There are inevitably all kinds of noises in weighing process for shearing-type track scale. Its seriously affect real-time and dynamic weighting results of high-speed vehicles. Based on traditional Fourier transform and low-pass filter method, these noises can hardly be deleted. According to this situation, Hilbert-Huang filtering method was proposed for signal de-noising. The signal was decomposed into four layers based on experience mode decomposition algorithm. Three IMF and RE were obtained. Based on given threshold rule, the last one and RE were taken as filter results. The experimental r...