Raman spectrum identification method and electronic equipment
By treating the Raman spectrum map as a sequence and using digital spiral matrix transformation to construct the Transformer model, the problem of high map similarity and low recognition accuracy in biotechnical drug identification was solved, and high-accuracy biotechnical drug identification was achieved, and through Virtual map generation improves model training efficiency.
CN118918358AActive Publication Date: 2024-11-08SHANGHAI INST FOR FOOD & DRUG CONTROL
Patent Information
- Application Number
- CN202410920266.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-09
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-07-09
Smart Images

Figure CN118918358A_ABST
Abstract
The embodiment of the invention provides a Raman spectrum identification method and electronic equipment, and relates to the technical field of spectrum identification. The Raman spectrum recognition method comprises the steps that longitudinal axis arrays of training maps of N target objects are obtained, and the training maps are Raman spectrum maps of the target objects; selecting a continuous feature array with a preset length from the longitudinal axis array of each training map, and correspondingly replacing a digital element in a preset digital spiral matrix with an intensity value in the feature array corresponding to each training map to obtain a training spiral matrix of the feature array corresponding to each training map; wherein the size of the digital spiral matrix is K * K, the preset length is K2, and K is an integer greater than 1; constructing a Transform model with an input length of K, and training the Transform model by using the training spiral matrix of the feature array of the training atlas to obtain a Transform model meeting conditions; and identifying the Raman spectrum of the to-be-detected object by taking the Transform model meeting the condition as a spectrum identification model to obtain an identification result of the to-be-detected object.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Cited By
Rapid bacillus plant type identification method based on Raman spectrum and deep learning
CN121453744A
