A Raman Spectral Image Unmixing Method Based on Non-negative Matrix Adjacency

Abstract

The invention discloses a Raman spectral image unmixing method based on non-negative matrix proximity. The method aims at the absence of impure pixels in the Raman spectral image, and unknown ANC, ASC and endmembers in the Raman spectral unmixing. When the Raman spectral image satisfies the LMM model, by adding a variety of regularization factors such as sparseness to the non-negative matrix, the components of the mixed pixels and their corresponding abundances can be effectively obtained. Compared with other existing methods, the present invention has the following advantages: (1) the number of endmembers (constituent substances) is not required as prior knowledge; (2) an iterative strategy is adopted, and the obtained result has nothing to do with the number of endmembers; (3) The graph relationship and 1 / 2 norm are added to improve the noise resistance and stability of the algorithm.

Description

technical field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a Raman spectral image unmixing method based on non-negative matrix proximity. Background technique

[0002] Raman spectroscopy is a three-dimensional image, including ordinary two-dimensional plane image information and wavelength information. While imaging the spatial characteristics of the target, each spatial pixel is dispersed to form dozens or even hundreds of narrow bands for continuous spectral coverage. A Raman image is a three-dimensional Raman image composed of two-dimensional images corresponding to several wavelengths.

[0003] Raman spectroscopy is widely used in industries such as cell engineering (detection of cell wall structural components) because of its unique spectral characteristics for different substances. However, most of the current Raman images are synthesized by mixing multiple different substances (endmembers). In order to analy...

Images