Talc grade classification device based on laser-induced breakdown spectroscopy

Abstract

The invention discloses a talc grade classification device based on laser-induced breakdown spectroscopy, which comprises an LIBS (laser-induced breakdown spectroscopy) detection module, an intelligent classification module and an automatic sorting module, the intelligent classification module is used for obtaining the quality grade of the talc sample according to the characteristic spectral line graph, and the automatic sorting module is used for sorting the talc sample according to the received quality grade. According to the method, laser-induced breakdown spectroscopy is adopted, so that the classification accuracy and the sorting accuracy of the quality grade of the talc are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of talc detection, and more particularly, relates to a talc grade classification device based on laser induced breakdown spectroscopy analysis. Background Art

[0002] As a key industrial mineral, talc's grade classification is crucial for quality control in downstream industries. Traditional methods, such as flotation and hand sorting, have been used for classification. These methods are not only inaccurate (resource utilization rate <60%) but also rely heavily on manual experience, making them difficult to adapt to industrial mass production. Existing detection technologies, such as X-ray fluorescence spectroscopy (XRF) and near-infrared spectroscopy (NIRS), can analyze composition, but they suffer from expensive equipment and delayed response times, making them inadequate for dynamic conveyor belt monitoring.

[0003] Laser-induced breakdown spectroscopy (LIBS) technology has been used for mineral composition detection and other industrial scenarios due to its advantages such as no need for preprocessing and rapid analysis. However, when detecting moving objects, LIBS technology still has some urgent problems that need to be solved, such as relatively large measurement errors and insufficient detection limits, which also cause significant interference with the detection accuracy and repeatability of LIBS technology. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a talc grade classification device based on laser induced breakdown spectroscopy analysis, which uses laser induced breakdown spectroscopy to improve the classification accuracy and sorting accuracy of talc quality grades.

[0005] In order to achieve the above-mentioned purpose of the invention, the talc grade classification device based on laser induced breakdown spectroscopy analysis of the present invention includes a LIBS detection module, an intelligent classification module and an automatic sorting module, wherein:

[0006] The LIBS detection module is used to excite the talc sample with laser, collect the plasma spectrum generated by the excitation, and analyze the The characteristic spectrum of the elements is generated, and the specific elements and quantities are set according to actual needs, and then the characteristic spectrum is sent to the intelligent classification module;

[0007] The intelligent classification module uses a pre-trained neural network to predict the quality grade of the talc sample based on the characteristic spectrum and sends the quality grade to the automatic sorting module;

[0008] The automatic sorting module is used to sort the talc samples according to the received quality grades.

[0009] Furthermore, the LIBS detection module includes a detection photoelectric sensor, a laser excitation device, a plasma light collection device and a spectrum analysis device, wherein:

[0010] The detection photoelectric sensor is used to detect the presence of the talc sample and generate a detection trigger signal, and send the detection trigger signal to the laser excitation device and the plasma light collection device;

[0011] The laser excitation device is used to emit laser and excite the talc sample after receiving the detection trigger signal;

[0012] The plasma light collection device is used to collect the plasma light generated by the excitation and transmit it to the spectrum analysis device after receiving the detection trigger signal;

[0013] The spectrum analysis device is used to perform spectrum analysis on the received plasma light to obtain Characteristic spectrum of elements.

[0014] Furthermore, the laser excitation device adopts Nd:YAG laser.

[0015] Furthermore, the plasma light collection device includes an optical lens and a fiber optic probe, wherein the optical lens is used to collimate and focus the plasma light generated by the excitation, and the fiber optic probe is used to directly transmit the plasma light passing through the optical lens to the spectrum analysis device.

[0016] Furthermore, the spectrum analysis device adopts a three-channel spectrometer configured with a charge coupled device or a complementary metal oxide semiconductor sensor.

[0017] Furthermore, the characteristic spectrum diagram contains 6 elements, namely Al, Ca, Mg, Si, Fe, and Na.

