Mixed-noise-resisting blind image source separating method based on feedback mechanism

Abstract

The invention relates to a mixed-noise-resisting blind image source separating method based on a feedback mechanism. The method comprises the steps of utilizing wavelet transform to enable mixed signals to be sparse, carrying out linear cluster on sparse wavelet coefficients to estimate a hybrid matrix of a system, carrying out primary separation according to mixed images, then respectively calculating mean values of separating branches, taking out the most value to be outputted, utilizing a 0-setting feedback method to remove signals of the branch out of original mixed signals, carrying out next blind separation on the residual mixed signals in the method, and repeatedly carrying out the process until only a noise branch is left, namely completely separating all mixed signals. The method can effectively and blindly separate image mixing where Gaussian white noise participates. Compared with classis FastICA, the method can achieve higher separation accuracy.

Description

Technical field: [0001] The invention belongs to the intersection field of digital image processing and blind signal processing, and is a blind image source separation method based on feedback mechanism and anti-mixing Gaussian white noise. Background technique: [0002] Effective denoising is one of the biggest problems encountered in the signal processing field. In the blind source separation (BSS, blind source separation) technology originated from the "cocktail party", most researchers only consider removing channel additive noise (refer to Compare with file 1). In fact, as the source branch signal, there are not only conventional signal branches, but also noise source branches, which participate in the mixing of the system together. The model is as follows figure 1 shown. In the existing blind image source separation algorithm, there is no effective solution to the situation where the noise source branch participates in the system mixing. In summary, it is of great s...

AI Technical Summary

Problems solved by technology

[0004] 1) The ICA-based algorithm can generally remove additive noise well, but the source signals are required to be independent and non-Gaussian. For image mixing with noise, it is difficult to ensure that only one branch is Gaussian property, thus causing unsatisfactory separation results (refer to comparative document 2);

[0005] 2) The SCA-based linear clustering algorithm has a good separation effect under the condition of no noise intervention and the mixed source satisfies the sparse type, but once the noise is involved, the sparsity is reduced, resulting in a sharp decline in the separation effect, and finally the source image cannot be correctly separated (refer to comparative document 3)

Images