Detection of weak activities by measurement of gamma-ray spectra

The presence of an activity in a measured sample may be verified by observing a relevant full-energy peak in a measured gamma-ray spectrum. Three decision procedures testing the presence of a full-energy peak are compared from the point of view of the test power as well as the sensitivity to background non-linearity. The better performance is achieved with the procedures which take into account an a priori information on the peak shape, rather than with a shape-independent one. The precedure which makes use of a generalized second difference filter is the least sensitive to the background non-linearity and should be preferred in routine use.