Fiber image highlight detection and elimination method based on hyperspectral imaging

Abstract

The invention discloses a fiber image highlight detection and elimination method based on hyperspectral imaging. The method comprises the steps of detecting a highlight region of a hyperspectral image based on a local space brightness difference method, and outputting a result as a constraint condition of highlight elimination; then, separating the mixed spectrum of the highlight region into a specular reflection component, a diffuse reflection component and a Gaussian noise component through repeated iteration separation based on an approximate sparse representation method of highlight constraint, and finally reserving the diffuse reflection component and using for replacing a mixed spectrum signal of the highlight region of the preprocessed hyperspectral image to obtain a hyperspectral image without highlights, eliminating the influence of highlight, and restoring spectral information, influenced by specular reflection, in the hyperspectral image. The method not only improves the highlight elimination effect, but also reduces the calculation complexity and the calculation time, and is more suitable for practical application.

Description

technical field [0001] The invention relates to the field of computer-aided image processing, in particular to a fiber image highlight detection and elimination method based on hyperspectral imaging. Background technique [0002] Each pixel in a hyperspectral image has a continuous wavelength spectrum with a huge amount of information, which is a major advantage of spectral analysis and greatly improves the efficiency of measurement and analysis. At present, microscopic hyperspectral imaging technology collects hyperspectral images of micron-scale objects in the form of spectral imaging, which is a non-contact and non-destructive color measurement method for micron-scale objects. However, when the hyperspectral imaging system collects hyperspectral images of textiles, due to the influence of sample characteristics, relative arrangement between the camera and the sample, sample geometry, illumination source and observation angle, a small part of the light is directly reflecte...

AI Technical Summary

Problems solved by technology

However, in the hyperspectral images actually collected by the hyperspectral imaging system, the prior information of the light source is not stable and consistent, and the efficiency of highlight removal needs to be improved

Images