Abstract
The paper proposes the new criterion of spatial resolution of an imaging system. This criterion considers shape and dimensions of a central peak and side lobes of a point spread function, standard deviation of noise. As a result, it helps to reach the optimal balance between the characteristics of a central peak, side lobes and noise. It differs from the widely known full width at half maximum and Sparrow criterion that mainly consider only characteristics of the central peak. There are discussed the digital filter optimal according to the proposed criterion and the limitations in maximization of spatial resolution of imaging systems.