Ultraviolet spectrum denoising method based on improved SWO optimization VMD-KPCA algorithm

Abstract

The invention discloses an ultraviolet spectrum denoising method based on an improved SWO optimization VMD-KPCA algorithm, and the method comprises the steps: firstly enabling a spectrum signal to serve as an input, importing the spectrum signal into an improved SWO algorithm (ISWO), and outputting an optimal parameter combination of VMD according to a fitness function through the algorithm; setting variational mode decomposition optimal parameters to perform signal decomposition to obtain intrinsic mode functions (IMF), and performing feature extraction on each IMF by using KPCA (Kernel Principal Component Analysis); and distinguishing the IMF by using a variance contribution rate (VCR) to screen out an effective signal, and finally reconstructing an effective signal component to finally obtain a denoised ultraviolet-visible absorption spectrum signal. The method has the advantages that through the improved SWO algorithm, optimal KPCA and VMD parameters can be selected in a self-adaptive mode, and the denoising effect is improved; the advantages of KPCA and VMD are combined, signals and noise can be effectively separated, the quality of spectral data is improved, the optimized algorithm has the good convergence speed, the calculation complexity is reduced, and the method is suitable for processing large-scale spectral data.

Images