Signal Denoising Technology of Debris Flow Based on Empirical Mode Decomposition

The debris-flow infrasound captured by the infrasound sensor often contains numerous interference signals, significantly impacting the signal analysis and evaluation. In order to effectively eliminate or reduce noise signals in the infrasound signals of debris flow disasters, this study proposes an infrasound signal de-noising EMD-based approach. Initially, EMD is employed to collect the Intrinsic Mode Function components (IMFs) of the mudslide infrasound signals. Subsequently, the Pearson coefficients in terms of correlation between the IMF component and the collected signals is computed, facilitating the determination of the IMF noise frequency band based on correlation magnitude. Ultimately, the processed signal is re-constructed to yield the de-noised signal. Experimental findings showcased the effectiveness of the proposed method in filtering out noise information from the signal, thereby presenting a novel approach for debris-flow infrasound signal de-noising.