On finding a threshold of useful signals in the analysis of magnetic and eddy current defectograms

To ensure traffic safety of railway transport, non-destructive testing of rails is regularly carried out by using various approaches and methods, including magnetic and eddy current flaw detection methods. The paper is devoted to the problem of automatic determination of a threshold level of amplitudes of useful signals (from defects and structural elements of a railway track) during the \linebreak analysis of defectograms (records) of magnetic and eddy current flaw detectors. A signal is considered useful (and is subject to further analysis) if a deviation of its value from an average of all signals is at least twice the threshold noise level of rails. The probability of obtaining a signal from a section without structural elements (a rail noise signal) is characterized by the normal distribution law. Thus, the rule of three sigma can be used to calculate the threshold noise level. And a signal is useful if its amplitude deviation from a sample mean exceeds twice the threshold noise level. The paper proposes an algorithm for finding the threshold level of a rail noise and gives its theoretical justification, and it also examines examples of its operation on several fragments of real magnetic and eddy current defectograms.