Hyperspectral image (HSI) restoration is a technique to inverse the information degradation process that occurs on a hyperspectral imaging system, i.e., spectrometer. Spectrometers can be classified as two types: plane-scanning and line-scanning spectrometers. It is necessary for a restoration algorithm to match the corresponding degradation process. However, most current restoration algorithms are only suitable to the former one. To solve such a mismatch of restoration algorithms to the imaging process in this paper, a new framework of HSI restoration is proposed. Compared to the existing frameworks, the proposed one is more applicable to a line-scanning spectrometer. Moreover, to solve the ill-posedness of such a framework, an anisotropy regularization term combining a vertical total variation and a linear spectral mixture is designed. Experimental results based on two simulation datasets, Pavia and San Diego, proved the effectiveness of the proposed framework and regularization term.
Due to the richness on high frequency components, hyperspectral image (HSI) is more sensitive to distortion like aliasing. Many methods aiming at removing such distortion have been proposed. However, seldom of them are suitable to HSI, due to low spatial resolution characteristic of HSI. Fortunately, HSI contains plentiful spectral information, which can be exploited to overcome such difficulties. Motivated by this, we proposed an aliasing removing method for HSI. The major differences between proposed and current methods is that proposed algorithm is able to utilize fractal structure information, thus the dilemma originated from low-resolution of HSI is solved. Experiments on real HSI data demonstrated subjectively and objectively that proposed method can not only remove annoying visual effect brought by aliasing, but also recover more high frequency component.
Mixed pixels are inevitable due to low-spatial resolutions of hyperspectral image (HSI). Linear spectrum mixture
model (LSMM) is a classical mathematical model to relate the spectrum of mixing substance to corresponding
individual components. The solving of LSMM, namely unmixing, is essentially a linear optimization problem
with constraints, which is usually consisting of iterations implemented on decent direction and stopping criterion
to terminate algorithms. Such criterion must be properly set in order to balance the accuracy and speed of
solution. However, the criterion in existing algorithm is too strict, which maybe lead to convergence rate
reducing. In this paper, by broaden constraints in unmixing, a new stopping rule is proposed, which can reduce
rate of convergence. The experiments results prove both in runtime and iteration numbers that our method can
accelerate convergence processing with only cost of little quality decrease in resulting.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.