Systematic approach to two-dimensional blind deconvolution by zero-sheet separation

Inspired by the work of Lane and Bates on automatic multidimensional deconvolution [ J. Opt. Soc. Am. A4, 180 ( 1987)], we have developed a systematic approach and an operational code for performing the deconvolution of multiply-convolved two-dimensional complex data sets in the absence of noise. We explain, in some detail, the major algorithmic steps, where noise or numerical errors can cause problems, our approach in dealing with numerical rounding errors, and where special noise-mitigating techniques can be used toward making blind de-convolution practical. Several examples of deconvolved imagery are presented, and future research directions are noted.