A novel de-noising method based on Independent Component Analysis(ICA) for DMD based Hadamard Transform Spectral Imager

A new de-noising method based on Independent Component Analysis (ICA) is proposed for imaging characteristics of Digital Micro-mirror Device (DMD) based Hadamard Transform Spectral Imager. As the ubiquitous Gaussian white noises caused by diffractions and other unknown factors in the optical instrument severely confine the usage of the spectral image. ICA is a powerful technique in recovering latent independent sources given only from the mixtures. Based on the fundamental analyzing mode of ICA, the projection of the spectral image is calculated under the transform bases. Then the de-noising processing is carried out by using the soft threshold arithmetic operators. The rebuild spectral image can be acquired by an inverse transform at last. Experiments demonstrate that the proposed ICA algorithm achieves a higher peak signal noise ration (PSNR) and subjective vision effects compared with traditional spectral image de-noising methods.