[Purposes] The paper aims to solve the problem that the defect signal on the key parts of the machinery is difficult to identify accurately under negative effect by the intermediate medium through the interface structure. [Methods] Water and machine oil have always been considered between interface structures as intermediate medium. Here, the relationship between different intermediate media and signal transmission characteristics was deeply studied by combining stochastic resonance theory and Kurtogram algorithm. [Findings] When the area of intermediate medium increases, the relative attenuation rate of the signal energy first decreases and then becomes constant. However, compared with oil, the energy attenuation speed of the signal passing through water is faster. [Conclusions] The signals passing through different intermediate media have been processed by stochastic resonance and fast Kurtogram algorithm. Comparing with the original signals, it is found that the methods can effectively improve the identify ability of signal characteristic parameters. In addition, the relative attenuation rate of the signal energy can be used to characterize the changing law of signals passing through different intermediate media in the frequency domain of signal concentration.