An eigenvector based method for point source localization in blurred images

We address the problem of resolving and localizing blurred point sources in intensity images. A new approach to image restoration is introduced which is a 2-D generalization of techniques originating from the field of direction of arrival estimation (DOA). It is shown that in the frequency domain, blurred point source images can be modeled with a structure analogous to the response of linear sensor arrays to coherent sources. Thus the problem may be cast into the form of DOA estimation, and modern eigenvector based subspace decomposition algorithms, such as MUSIC, may be adapted to search for these point sources. A generalization of array smoothing based on a regularization operator is introduced for 2-D arrays in order to achieve rank enhancement in the signal space of the covariance matrix