Abstract
Due to the instrumental and imaging optics limitations, it is difficult to acquire high spatial resolution hyperspectral imagery (HSI). Super-resolution (SR) imagery aims at inferring high quality images of a given scene from degraded versions of the same scene. This paper proposes a novel hyperspectral imagery super-resolution (HSI-SR) method via dictionary learning and spatial-spectral regularization. The main contributions of this paper are twofold. First, inspired by the compressive sensing (CS) framework, for learning the high resolution dictionary, we encourage stronger sparsity on image patches and promote smaller coherence between the learned dictionary and sensing matrix. Thus, a sparsity and incoherence restricted dictionary learning method is proposed to achieve higher efficiency sparse representation. Second, a variational regularization model combing a spatial sparsity regularization term and a new local spectral similarity preserving term is proposed to integrate the spectral and spatial-contextual information of the HSI. Experimental results show that the proposed method can effectively recover spatial information and better preserve spectral information. The high spatial resolution HSI reconstructed by the proposed method outperforms reconstructed results by other well-known methods in terms of both objective measurements and visual evaluation.
Highlights
Hyperspectral imagery (HSI) has high spectral resolution, which is very important for many applications, such as land use analysis, environment studies, military surveillance, and so on
In [21,22], the authors only establish one HR dictionary, which needed a few of atoms to resolve the SR problem well, but they fix the linear measurement matrix which is limited only to a scale factor of two. If these methods were directly applied to improve the spatial resolution of every spectral band of HSI individually without considering spectral information, they will destroy the spectral information of the HSI, which is extremely important for the applications of the HSI
Based on the LR HSI generation model, we referred to the hyperspectral imagery super-resolution (HSI-SR) problem as an ill-posed inverse problem
Summary
Hyperspectral imagery (HSI) has high spectral resolution (containing about 200 spectral band in the visible and infrared wavelength regions, i.e., 400–2500 nm), which is very important for many applications, such as land use analysis, environment studies, military surveillance, and so on. The typical methods are the intensity-hue-saturation (IHS) method [10], the principal component analysis (PCA) method [11,12], the wavelet transform (WT) method [13,14], and a variational model for P + XS image fusion [15] These pan-sharpening techniques perform a tradeoff between the spatial resolution and spectral resolution of the HSI. In [21,22], the authors only establish one HR dictionary, which needed a few of atoms to resolve the SR problem well, but they fix the linear measurement matrix which is limited only to a scale factor of two If these methods were directly applied to improve the spatial resolution of every spectral band of HSI individually without considering spectral information, they will destroy the spectral information of the HSI, which is extremely important for the applications of the HSI.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
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 call