[0018] Furthermore, the intelligent classification module includes a data preprocessing module, an element feature extraction module, a PCA dimension reduction module and a classifier, wherein:

[0019] The data preprocessing module is used to filter out the spectrum data at the stable stage at the laser focus from the characteristic spectrum line diagrams received at all times, and then send it to the element feature extraction module;

[0020] The element feature extraction module is used to extract the element features from the received spectral data. The characteristic peak of the element is The characteristic peak matrix , Indicates the number of characteristic peaks, and then the characteristic peak matrix Send to PCA dimensionality reduction module;

[0021] The PCA dimensionality reduction module is used to perform PCA dimensionality reduction on the feature peak matrix. Send to classifier;

[0022] The classifier is used to classify the feature peak matrix The quality grade of talc samples was predicted.

[0023] Furthermore, the spectral data screening in the data preprocessing module adopts the descending minimum fluctuation screening method, and the specific method is: for the preprocessed talc spectral data, the spectral characteristic peak with a wavelength of 396.254 nm is selected, and the peak intensity is arranged in descending order. The first 5 data are selected to form a data set, and the relative standard deviation of the data set is calculated. One new data is added in sequence to form a new data set and the relative standard deviation is calculated. The data set with the smallest relative standard deviation is selected and used as the spectral data in the stable stage.

[0024] Furthermore, the classifier is a support vector machine.

[0025] Furthermore, the automatic sorting module includes a sorting photoelectric sensor, a programmable logic controller and a sorting actuator, wherein:

[0026] The sorting photoelectric sensor is used to generate a sorting trigger signal when the talc sample passes through and send it to the programmable logic controller;

[0027] After receiving the sorting trigger signal, the programmable logic controller is used to calculate the time when the talc sample arrives at the sorting actuator according to the conveyor belt speed and the distance between the photoelectric sensor and the sorting actuator, and generate a sorting instruction based on the quality grade and send it to the sorting actuator;

[0028] The sorting execution mechanism is used to complete the sorting action according to the sorting instruction.

[0029] The present invention discloses a talc grade classification device based on laser induced breakdown spectroscopy analysis, comprising a LIBS detection module, an intelligent classification module and an automatic sorting module. The LIBS detection module is used to excite a talc sample with a laser and obtain a characteristic spectrum line graph. The intelligent classification module is used to obtain the quality grade of the talc sample according to the characteristic spectrum line graph. The automatic sorting module is used to sort the talc sample according to the received quality grade.

[0030] The present invention has the following beneficial effects:

[0031] 1) The present invention obtains the quality grade of talc based on the analysis results of laser-induced breakdown spectroscopy with high accuracy;

[0032] 2) The device of the present invention can perform real-time quality grade detection and subsequent sorting of talc samples on the conveyor belt, effectively adapting to the dynamic scene of the conveyor belt and can be directly integrated into the existing production line to improve overall processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a structural diagram of a specific embodiment of a talc grade classification device based on laser-induced breakdown spectroscopy analysis according to the present invention;

[0034] Figure 2 is an example diagram of the characteristic spectrum of the talc sample in this embodiment;

[0035] Figure 3 is an example diagram of PCA analysis in this embodiment;

[0036] Figure 4 This is an example diagram of quality level prediction using the support vector machine in this embodiment. DETAILED DESCRIPTION

[0037] The following describes the specific embodiments of the present invention in conjunction with the accompanying drawings so that those skilled in the art can better understand the present invention. It should be noted that in the following description, when detailed descriptions of known functions and designs may dilute the main content of the present invention, such descriptions will be omitted here.

[0038] Example

[0039] Figure 1 This is a structural diagram of a specific embodiment of the talc grade classification device based on laser induced breakdown spectroscopy analysis of the present invention. Figure 1 As shown, the talc grade classification device based on laser induced breakdown spectroscopy analysis of the present invention includes a LIBS detection module 1, an intelligent classification module 2 and an automatic sorting module 3. Each module will be described in detail below.

