Image restoration by Gauss LU decomposition
Abstract
The application of a truncated Gauss LU decomposition technique to image restoration problems is demonstrated. The demonstration involves a computer simulation in which the point spread function is spatially separable. The Gauss LU decomposition method requires solution of 16 sets of equations, each set containing eight equations in eight unknowns; the technique is thus simpler than the eigenvalue and the singular value decomposition approaches to image restoration.
 Publication:

Applied Optics
 Pub Date:
 May 1979
 DOI:
 10.1364/AO.18.001684
 Bibcode:
 1979ApOpt..18.1684A
 Keywords:

 Image Enhancement;
 Imaging Techniques;
 Matrix Theory;
 Fredholm Equations;
 Matrices (Mathematics);
 Random Noise;
 Instrumentation and Photography;
 IMAGE RESTORATION;
 IMAGE PROCESSING