The proposed Bayesian method still provides interesting res... |

7 |
Support-based implementation of Bayesian data fusion for spatial enhancement: Application to ASTER thermal images,
- Fasbender, Tuia, et al.
- 2008
(Show Context)
Citation Context ...odel Zp = Fp (X) + Ep. (1) In (1), Fp (·) is a linear or nonlinear transformation that models the degradation operated on X. As previously assumed in numerous works (see for instance [3], [19], [24], =-=[28]-=-, [29] among some recent contributions), these degradations may include spatial blurring, spatial decimation and spectral mixing which can all be modeled by linear transformations. In what follows, th... |

7 |
Generative Bayesian image super resolution with natural image prior,”
- Zhang, Zhang, et al.
- 2012
(Show Context)
Citation Context ...be detailed in Section V. To sample according to a high-dimension Gaussian distribution such as f ( u|Σu, s2, z ) , one might think of using other simulation techniques such as the method proposed in =-=[40]-=- to solve super resolution problems. Similarly, Orieux et al. have proposed a perturbation approach to sample high-dimensional Gaussian distributions for general linear inverse problems [41]. However,... |

6 |
Estimation of covariance matrices based on hierarchical inverse-Wishart priors,” J. of Stat. Planning and Inference,
- Bouriga, Feron
- 2012
(Show Context)
Citation Context ...rparameter Σu: Assigning a conjugate a priori inverse-Wishart distribution to the covariance matrix of a Gaussian vector has provided interesting results in the signal and image processing literature =-=[38]-=-. Following these works, we have chosen the following prior for Σu Σu ∼ W−1(Ψ, η) (9) whose density is f(Σu|Ψ, η) = |Ψ| η 2 2 ηm̃λ 2 Γm̃λ( η 2 ) |Σu|− η+m̃λ+1 2 e− 1 2 tr(ΨΣ −1 u ). Again, the hyper-h... |

6 |
A Bayesian restoration approach for hyperspectral images,”
- Zhang, Duijster, et al.
- 2012
(Show Context)
Citation Context ...efined below. These measures have been widely used in the HS image processing community and are appropriate for evaluating the quality of the fusion in terms of spectral and spatial resolutions [18], =-=[46]-=-, [47]. a) RSNR: The reconstruction SNR (RSNR) is related to the difference between the actual and fused images RSNR(x, x̂) = 10 log10 ( ‖x‖2 ‖x−x̂‖22 ) . The larger RSNR, the better the fusion qualit... |

5 | Pan-sharpening with a Bayesian nonparametric dictionary learning model,” in
- Ding, Jiang, et al.
- 2014
(Show Context)
Citation Context ...l resolution panchromatic (PAN) image and low spatial resolution multispectral (MS) image. Pansharpening has been addressed in the literature for several decades and still remains an active topic [2]–=-=[4]-=-. More recently, hyperspectral (HS) imaging, which consists of acquiring a same scene in several hundreds of contiguous spectral bands, has opened a new range of relevant applications, such as target ... |

5 |
A variational approach to hyperspectral image fusion,” in
- Moeller, Wittman, et al.
- 2009
(Show Context)
Citation Context ... HS and PAN images has been explored [7]–[9]. Capitalizing on decades of experience in MS pansharpening, most of the HS pansharpening approaches merely adapt existing algorithms for PAN and MS fusion =-=[10]-=-, [11]. Other methods are specifically designed to the HS pansharpening problem (see, e.g., [8], [12], [13]). Conversely, the fusion of MS and HS Copyright (c) 2014 IEEE. Personal use of this material... |

5 |
Super-resolution of hyperspectral imagery using complex ridgelet transform,”
- Chen, Qian, et al.
- 2012
(Show Context)
Citation Context ...st of the HS pansharpening approaches merely adapt existing algorithms for PAN and MS fusion [10], [11]. Other methods are specifically designed to the HS pansharpening problem (see, e.g., [8], [12], =-=[13]-=-). Conversely, the fusion of MS and HS Copyright (c) 2014 IEEE. Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from the IE... |

4 | Bayesian fusion of hyperspectral and multispectral images,” in
- Wei, Dobigeon, et al.
- 2014
(Show Context)
Citation Context ...ct n◦ANR-12-BS03-003 and by ANR-11-LABX-0040-CIMI within the program ANR-11-IDEX-0002-02 within the thematic trimester on image processing. Part of this work was presented during the IEEE ICASSP 2014 =-=[1]-=-. Qi Wei, Nicolas Dobigeon and Jean-Yves Tourneret are with University of Toulouse, IRIT/INP-ENSEEIHT, 2 rue Camichel, BP 7122, 31071 Toulouse cedex 7, France (e-mail: {qi.wei, nicolas.dobigeon, jeany... |

4 |
Superresolution construction of multispectral imagery based on local enhancement,”
- Elbakary, Alam
- 2008
(Show Context)
Citation Context ...p = Fp (X) + Ep. (1) In (1), Fp (·) is a linear or nonlinear transformation that models the degradation operated on X. As previously assumed in numerous works (see for instance [3], [19], [24], [28], =-=[29]-=- among some recent contributions), these degradations may include spatial blurring, spatial decimation and spectral mixing which can all be modeled by linear transformations. In what follows, the remo... |

