PSF fitting (for point-spread function fitting) is a type the analysis of overlapping Airy disks to determine the likely scenario that produced them. Through such analysis, stars and other unresolved astronomical objects can be identified even when the Rayleigh criterion for the telescope's angular resolution suggests they are too close together. It constitutes a type of speckle suppression, the suppression of noise due to characteristics of the optical instrument (i.e., telescope). It is used for resolving stars that are close together in the celestial sphere, or similarly close to other objects such as galaxies. It is also used in direct imaging of extra-solar planets.
The technique is typically carried out by software processing of the observation data using it can require considerable computation. A general method is to produce "candidate" fully-resolved images by some means, and for each of these, calculate the observation data this telescope would make of it (given its limits), then using a chi-squared test or similar statistical test to compare the calculated result with the actual observation data, then choose the best match. Various means are used to optimize such "guess and test" methods.