Spectrum reconstruction with taking into account errors of initial data

The technique is proposed for processing of data from integral-type detectors (ionization and vacuum chambers or calorimeters) taking into account the errors of detected signals and detector sensitivity factor. The problem of spectrum reconstruction is successfully solved by the maximum likelihood method (MLM). The iterative process of the MLM solution search has shown that the accumulation of approximation errors is absent when computer numerical solution. The correctness of spectrum form description with the help of descrete representation depends on node points distribution. A condition of optimum position of nodes is obtained. The data processing technique for analogous detectors is tested by comparison of determined spectra, obtained with the help of a set of ionization chambers, with measurements by a scintillation counter. Maximum difference constituted 20%. Spectrum, reconstructed according to the proposed technique, is free from false peaks and oxillations, and its histogram is useful for practical application