Restoration of blurred and noisy images is a rich source of research material for years to come. The reason behind this is that; sources of degradation are too many and each of them behaves differently. Known, restoration techniques, in general, requires some a priori knowledge about the causing degradation's effects, this knowledge contributes in placing constraints on the solution of the restoration algorithm. One of the most successfully implemented methods is that known as Kalman filtering. The strategy of this filtering method is to obtain the optimum estimates of the unaffected image signal, using correlation coefficients that were estimated from the image correlation function. In this work, new estimation methods are suggested to recover the original unaffected image signal, using only one-pass filtering process. These introduced methods involved many useful procedures to enhance the final restored results; i.e. covariance matrix, correlation coefficients, driven matrix, gain matrix, image rotation. The estimation of the correlation matrix for a random field is, for example, obtained from the inverse Fourier transform of the power spectrum, STD of an additive noise is computed from the image itself (considering a smoothed image region), the driven matrix is calculated from the correlation coefficients. Moreover, rotational process is applied to perform image restoration in both row and column directions of an image. The introduced filter is applied on different types of degradation effects such as blurring by a Gaussian function in the presence and absence of the noise, and also, motion effects is considered. However, certain problems has been associated with the implementation of this filtering method; i.e. it involves a huge number of ':,' multiplication operations and an inversion of a large size matrices, which in turns, requires a large amount of computer memory and along processing time This problem is overcome by segmenting the processed image into several smaller blocks.Indeed, the image segmentation create artifacts which are manipulated by utilizing the quasi-toplitz-gain matrix. Finally, the obtained results, by our adaptive restoration method, have been compared with those obtained from the inverse and Wienter filters, using the peak signal to noise ratio PSNR quantitative test measure.