Abstract

Gamma-ray and X-ray imaging systems are required in many fields, including astronomy, industrial radiography, nuclear medicine and reactor safety research. Because hard X rays and gamma rays cannot be efficiently reflected or refracted, rather crude imaging systems are required, often with the result that very few photons are detected. Quantum noise is thus almost always the limiting factor in system performance. Nevertheless, there is considerable debate in the literature about how to specify and quantify these quantum limits. Widely differing conclusions are reached by different authors on questions such as: When is a pinhole preferable to a coded aperture in imaging a particular object? It is the goal of this paper to show that these discrepancies disappear when one carefully specifies the nature of the imaging system, the object or class of objects to be studied, the nature of any a priori information about the object, and the purpose of the imaging procedure. When the problem to be solved is thus carefully stated, general procedures can be given for quantifying the quantum limits to system performance.

© 1986 Optical Society of America