A novel entropy-based method for signal and noise classification in time-frequency analysis

This paper presents a novel entropy-based method for the classification of detection signals and noise, utilizing timefrequency analysis to distinguish different types of detection signals and noise. We first apply the Short-Time Fourier Transform (STFT) to transform signals into the time-frequency domain, followed by the Fractional Fourier Transform (FRFT) to extract features at various fractional orders. The spectral entropy of the signals in the fractional Fourier domain is then computed to quantitatively assess the sparsity of the signals, which serves as the criterion for signal and noise classification. To validate the effectiveness of the proposed algorithm, we conduct experimental analyses on different types of signals, including Continuous Wave (CW), Linear Frequency Modulated (LFM), and Hyperbolic Frequency Modulated (HFM) signals. The experimental results demonstrate that the proposed method can accurately classify different types of detection signals under noisy conditions and further distinguish the modulation schemes through feature extraction. Compared to traditional classification methods, the entropy-based classification strategy exhibits significant advantages in terms of noise resistance and classification accuracy. The findings of this study provide an efficient and robust classification tool for signal processing in complex environments.