一种拉曼光谱下利用神经网络对混合物各组分分析的方法
By designing an RH-CNN neural network and combining Raman characteristic peaks and spectral features, the difficulties in qualitative and quantitative analysis of mixture components were solved, and high-precision spectral analysis was achieved.
CN115828077BActive Publication Date: 2026-07-17PLANTS & ANIMALS & FOOD TESTING QUARANTINE TECH CENT SHANGHAI ENTRY EXIT INSPECTION & QUARANTINE BUREAU +1
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- PLANTS & ANIMALS & FOOD TESTING QUARANTINE TECH CENT SHANGHAI ENTRY EXIT INSPECTION & QUARANTINE BUREAU
- Filing Date
- 2021-09-17
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing neural networks fail to effectively consider the importance of Raman characteristic peaks in Raman spectroscopy analysis, making qualitative and quantitative analysis of mixture components difficult.
Method used
The RH-CNN neural network is designed, and the characteristic matrix of Raman feature peaks and the Raman spectral feature matrix are combined and trained using an adaptive weight update method to obtain a classification model.
Benefits of technology
It improves the identification effect of spectral vibrations of mixtures, with a qualitative analysis accuracy of 98.30% and a quantitative analysis accuracy of 93.10%, and can effectively obtain fine-grained characteristics of Raman spectra.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN115828077B_ABST
Abstract
本发明公开了一种拉曼光谱下利用神经网络对混合物各组分分析的方法，先获取拉曼特征峰的特性矩阵N与拉曼光谱特征矩阵V，利用神经网络将拉曼特征峰的特性矩阵N与拉曼光谱特征矩阵V组合使用，采用自适应的方法进行训练，从而得到分类模型。本发明的方法，可以提高对混合物光谱振动的识别效果，以及有益于获取拉曼光谱特征的细粒度。
Need to check novelty before this filing date? Find Prior Art