CN-122023901-A - AI-based infrared spectrum mineral rapid identification system

Abstract

The invention provides an AI-based infrared spectrum mineral rapid identification system. The rapid identification system for the infrared spectrum minerals based on the AI comprises an infrared spectrum data acquisition module and a data preprocessing module, wherein the infrared spectrum data acquisition module is used for acquiring infrared spectrum data of mineral samples, the data preprocessing module is used for denoising, smoothing and normalizing the acquired infrared spectrum data, the characteristic extraction and selection module is used for extracting effective characteristics from the preprocessed spectrum data, and the AI identification and classification module is used for carrying out mineral classification identification on the extracted characteristics by using a deep learning algorithm. According to the AI-based infrared spectrum mineral rapid identification system, high-quality mineral spectrum data are collected through the infrared spectrum data collection module, and denoising, smoothing and standardization processing are carried out through the data preprocessing module, so that noise interference is reduced, data quality is improved, and the problem that the system excessively depends on spectrum data quality is avoided.

Inventors

• XIONG ZILI
• YU LIANG

Assignees

• 韩山师范学院

Dates

Publication Date

20260512

Application Date

20260129

Claims (8)

1. 1. An AI-based infrared spectrum mineral rapid identification system, comprising: The infrared spectrum data acquisition module is used for acquiring infrared spectrum data of the mineral sample; The data preprocessing module is used for carrying out denoising, smoothing and normalization processing on the collected infrared spectrum data; the feature extraction and selection module is used for extracting effective features from the preprocessed spectrum data; the AI recognition and classification module is used for carrying out mineral classification recognition on the extracted features by using a deep learning algorithm; and the result output and feedback module is used for outputting the identification result and providing feedback.
2. 2. The rapid identification system of AI-based infrared spectrum minerals as set forth in claim 1, wherein the infrared spectrum data acquisition module uses an infrared spectrometer with an acquisition range of 1000cm -1 to 4000cm -1 .
3. 3. The rapid identification system of infrared spectrum minerals based on AI of claim 1, wherein the data preprocessing module comprises a denoising algorithm, a wavelet transform smoothing algorithm and a standardized processing module, wherein the data preprocessing module removes environmental noise in spectrum data, smoothes fluctuation and normalizes data range, and the feature extraction and selection module adopts a principal component analysis method to reduce dimension of the spectrum data.
4. 4. The system for rapidly identifying minerals by infrared spectrum based on AI according to claim 1, wherein said AI identification and classification module is trained by self-supervised learning method using unlabeled mineral samples, and said AI identification and classification module adopts an algorithm combining convolutional neural network and support vector machine.
5. 5. The rapid identification system of minerals by infrared spectrum based on AI of claim 1 wherein the result output and feedback module comprises a graphical user interface for displaying the results of the identification of minerals.
6. 6. The system of claim 1, wherein the data preprocessing module is configured to automatically detect and filter noise.
7. 7. The rapid identification system of minerals by infrared spectrum based on AI of claim 1 wherein the result output and feedback module comprises a multi-mode feedback mechanism capable of providing graphical, textual and speech feedback based on the identification result.
8. 8. The rapid identification system of AI-based infrared spectrum minerals as set forth in claim 7 wherein the multi-modal feedback mechanism comprises graphical feedback, text feedback, voice feedback, alarm feedback, and a combination of graphics and text feedback.

Description

AI-based infrared spectrum mineral rapid identification system Technical Field The invention relates to the technical field of rapid identification of infrared spectrum minerals, in particular to an AI-based rapid identification system of infrared spectrum minerals. Background The AI-based infrared spectrum mineral rapid identification system collects spectrum data of mineral samples through an infrared spectrometer and performs pretreatment such as denoising and smoothing to improve data quality. The system then extracts key features, typically including peak position, absorption intensity, etc., associated with the mineral features from the spectral data. Through machine learning or deep learning algorithms, such as support vector machines or convolutional neural networks, the system can input the features into a trained model, and compare the model with a known mineral spectrum library, so that automatic classification and identification of minerals are realized. Finally, the recognition result is presented to the user through a visual interface or other feedback means, providing detailed information of the minerals, such as chemical composition or potential use. Although this system has high efficiency and accuracy in mineral identification, there are some drawbacks. The system is highly dependent on the quality of the spectral data, and any errors in the acquisition process or environmental disturbances may lead to inaccurate recognition results. AI models often rely on a large number of labeled samples during training, and if the training dataset is inadequate, particularly for some rare minerals, the model's ability to identify may be limited and may not be able to handle the type of mineral that is not found or the complex sample. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an AI-based infrared spectrum mineral rapid identification system, which solves the problem that the identification result is inaccurate due to the fact that the system is highly dependent on the quality of spectrum data, and the problem that the type of the mineral which is not seen or the complex sample cannot be processed due to the fact that a large number of marked samples are relied on during training. In order to achieve the purpose, the invention is realized by the following technical scheme that the AI-based infrared spectrum mineral rapid identification system comprises: The infrared spectrum data acquisition module is used for acquiring infrared spectrum data of the mineral sample; The data preprocessing module is used for carrying out denoising, smoothing and normalization processing on the collected infrared spectrum data; the feature extraction and selection module is used for extracting effective features from the preprocessed spectrum data; the AI recognition and classification module is used for carrying out mineral classification recognition on the extracted features by using a deep learning algorithm; and the result output and feedback module is used for outputting the identification result and providing feedback. Preferably, the infrared spectrum data acquisition module uses an infrared spectrometer, and the acquisition range of the infrared spectrometer is 1000cm -1 to 4000cm -1, and the infrared spectrometer has higher resolution to adapt to various mineral samples. Preferably, the data preprocessing module comprises a denoising algorithm, a wavelet transform smoothing algorithm and a standardized processing module, removes environmental noise in spectrum data, smoothes fluctuation and standardizes a data range so as to improve data quality, and the feature extraction and selection module adopts a principal component analysis method to reduce dimension of the spectrum data and reduce redundant features. Preferably, the AI identification and classification module trains by using unlabeled mineral samples through a self-supervision learning method to expand the identification range of the system and improve the adaptability to new samples, and adopts an algorithm combining a convolutional neural network and a support vector machine, wherein the convolutional neural network is used for automatically extracting high-dimensional features, and the support vector machine is used for carrying out classification identification. Preferably, the result output and feedback module comprises a graphical user interface for displaying the result of mineral identification, and has a data storage function for recording historical identification data for subsequent analysis. Preferably, the data preprocessing module further improves the processing capacity of the system on unstable data through an automatic noise detection and filtering method, and reduces the requirement of human intervention. Preferably, the result output and feedback module comprises a multi-mode feedback mechanism capable of providing graphic, text and voice feedback according to the recognition result, so that th