An improved EMD adaptive denoising and feature extraction algorithm

The noise is often accompanied in the process of extracting the signal from the engineering application which will cause a large error in the processing of the signal and interferer the judgment. Aiming at this problem, this paper proposes an improved EMD adaptive removing noise and feature extraction algorithm. The method firstly uses the nonlinear non-stationary signal adaptive analysis method EMD to decompose the signal, and then selects the appropriate threshold to generate EMD decomposition. The high-frequency component is removed noise, and finally the component of removed noise is filtered by an adaptive filter and combined to generate a new signal, and then the new signal is subjected to frequency domain analysis to extract features. The effectiveness and feasibility of the proposed method are verified by simulation signals and example simulations.