A method for reconstructing two-dimensional distributions from a limit
ed number of integrals is presented. The method can be applied for arb
itrary expansions of the distribution; thus different types of distrib
utions can be processed with different expansions. The method is numer
ically simple, and is especially useful for reconstructing time evolut
ion of two-dimensional images. A comparison with maximum entropy tomog
raphy and an example of reconstructing tokamak radiation profiles in t
he case of impurity injection are presented.