Spectral model transfer method based on CNN-SVR model and transfer learning

Abstract

The invention belongs to the field of spectrum detection and spectrograph model transmission, and provides a spectrum model transfer method based on a CNN-SVR model and transfer learning. The method comprises the following steps: acquiring and preprocessing spectral data, dividing the processed data into a training set and a test set, constructing a main instrument CNN-SVR model, inputting the main instrument training set into the model for training and optimizing to obtain an optimal main instrument CNN-SVR model and hyper-parameters thereof; migrating the model to a slave instrument, freezing a CNN network hyper-parameter value, inputting a slave instrument training set to train and finely adjust SVR parameters, obtaining a migration model based on the CNN-SVR network, and inputting a slave instrument test set into the migration model to predict the model transmission performance. The invention can automatically extract the essential characteristics of the high-dimensional wavelength variables, is suitable for small sample spectrum prediction, and realizes the transmission of the spectrum model among different spectrum instruments by using the characteristics of transfer learning.

Description

technical field [0001] The invention belongs to the field of spectral detection and spectrometer model transfer, and provides a spectral model transfer method based on CNN-SVR model and transfer learning. Background technique [0002] The accuracy of spectral detection is affected by external factors such as spectrometer, measurement environment, and auxiliary materials. When the spectrometer to be tested or the measurement environment changes, the original model may not be able to predict correctly. The most direct way to solve this problem is to re-model, but Due to the high cost of spectral data acquisition, the small number of spectral samples and the variable detection conditions, the cost of remodeling is high and the applicability of the model has not been improved. An effective means to solve this problem is to use the model transfer method, so that the original model can achieve good prediction accuracy under new measurement conditions. There are two main categorie...

AI Technical Summary

Problems solved by technology

[0003] The CNN network has been widely used in many fields and has been verified to have a strong feature extraction ability. For one-dimensional spectral data, the CNN network has been successfully used to achieve spectral classification and regression, but CNN is easy to train on small sample data sets. There is a problem of overfitting

Images