Space-variant and color image restoration using Kalman filtering

A two-dimensional space-variant Kalman filtering approach for the restoration of images degraded by linear, space-variant blurs, and an efficient two-dimensional multichannel Kalman filtering approach for the restoration of color images on the basis of a coupled AR model used to represent both the intrachannel and interchannel correlations in multispectral image data are presented. In both cases, the authors assume that the degradation function is completely known. It is also possible to combine the proposed algorithms into one procedure for the restoration of color images degraded by space-variant blurs. The results are compared with those obtained using other methods. Experimental results are included.