Deconvolution for uncertain systems

The degradation of signals and images can be caused by both natural perturbations and electronic systems, recording linear systems, in which parameters are slowly time-varying such as sensors or other systems of storage. Treatment of the above mentioned systems are discussed. For this purpose, Sekko & al. developed a structure, which is improved later by Neveux, in order to produce an inverse computing filter with constant gain. The disadvantage of this approach is the resulting Kalman filter has to be used on line. In order to solve this problem, we propose a combination of the idea that has been proposed by Biemond, which gives the advantage to decorrelate lines (or columns) of the image, with the theory of the world in torus and the developed tools for uncertain systems. This work enables the realization of the deconvolution of 1D and 2D slowly time-varying systems.