[0040] LIBS detection module 1 is used to excite the talc sample with laser, collect the plasma spectrum generated by the excitation, and analyze the The characteristic spectrum line diagram of the elements is generated, and the specific elements and quantities are set according to actual needs, and then the characteristic spectrum line diagram is sent to the intelligent classification module 2.

[0041] like Figure 1 As shown, the LIBS detection module 1 in this embodiment includes a detection photoelectric sensor 11, a laser excitation device 12, a plasma light collection device 13 and a spectrum analysis device 14, wherein:

[0042] The detection photoelectric sensor 11 is used to detect the presence of the talc sample and generate a detection trigger signal, and send the detection trigger signal to the laser excitation device 12 and the plasma light collection device 13.

[0043] The laser excitation device 12 is used to emit laser light upon receiving a detection trigger signal to excite the talc sample. In this embodiment, the laser excitation device 12 uses an Nd:YAG laser with a wavelength of 1064 nm, a pulse energy of 120 mJ, and a repetition rate of 20 Hz. In practice, the laser pulse, energy, frequency, and delay time can be controlled based on the actual quality of the talc sample to effectively excite the sample.

[0044] When the talc sample placement platform is a fixed platform, an optical path adjustment module can be configured in the laser excitation device 12 to automatically adjust the distance between the laser focusing lens and the talc sample to ensure that the laser can act on the surface of the talc sample, thereby ensuring efficient excitation. When the talc sample placement platform is a conveyor belt on a production line, a tracking module can be configured in the laser excitation device 12 to synchronize the laser light emitted by the laser excitation device 12 with the talc sample on the conveyor belt, so that the laser light can be stably focused on the surface of the talc sample within a certain period of time. In addition, in actual applications, to ensure operational safety, it is generally necessary to configure the laser excitation device 12 with safety devices such as a laser protective cover (protection level CLASS IV), an emergency stop button, and an overheating protection system.

[0045] After receiving a detection trigger signal, the plasma light collection device 13 collects the plasma light generated by the excitation and transmits it to the spectral analysis device 14. In this embodiment, the spectral sampling device 13 includes an optical lens (focal length 75 mm) and a fiber optic probe (core diameter 200 μm). The optical lens is used to collimate and focus the plasma light generated by the excitation, and the fiber optic probe is used to transmit the plasma light directly to the spectral analysis device 14 after passing through the optical lens, reducing light scattering and absorption during transmission. In practical applications, the design of the plasma light collection device 12 can be optimized according to the actual application scenario to reduce light loss and distortion while improving light collection efficiency.

[0046] The spectrum analyzer 14 is used to perform spectrum analysis on the received plasma light to obtain Characteristic spectrum of the elements. This device is the core of the entire LIBS detection module 1 and can decompose the light emitted by the plasma into spectra of different wavelengths. A highly sensitive spectrometer can detect weak spectral signals, which is crucial for analyzing trace elements or low-concentration substances in the plasma. In this embodiment, the spectral analysis device 14 uses a three-channel spectrometer equipped with a charge-coupled device (CCD) or complementary metal oxide semiconductor (CMOS) sensor, with a wavelength range of 200-930 nm, a spectral integration time of 50 ms, and a signal-to-noise ratio of >30:1, thereby quickly responding to and accurately recording optical signals.

[0047] The elements included in the characteristic spectrum have a great influence on the accuracy of subsequent talc grade classification. In order to improve the accuracy, the characteristic spectrum in this embodiment includes 6 elements, namely Al, Ca, Mg, Si, Fe, and Na. Figure 2 This is an example of the characteristic spectrum of the talc sample in this embodiment. Figure 2 As shown, in this embodiment, three quality grades are set for talc samples. The characteristic values ​​of the six elements in the characteristic spectrum of different qualities are obviously different. Therefore, a combination of the six element characteristic values ​​can be used to achieve accurate quality grade classification.

[0048] The intelligent classification module 2 uses a pre-trained neural network to predict the quality grade of the talc sample based on the characteristic spectrum, and sends the quality grade to the automatic sorting module 3.

