Two-dimensional adaptive block Kalman filtering of radar imagery

Abstract

A 2-D adaptive block Kalman filtering scheme is introduced for processing synthetic aperture radar (SAR) imagery. The image process is represented by an autoregressive (AR) model with a nonsymmetric-half-plane (NSHP) region of support arranged in the block diagonal form. 2-D Kalman filtering equations are derived which take into account not only the effects of the multiplicative speckle noise but also the additive receiver thermal noise. This method assumes local stationarity within one block, whereas the image can be assumed to be globally nonstationary. A recursive identification process using the stochastic Newton approach is also presented which can be used online to estimate the filter parameters based upon the information within each new block of data received. An image processing example is provided to examine the effectiveness of the proposed method when used to remove the effects of speckle noise in an SAR image. >