A method for denoising and reconstruction of EM leakage information based on optimized VMD joint singular spectrum analysis

The electromagnetic(EM) waves unintentionally emitted by computer display devices contain valuable information, but they are easily interfered with by noise during propagation, seriously affecting the recovery of the leaked information. To address this problem, this paper proposes a video EM leakage signal preprocessing method using improved Triangular Topology Aggregation Optimized VMD combined with SSA to suppress noise. The method enhances the TTAO algorithm to optimize VMD’s penalty factor and mode number. It then performs permutation entropy analysis on the decomposed modes and utilizes a relative energy threshold criterion to adaptively improve Singular Spectrum Analysis for singular value screening and denoising. Simulation results demonstrate this method’s superiority over VMD, SSA, and VMD-SSA in SNR, RMSE, and CC metrics. For measured EM leakage signals, the proposed method achieved an average noise rejection ratio of 13.739 dB, outperforming other techniques. Image reconstruction experiments confirm its effectiveness in restoring EM leakage information.