Iterative realization of the 2-D Capon method applied in SAR image processing

The two-dimensional (2-D) Capon method is commonly used in the post-processing of the synthetic aperture radar images. In applications, the covariance matrix of the observation is estimated by vectorizing the 2-D data along one direction and then calculating the matrix in the same way as the 1-D case. This operation may arouse the problem of memory overload and estimation inaccuracy. By enduing the Capon filter with the assumption of separability, this paper decomposes the 2-D problem into two one-dimensional (1-D) ones and solve them iteratively to get the final estimation. The convergence of the iteration is demonstrated by the Monte Carlo simulation. Moreover, the comparison with the traditional method is carried out, with simulated images.