Improvement of rank one phase estimation (ROPE) autofocusing technique

This paper presents an improved rank one phase error estimate (ROPE) autofocus technique. Compared with ROPE, the improved algorithm (IROPE) has two distinct advantages. Firstly, its model is more practical, so its property of robustness is better. Secondly, at the beginning of its iteration process, the initialized values are set up approximately. The approximation is more accurate than that of the blind initialization of ROPE. The IROPE estimator of phase error is the maximum-likelihood (ML) estimator, and, under some special conditions, its estimator is the same as that of phase gradient autofocus (PGA), which is a linear unbiased minimum variation (LUMV) estimator. Both ROPE and PGA are subsets of IROPE.