Multichannel image restoration based on a pseudo-block-diagonalized Wiener filter
Ki-Woon Na, Joonki Paik · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 1996
In the present paper we propose a multichannel image restoration method based on a pseudo- block-diagonalized Wiener filter. The proposed algorithm assumes that a multichannel imaging system includes both intra and inter-channel degradation, and utilizes correlation information among channels in the restoration process. Three types of the spectrum estimation techniques are also proposed for implementing the proposed restoration algorithm. The proposed restoration method can be implemented in the Fourier transform domain, and it neither assumes spectral or spatial separability used by Hunt, nor computes the inverse of a huge block matrix.