[0049] like Figure 1 As shown, the intelligent classification module 2 in this embodiment includes a data preprocessing module 21, an element feature extraction module 22, a PCA dimension reduction module 23 and a classifier 24, wherein:

[0050] The data preprocessing module 21 is used to filter out spectral data at the stable phase at the laser focus from the received characteristic spectral line graphs at all times, and then send it to the element feature extraction module 22. In this embodiment, the spectral data screening adopts the descending minimum fluctuation screening method. The specific method is as follows: for the preprocessed talc spectral data, the spectral characteristic peak with a wavelength of 396.254 nm is selected, and the peak intensity is sorted in descending order. The first five data are selected to form a data set, and the relative standard deviation (RSD) of this data set is calculated. New data are then added one by one to form a new data set, and the relative standard deviation is calculated. The data set with the smallest relative standard deviation is selected and used as the spectral data for the stable phase.

[0051] The element feature extraction module 22 is used to extract the element feature from the received spectral data. The characteristic peak of the element is The characteristic peak matrix , Indicates the number of characteristic peaks, and then the characteristic peak matrix The data are sent to the PCA dimension reduction module 23. In this embodiment, the extraction window of the characteristic peak is ±0.1 nm. Figure 3 This is an example diagram of PCA analysis in this embodiment.

[0052] The PCA dimension reduction module 23 is used to perform PCA dimension reduction on the characteristic peak matrix, and the reduced characteristic peak matrix Sent to classifier 24.

[0053] The classifier 24 is used to classify the peak value matrix according to the characteristic peak value matrix. The quality grade of the talc sample is predicted. The specific model of the classifier 24 can be set according to actual needs. In this embodiment, three classifiers were compared: Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Random Forest (RF). Ultimately, the Support Vector Machine was used with hyperparameters γ = 0.1 and C = 16. Figure 4 This is an example diagram of the quality level prediction of the support vector machine in this embodiment. Figure 4 As shown, the talc sample is a grade 2 product.

[0054] A number of talc grade classification samples are collected in advance, each talc grade classification sample includes a characteristic spectrum and a corresponding quality grade, and the intelligent classification module 2 is trained, and the loss function adopts the cross entropy loss.

[0055] The automatic sorting module 3 is used to sort the talc samples according to the received quality grades. As the key link in the present invention for realizing the automated processing of talc after grade classification, the automatic sorting module 3 needs to quickly and accurately sort the talc on the conveyor belt according to the talc quality grade output by the classification model, ensuring that talc samples of different quality grades are transported to the corresponding collection areas, meeting the efficiency and accuracy requirements of the grading and processing of talc samples in industrial production. When the talc samples are placed on a fixed platform, the implementation of the automatic sorting module 3 is relatively simple. In order to adapt to the application in the production line, an automatic sorting module 3 suitable for a conveyor belt is designed in this embodiment, including a sorting photoelectric sensor 31, a programmable logic controller (PLC) 32 and a sorting actuator 33, wherein:

[0056] The sorting photoelectric sensor 31 is used to generate a sorting trigger signal when a talc sample passes through and send the signal to the programmable logic controller 32 .

[0057] After receiving the sorting trigger signal, the programmable logic controller 32 is used to calculate the time when the talc sample reaches the sorting actuator 33 based on the conveyor belt speed and the distance between the photoelectric sensor 31 and the sorting actuator 33, and generate a sorting instruction based on the quality grade and send it to the sorting actuator 33.

[0058] The sorting actuator 33 is used to perform sorting according to the sorting instructions. In this embodiment, the talc has three quality grades. Therefore, the sorting actuator 33 includes three sets of pneumatic push rods (response speed 50 ms) and bins. The pneumatic push rods corresponding to the quality grade extend (stroke 30 mm) when the talc sample reaches the predetermined position, pushing the talc sample into the corresponding bin.

[0059] Next, the workflow of the talc grade classification device in this embodiment is described using a production line application scenario as an example. The specific steps are as follows:

[0060] Step 1: The talc sample is transported to the detection area via a conveyor belt. The photoelectric sensor triggers the laser excitation device, and the laser is focused on the sample surface to generate plasma.

