35 Constrained Spectrum Denoising Based on Sparse Representation

A method based on a combined use of sparse representation and data fidelity constraints was proposed to remove the noise in the measured spectrum signals. Experiment with signal to noise ratio (SNR), root mean square error (RMSE), and normal correlation coefficient (NCC) three indicators evaluation de-noising performance. Simulation results show that: in comparison with the wavelet denoising methods, the method in this paper has better de-noising ability, and gain better evaluation norm.