A Spectral Reconstruction Method Based on Single RGB Image of Color Digital Camera

Abstract

The invention belongs to the technical field of spectral imaging and relates to a color digital camera single RGB (red green blue) image based spectral reconstruction method. The method includes: measuring the spectral reflectance of training samples; setting up a shooting environment, shooting the training samples and an RGB image of a reconstruction object and correcting the image; extracting RGB response values of the training samples and reconstruction object pixels, calculating Euclidean distance of the RGB value between any pixel and the training samples, performing ascending sort on the training samples from small to large according to the Euclidean distance, and taking the first p training samples as optimal training sample sets; calculating inverse distance weighting matrixes of the optimal training sample sets; expanding the response values of the optimal training sample sets and the pixels to be reconstructed; weighting response value extension matrixes and spectral reflectance matrixes of the optimal training sample sets by the aid of the inverse distance weighting matrixes; calculating spectral reconstruction transformation matrixes by the aid of a pseudoinverse method, and reconstructing the spectral reflectance of the corresponding pixels to acquire the spectral image of the reconstruction object.

Description

Technical field [0001] The invention belongs to the technical field of spectral imaging, and specifically relates to a spectral reconstruction method based on a single RGB image of a color digital camera. Background technique [0002] Because spectral imaging technology has high imaging quality capabilities in terms of spectrum and chromaticity, and has the function of unifying maps, it has developed rapidly in the fields of remote sensing, color imaging, biomedicine, cultural heritage, and art research. In order to meet the specific applications in the above different fields, various types of spectral imaging systems have been produced, such as camera + line scan spectrometer, camera + filter wheel, camera + bandpass filter (broadband or narrowband), camera + multiple Spectral light sources, monochrome cameras + acousto-optic or liquid crystal tunable filters, etc. are often divided into two categories in the field of color science, narrowband and broadband. For the above-menti...

AI Technical Summary

Problems solved by technology

However, most of the existing studies use standard polynomial expansion models, ignoring the influence of camera response value expansion items on spectral reconstruction accuracy

In addition, the current method mostly uses the entire training sample set to reconstruct all test samples, and all training samples have the same weight, ignoring the influence of the number of training samples and the correlation between them and the test samples on the spectral reconstruction accuracy, which restricts the Overall Spectral Reconstruction Accuracy

[0004] For the above problems, there are no reasonable and effective solutions in academia and industry.

Images