4 | Sampling high-dimensional Gaussian distributions for general linear inverse problems,”
- Orieux, Feron, et al.
- 2012
(Show Context)
Citation Context ...oposed in [40] to solve super resolution problems. Similarly, Orieux et al. have proposed a perturbation approach to sample high-dimensional Gaussian distributions for general linear inverse problems =-=[41]-=-. However, these techniques rely on additional optimization schemes included within the Monte Carlo algorithm, which implies that the generated samples are only approximately distributed according to ... |

3 |
Fusion of hyperspectral and multispectral images: A novel framework based on generalization of pan-sharpening methods,”
- Chen, Pu, et al.
- 2014
(Show Context)
Citation Context ...d PAN images has been explored [7]–[9]. Capitalizing on decades of experience in MS pansharpening, most of the HS pansharpening approaches merely adapt existing algorithms for PAN and MS fusion [10], =-=[11]-=-. Other methods are specifically designed to the HS pansharpening problem (see, e.g., [8], [12], [13]). Conversely, the fusion of MS and HS Copyright (c) 2014 IEEE. Personal use of this material is pe... |

3 |
Hyperspectral and multispectral data fusion mission on hyperspectral imager suite (HISUI),” in
- Yokoya, Iwasaki
- 2013
(Show Context)
Citation Context ...d by the H-IIA rocket in 2015 or later as one of mission instruments onboard JAXA’s ALOS-3 satellite. Some research activities have already been conducted for this practical multi-band fusion problem =-=[15]-=-. Noticeably, a lot of pansharpening methods, such as component substitution [2], relative spectral contribution [16] and high-frequency injection [17] are inapplicable or inefficient for the HS+MS fu... |

3 |
Dimension reduction of optical remote sensing images via minimum change rate deviation method,”
- Dianat, Kasaei
- 2010
(Show Context)
Citation Context ...ch as locally linear embedding (LLE) [42], independent component analysis (ICA) [43], hyperspectral signal subspace identification by minimum error (HySime) [32], minimum change rate deviation (MCRD) =-=[44]-=- and so on. In this work, we propose to use the principal component analysis (PCA), which is a classical DR technique used in HS imagery. It maps the original data into a lower dimensional subspace wh... |

3 | Hyperspectral and multispectral image fusion based on a sparse representation,”
- Wei, Dias, et al.
- 2015
(Show Context)
Citation Context ...g constraints for a possible improved spectral accuracy and the generalization to nonlinear degradations would also deserve some attention. Finally, a comparison with very recent fusion methods [47], =-=[52]-=- would be clearly interesting. ACKNOWLEDGMENTS The authors thank Dr. Paul Scheunders and Dr. Yifan Zhang for sharing the codes of [19] and Jordi Inglada, from Centre National d’Études Spatiales (CNES... |

2 |
A bayesian approach to visualization-oriented hyperspectral image fusion,”
- Kotwal, Chaudhuri
- 2013
(Show Context)
Citation Context ... spectral bands, has opened a new range of relevant applications, such as target detection, classification and spectral unmixing [5]. The visualization of HS images is also interesting to be explored =-=[6]-=-. Naturally, to take advantage of the newest benefits offered by HS images, the problem of fusing HS and PAN images has been explored [7]–[9]. Capitalizing on decades of experience in MS pansharpening... |

2 |
Japanese hyper-multi spectral mission,” in
- Ohgi, Iwasaki, et al.
- 2010
(Show Context)
Citation Context ...range [450 ∼ 800]nm. Another interesting example is the HS+MS suite (called hyperspectral imager suite (HISUI)) that has been developed by the Japanese ministry of economy, trade, and industry (METI) =-=[14]-=-. HISUI is the Japanese nextgeneration Earth-observing sensor composed of HS and MS imagers and will be launched by the H-IIA rocket in 2015 or later as one of mission instruments onboard JAXA’s ALOS-... |

2 |
resolution enhancement using high-resolution multispectral imagery with arbitrary response functions
- “Hyperspectral
- 2005
(Show Context)
Citation Context ...OCESSING, VOL. ??, NO. ??, ?? random field (IGMRF). The parameters of this IGMRF are empirically estimated from a panchromatic image in the first step of the analysis. In [18] and related works [20], =-=[21]-=-, a multivariate Gaussian distribution is proposed as prior distribution for the unobserved scene. The resulting conditional mean and covariance matrix can then be inferred using a standard clustering... |

1 |
Introducing hyperspectral pansharpening
- Loncan, Almeida, et al.
(Show Context)
Citation Context ...ualization of HS images is also interesting to be explored [6]. Naturally, to take advantage of the newest benefits offered by HS images, the problem of fusing HS and PAN images has been explored [7]–=-=[9]-=-. Capitalizing on decades of experience in MS pansharpening, most of the HS pansharpening approaches merely adapt existing algorithms for PAN and MS fusion [10], [11]. Other methods are specifically d... |

1 |
Bayesian fusion of multiband images – Complementary results and supporting materials
- Wei, Dobigeon, et al.
- 2014
(Show Context)
Citation Context ... related HS analysis 4 IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING, VOL. ??, NO. ??, ?? references, e.g., [23], [32]. More experimental justifications for the necessity of DR can be found in =-=[33]-=-. Using the notation u = [ uT1 ,u T 2 , · · · ,uTmxmy ]T , we have u = Vx, where V is an M̃×M block-diagonal matrix whose blocks are equal to V and M̃ = mxmym̃λ. Instead of assigning a prior distribut... |