Rail head surface defect magnetic flux leakage signal filtering based on the relativity among the adjacent sensors

Rail is the infrastructure of railway traffic and magnetic flux leakage (MFL) is commonly used to detect the surface defects of ferromagnetic materials such as rail. A correlation-based filtering method is proposed to inhibit the lift-off jamming in MFL detection for the defects in the rail head surface. The detection data is segmented and the two sensors with the minimum correlation coefficient of their outputs in the x direction are found. The possible defect width and depth according to the maximum of each output is estimated and the leakage magnetic field (LMF) of it is calculated. The output of the sensor with smaller correlation coefficient between the calculation LMF and the sensor output is chosen as the reference signal to filter. An experimental system was constructed to detect the artificial defects in the rail head surface, and the experimental results shown that the method reduces the lift-off jamming.