Subsurface defects recognition using thermography based on second order differential peak method of logarithmic coordinates

It is difficult to detect subsurface defects on the specimen using thermography. For this problem, different cracks on ferromagnetic 45# steel and nonferromagnetic material aluminum were studied. Cooling curve of different defects on both 45# steel and aluminum were extracted and normalized and then turned into the logarithmic temperaturelogarithmic time curves to obtain a second order differential curve by fitting, furthermore, the minimum peak time of the second order differential was used as the feature quantity to quantified the defect depth. Results show that, as for 45# steel,when the ratio of width to depth is equal to or larger than 2, the relationship between the time corresponding to the minimum peak of second order differential and the square of depth is linear,its possible to measure the depth of the crack. As for aluminum, the time is independent of the depth, and the depth cannot be measured. Finally, the simulation results were verified by experiment.