Fast iterative and non-iterative reconstruction of band-limited functions from irregular sampling values

Three approaches to a fast solution of the irregular sampling problem are described. The first is a direct method making use of pseudo-inverse matrices. The second is useful for cases where the number of sampling points is small compared to the total signal length; it is based on matrix iteration instead of iterated low pass filtering. The third approach can be applied in situations were the sampling geometry and the spectrum are given before the sampling data are available. The appropriate choice of a method in a given situation depends very much on the spectral size, the signal length, and the sampling density