Two-Dimensional Adaptive Filtering using the Kalman Algorithm
Mario Gonsalves, Rayavarapu Lakshmi Narasimha Dinesh, K.R. Santha · 2006
Kalman Filters have been used in a wide range of one-dimensional signal processing applications. This paper deals with the application of the Kalman adaptive algorithm to the field of two-dimensional (2-D) signal processing. The results obtained on applying the aforesaid algorithm for the enhancement of an image distorted by noise are discussed. The Kalman algorithm is used to first estimate the coefficients of the unknown 2-D 3x3 tap FIR channel across which the image is assumed to be transmitted, and then to estimate the image itself. The results presented show that there is an improvement of more than 8 dB in the signal to noise ratio (SNR) of the image, measured before and after filtering. Further, the mean square error (MSE) and minimum mean square error (MMSE) plots show that the error in the estimation process converges quickly to a very low value