Channel Noise Characterization and Classification Modeling Based on BG Models

The strength of the Abstract-BG (Bernoulli-Gaussian) model lies in its ability to accurately characterize complex noise environments. By combining Gaussian noise and sparse impulse noise, this model effectively captures the common noise characteristics of real signals. When dealing with signals that are disturbed by both Gaussian noise and occasional strong noise, the BG model provides a more realistic description of the noise, which improves the robustness and signal recovery of the processing system. This makes the BG model better able to cope with complex noise situations encountered in practice in applications such as image processing and communication systems. In this paper, based on the PLC (Power Line Carrier) communication channel, we introduce the power line channel noise characteristics and use the BG model to classify and statistically model the noisy signals.