MMSE Restoration of Image Degraded by Noise and Separable Blur Based on Biorthogonal Multirestoration Analysis
碩士 === 國立臺灣科技大學 === 工程技術研究所 === 83 === In this thesis, we assume image degraded process to be due to the additive white noise and the space invariant separable blur. In this assumption, we propose a method which uses the minimum mean square...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
1994
|
Online Access: | http://ndltd.ncl.edu.tw/handle/78303388071596907721 |
Summary: | 碩士 === 國立臺灣科技大學 === 工程技術研究所 === 83 === In this thesis, we assume image degraded process to be due to
the additive white noise and the space invariant separable
blur. In this assumption, we propose a method which uses the
minimum mean square error (MMSE) criteria to restore the
degraded image. The basic ideal is to consider the point spread
function as a low pass filter of a biorthogonal
multirestoration analysis,then uses the minimum mean square
error criteria to restore the degraded image and gets a fast
nonrecursive algorithm. The algo- rithm can be implemented by
using the fast Fourier transform (FFT) computation and doesn't
require any statistical property of the original image.
|
---|