Denoising of electromagnetic data from different geological blocks using a hybrid PSO-GWO algorithm and CNN

The wide-field electromagnetic (WFEM) method is widely used in resource exploration in complex geological environments due to its effectiveness in deep detection, interference resistance, and adaptability to various surface conditions. Noise interference remains a significant challenge in areas with high human activity, such as mining regions and urban environments. In these regions, electromagnetic noise is often complex, non-periodic, and pseudo-random, severely degrading the signal-to-noise ratio (SNR) and complicating subsurface interpretation. Traditional denoising methods struggle to effectively handle such complex, non-periodic noise. We propose a novel denoising approach that combines Particle Swarm Optimization (PSO) and Grey Wolf Optimization (GWO) to optimize a Convolutional Neural Network (CNN). This hybrid algorithm enhances CNN’s ability to extract nonlinear features and effectively separate multiple noise types, such as Gaussian white noise, impulse noise, and attenuation noise. The method was tested on WFEM data from the Sichuan Changning Block and the Shanxi Xinyuan Coal Mine area. Simulation results demonstrate that the PSO-GWO-optimized CNN effectively handles complex, non-periodic noise, improving signal clarity and successfully reducing noise interference, thereby enhancing data quality and enabling more accurate geological interpretations. The integration of CNN with PSO-GWO optimization not only improves the SNR but also enhances interpretative accuracy in resource exploration. The successful application of this method in regions with strong electromagnetic interference highlights its broader applicability and practicality.