Filtration of an accompanying signal of a radiometric defectoscope using a digital computer

The study has been aimed at development of the method for discrimination of signals caused by defects from the mixture of signal-noise when testing thick-walled articles of a complex shape confined within the cylindrical, conical, and spherical surfaces. Algorithms are developed for filtration of accompanying signals on the base of selected characteristic features. Suggested are the criteria for evaluation of the various filtration algorithms. A set-up is made, which includes a recording channel of a radiometric flaw detector operating in the logarithmic mode and a puncher. Used as the articles under test have been their defective simulators. Compared are various algorithms and a procedure is suggested for processing the signals from the flaw detector, which insures discrimination of the signal induced by a defect when testing the articles with a complex shape