A Spectral Reconstruction Method Based on Sample Feature Matching

Abstract

The invention provides a method for spectral reconstruction based on sample feature matching, which is characterized in that it includes the following steps: firstly obtain the color response values ​​of the training sample set and the test sample through a digital imaging device under an ambient light source; then use the sample feature matching method Filter the training sample set to obtain the clustering training sample color response value set and spectrum set; then normalize the clustering training sample color response value and test sample color response value, and use the least square method to calculate the coefficient matrix; finally, use The coefficient matrix and curve smoothing methods are used to reconstruct the spectral reflectance of the object surface. The invention provides a spectral reconstruction method with high reconstruction accuracy, simple calculation, full consideration of the physical characteristics of the spectral reflectance itself, and screening by the characteristics of the sample itself.

Description

technical field [0001] The invention relates to a spectral reconstruction method, in particular to a spectral reconstruction method based on sample feature matching, which can be widely used in the fields of textile color, printing detection, art reproduction, e-commerce and the like. Background technique [0002] There are two ways to quantitatively represent object color information: color information based on chromaticity and color information based on object surface spectral reflectance. Color information based on chromaticity is usually represented by color tristimulus values. For example, colors in ordinary digital cameras, LCD monitors or multimedia projectors are usually represented by R, G, and B tristimulus values, and these tristimulus values ​​can be converted to Other corresponding tristimulus values, such as CIE tristimulus values ​​X, Y, Z or Lab values, etc. The color tristimulus value is the color value obtained by integrating the relative spectral power di...

AI Technical Summary

Problems solved by technology

However, the above spectral reconstruction method only considers the optimal value in the mathematical sense, ignoring the characteristics of the sample itself and the physical meaning of the spectral reflectance (for example, the value of the spectral reflectance has a value not greater than 1 or not less than 0), resulting in reconstruction There is a large deviation between the physical value and smoothness of the spectral reflectance and the actual object

Images