Near infrared spectrum classification model training method and system and classification method and system

A technology of near-infrared spectroscopy and training methods, applied in biological neural network models, character and pattern recognition, instruments, etc., can solve problems such as weak useful information intensity, high noise interference, and inability to ensure the accuracy of classification results

Pending Publication Date: 2021-09-10
YANSHAN UNIV
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, near-infrared spectroscopy has a wide spectral range, a lot of noise interference, and weak useful information. The traditional qualitative model must combine a large number of spectral preprocessing, feature extraction, and dimensionality reduction.
The feature change brought about by this feature extraction and dimensionality reduction is often achieved through data compression, which will destroy the integrity of the spectrum, lose useful information, and fail to ensure the accuracy of the classification results.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Near infrared spectrum classification model training method and system and classification method and system
  • Near infrared spectrum classification model training method and system and classification method and system
  • Near infrared spectrum classification model training method and system and classification method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0045] This embodiment is used to provide a classification model training method for near-infrared spectroscopy, such as figure 1 As shown, the training method includes:

[0046] S1: Obtain multiple near-infrared spectrum sequences for training corresponding to various types of samples;

[0047] Each type of sample corresponds to a preset wavelength range, and the near-infrared spectrum sequence of the sample in the preset wavelength range can be expressed as X={x 1 , x 2 ,...,x i ,...,x n}, n is the number of sequence points in the near-infrared spectrum sequence, and is also the total number of characteristic wavelengths; x i is the absorbance value corresponding to the sample when the wavelength is i.

[0048] For each type of sample, multiple near-infrared spectrum sequences are obtained as training near-infrared spectrum sequences, and the number of training near-infrared spectrum sequences corresponding to different types of samples can be the same or different. Th...

Embodiment 2

[0103] This embodiment is used to provide a kind of near-infrared spectrum classification model training system, such as Figure 12 As shown, the training system includes:

[0104] The first acquisition module M1 is used to acquire multiple training near-infrared spectrum sequences corresponding to various types of samples;

[0105] The first conversion module M2 is used to convert each of the training near-infrared spectral sequences into training two-dimensional images using MTF to obtain training samples; all the training samples and the labels corresponding to the training samples form a training data set ;

[0106] A building block M3 for building an initial classification model;

[0107] The training module M4 is configured to use the training data set to train the initial classification model to obtain a classification model.

[0108] A near-infrared spectrum classification model training system based on Markov transition field (MTF) image coding and residual network...

Embodiment 3

[0110] This embodiment is used to provide a classification method of near-infrared spectrum, such as Figure 13 As shown, the method includes:

[0111] T1: Obtain the near-infrared spectrum sequence to be classified;

[0112] The form of the obtained near-infrared spectrum sequence to be classified is the same as that in S1 in Example 1.

[0113] T2: Using MTF to convert the near-infrared spectrum sequence to be classified into a two-dimensional image;

[0114] As an optional implementation, the near-infrared spectrum sequence to be classified can be clipped and normalized first, which is the same as the method described in Example 1, and the method of converting into a two-dimensional image is also the same as the method used in Example 1 , after being converted into a two-dimensional image, the two-dimensional image can also be further cropped, and the cropping method used is the same as that used in Embodiment 1, and will not be repeated here.

[0115] T3: Using the two-di...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a near infrared spectrum classification model training method and system, and a classification method and system, and the method comprises the steps of converting each training near infrared spectrum sequence into a training two-dimensional image through an MTF, and obtaining a training data set; and constructing an initial classification model, and training the initial classification model by using the training data set to obtain a classification model. A near-infrared spectrum sequence is integrally converted into an image by introducing an MTF method, the dependency of an original spectrum sequence on wavelength and the integrity of characteristics are reserved, a one-dimensional spectrum sequence is encoded into an image, the powerful advantage of machine vision image processing can be applied to classification and recognition of one-dimensional near-infrared spectrums, and the classification accuracy of the classification model can be obviously improved. By using the classification model, the near infrared spectrum sequence can be classified more accurately.

Description

technical field [0001] The present invention relates to the technical field of one-dimensional near-infrared spectrum data analysis and processing, in particular to a classification model training method, system and method for near-infrared spectrum based on MTF (Markov Transition Field) image coding and residual network. Classification methods and systems. Background technique [0002] Due to its advantages of simple operation, high detection efficiency, simultaneous detection of multiple indicators, no damage to the sample, low cost and no pollution in the experiment, near-infrared spectroscopy gradually stands out among many detection technologies mainly based on chemical inspection. The key to classification, recognition and detection of near-infrared spectroscopy is to use the rich spectral information in the near-infrared spectral region to establish a mathematical model with classification ability, and use the generalization ability of the model to detect the spectrum...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/04
CPCG06N3/045G06F18/24G06F18/295G06F18/214
Inventor 王书涛刘诗瑜孔德明
Owner YANSHAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap