Method and System for Quantitatively Detecting Copper in Rice Leaves

摘要

#$%^&*AU2020101288A420200813.pdf#####ABSTRACT The present invention provides a method and system for quantitatively detecting copper (Cu) content in rice leaves. The method includes: obtaining rice leaf samples, laser-induced breakdown spectroscopy (LIBS) data and a true Cu content Y; establishing a simple linear regression (SLR) model of Cu emission line intensity and Cu content, and calculating a predicted Cu content Yi based on the model; establishing an exponential regression (ER) model of Cu emission line intensity and Cu content, and obtaining a maximum ReJ; calculating a predicted Cu content Yj based on the model; establishing a multiple linear regression (MLR) model of Cu emission line intensity and Cu content Yd, and calculating a predicted Cu content based on the model; and establishing a linear regression model of multi-equation combination predicted content and true Cu content, and finally determining a Cu content. The method of the present invention contains mathematical relationships such as linear, exponential and non-linear relationships between the LIBS spectra and the Cu content of the rice leaves to the maximum extent. This method retards matrix interference information and thus achieves accurate quantification of the Cu content.1/2 101 Obtain rice leaf samples 102 Acquire laser-induced breakdown spectroscopy (LIBS) data of the rice leaf samples 103 Measure a true copper (Cu) content y in the rice leaf samples by using inductively coupled plasma mass spectrometry (ICP-MS) Obtain a characteristic band with a highest correlation with the true content from the 104 spectral data by using a characteristic variable screening method 105 Quickly locate n Cu emission lines in the characteristic band Use a simple linear regression (SLR) method to establish n SLR models of Cu emission 106 line intensity and Cu content of the test set samples, and obtain correlations RI, R2, ... , Rn between the true Cu content y and predicted contents yl, y2, ... , yn 107 Obtain a maximum correlation Ri among RI, R2, ... , Rn 108 Calculate a predicted Cu content yi based on an SLR model of Cu emission line intensity and Cu content corresponding to the maximum correlation Ri Use an exponential regression (ER) equation to establish n ER models of Cu emission 109 line intensity and Cu content of the test set samples, and obtain correlations Rel, Re2, Ren between the true Cu content y and predicted contents yl, y2, ... , yn 110 Obtain a maximum correlation Rej among Re, Re2, ... , Ren 111 Calculate a predicted Cu content yj of an ER model of Cu emission line intensity and Cu content corresponding to the maximum correlation Rej Use a multiple linear regression (MLR) method to establish n MLR models of Cu 112 emission line intensity and Cu content of the test set samples, and obtain a correlation Rd between the true Cu content y and a predicted content yd; 113 Calculate the predicted Cu content yd based on an MLR model of Cu emission line intensity and Cu content corresponding to the correlation Rd; Establish a linear regression model of test multi-equation combination predicted content 114 and true Cu content by using a linear regression method by taking Cu contents yi, yj and yd as an input vector and the true content y as an output vector 115 Determine a Cu content based on the linear regression model of multi-equation combination predicted content and true Cu content FIG. 1