Abstract
A hyperspectral image (HSI) contains abundant spatial and spectral information, but it is always corrupted by various noises, especially Gaussian noise. Global correlation (GC) across spectral domain and nonlocal self-similarity (NSS) across spatial domain are two important characteristics for an HSI. To keep the integrity of the global structure and improve the details of the restored HSI, we propose a global and nonlocal weighted tensor norm minimum denoising method which jointly utilizes GC and NSS. The weighted multilinear rank is utilized to depict the GC information. To preserve structural information with NSS, a patch-group-based low-rank-tensor-approximation (LRTA) model is designed. The LRTA makes use of Tucker decompositions of 4D patches, which are composed of a similar 3D patch group of HSI. The alternating direction method of multipliers (ADMM) is adapted to solve the proposed models. Experimental results show that the proposed algorithm can preserve the structural information and outperforms several state-of-the-art denoising methods.
Highlights
A hyperspectral image (HSI) consists of hundreds of contiguous bands at specific wavelengths
The original HSI is treated as a 3-order tensor, so N=3, and through some experiments, we find that α1, α2 should be equal and they should be smaller than α3 when the denoising result is good
erreur relative globale adimensionnelle de synthese (ERGAS) measures fidelity of the restored image based on the weighted sum of mean squared error (MSE) in each band, and spectral angle mapper (SAM) calculates the average angle between spectrum vectors of the target HSI and the reference one across all spatial positions, so it fully reflects the fidelity of the spectral reflectance of the target HSI
Summary
A hyperspectral image (HSI) consists of hundreds of contiguous bands at specific wavelengths. Tensor-based approaches have achieved promising results in HSI denoising [25,26,27,28,29,30,31,32,33,34,35,36] as they can process the spatial and spectral information jointly. Liu et al [27] designed the parallel factor analysis (PARAFAC) method by utilizing the parallel factor analysis [28] to denoise HSI It regards the two spatial dimensions as two modes of tensor, this will lead to vertical and horizontal artifacts. Considering the nonlocal self-similarity across space in an MSI, a tensor dictionary learning-based (TDL) model [31] is proposed for denoising by enforcing hard constraints on the rank of the core tensor.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
