Model transferring method between hyperspectral camera and ordinary spectrograph

Abstract

The invention belongs to the technical field of spectrum detection, and relates to a model transferring method between different kinds of spectrographs. A model transferring method between a hyperspectral camera and an ordinary spectrograph comprises the following steps of (1) collecting spectrum data and a sample concentration value of a sample under the ordinary spectrograph; (2) collecting image data of the sample under the hyperspectral camera; (3) obtaining spectrum data of common use wavelength parts of the spectrum data of the ordinary spectrograph and the average spectrum data of the hyperspectral camera; (4) building an ordinary spectrograph correcting model; (5) dividing a standard set and an unknown set of the two spectrographs; and (6) performing model transferring on the ordinary spectrograph and the hyperspectral camera by using a repeated sample method through being combined with a segmented direct correction and slope / intercept correction method; and substituting the result after the unknown set transferring of the hyperspectral camera into the original ordinary spectrograph correction model to obtain a prediction result after the hyperspectral camera transferring.The method provided by the invention solves the problem of model transferring between the ordinary spectrograph and the hyperspectral camera; a great amount of image information data collected by thehyperspectral camera can be fast predicted; the workload is greatly reduced; and the powerful foundation is provided for the wide application of the hyperspectral camera to quantitative analysis and fast measurement technologies.

Description

technical field [0001] The invention belongs to the technical field of spectrum detection and relates to a model transfer method between different types of spectrometers. Background technique [0002] Spectroscopy has been widely used in quantitative analysis, such as in agriculture, food, industry and other fields. The conventional method is to establish a correlation model for spectral data and concentration data through chemometric methods to realize the prediction of unknown samples. Ordinary spectrometers are very mature for measuring spectral data and establishing related models. Most of them measure the spectral data of the required substances through optical fibers, etc., and then establish models according to needs. Under the condition of using ordinary spectrometers, there have been some studies on the model transfer of factors such as different temperatures, different ordinary spectrometers, different measurement conditions, and samples between different regions....

AI Technical Summary

Problems solved by technology

The above methods cannot well solve the problem of model transfer between hyperspectral cameras and ordinary spectrometers

Compared with the problem of model transfer between ordinary spectrometers, the model transfer between hyperspectral cameras and ordinary spectrometers is more difficult. What a hyperspectral camera acquires is an image data. First, the image data must be processed and analyzed. Hyperspectral cameras are multi-purpose Taking data in the field, not only differences between instruments, but also factors such as illumination and environment will also cause differences in the spectral information of the same sample, which brings great difficulties to the model transfer between ordinary spectrometers and hyperspectral cameras. It is necessary to find a fast, accurate and effective method

Images