[0061] Step 2: The plasma light collection device captures the plasma light, and the spectrum analysis device performs spectrum analysis to obtain a characteristic spectrum line diagram, which is then transmitted to the intelligent classification module.

[0062] Step 3: The intelligent classification module pre-processes the characteristic spectrum and extracts element features. After PCA dimension reduction and SVM model classification, the talc quality grade is obtained and sent to the automatic sorting module.

[0063] Step 4: When the talc sample arrives at the sorting area, the photoelectric sensor in the automatic sorting module generates a sorting trigger signal, the programmable logic controller generates a sorting stall, and the corresponding pneumatic push rod in the sorting actuator moves to push the talc sample into the corresponding material box to complete the sorting.

[0064] Table 1 is a statistical table of the talc sample sorting results in this embodiment.

[0065] sample Number of real samples Number of correct predictions Number of incorrect predictions Grade 1 21 19 2 Grade 2 22 22 0 Grade 3 22 21 1 total 65 62 3 Test set accuracy —— —— 95.38%

[0066] Table 1

[0067] After experimental statistics, the classification accuracy of the present invention reached 95.38%, and the sorting accuracy was 100%, which is significantly better than the existing device. Although the above description of the illustrative embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations based on the concept of the present invention are protected.

Claims

1. A talc grade classification device based on laser induced breakdown spectroscopy analysis, characterized in that: It includes LIBS detection module, intelligent classification module and automatic sorting module, among which: The LIBS detection module is used to excite the talc sample with laser, collect the plasma spectrum generated by the excitation, and analyze the The characteristic spectrum of the elements is generated, and the specific elements and quantities are set according to actual needs, and then the characteristic spectrum is sent to the intelligent classification module; The intelligent classification module uses a pre-trained neural network to predict the quality grade of the talc sample based on the characteristic spectrum and sends the quality grade to the automatic sorting module; The automatic sorting module is used to sort the talc samples according to the received quality grades.

2. The talc grade classification device according to claim 1, characterized in that: The LIBS detection module includes a detection photoelectric sensor, a laser excitation device, a plasma light collection device and a spectrum analysis device, wherein: The detection photoelectric sensor is used to detect the presence of the talc sample and generate a detection trigger signal, and send the detection trigger signal to the laser excitation device and the plasma light collection device; The laser excitation device is used to emit laser and excite the talc sample after receiving the detection trigger signal; The plasma light collection device is used to collect the plasma light generated by the excitation and transmit it to the spectrum analysis device after receiving the detection trigger signal; The spectrum analysis device is used to perform spectrum analysis on the received plasma light to obtain Characteristic spectrum of elements.

3. The talc grade classification device according to claim 2, characterized in that: The laser excitation device adopts Nd:YAG laser.

4. The talc grade classification device according to claim 2, characterized in that: The plasma light collection device includes an optical lens and a fiber optic probe, wherein the optical lens is used to collimate and focus the plasma light generated by excitation, and the fiber optic probe is used to directly transmit the plasma light passing through the optical lens to the spectrum analysis device.

5. The talc grade classification device according to claim 2, characterized in that: The spectrum analysis device adopts a three-channel spectrometer configured with a charge coupled device or a complementary metal oxide semiconductor sensor.

6. The talc grade classification device according to claim 1, characterized in that: The characteristic spectrum line diagram contains 6 elements, namely Al, Ca, Mg, Si, Fe, and Na.

7. The talc grade classification device according to claim 1, characterized in that: The intelligent classification module includes a data preprocessing module, an element feature extraction module, a PCA dimension reduction module and a classifier, wherein: The data preprocessing module is used to filter out the spectrum data at the stable stage at the laser focus from the characteristic spectrum line diagrams received at all times, and then send it to the element feature extraction module; The element feature extraction module is used to extract the element features from the received spectral data. The characteristic peak of the element is The characteristic peak matrix , Indicates the number of characteristic peaks, and then the characteristic peak matrix Send to PCA dimensionality reduction module; The PCA dimensionality reduction module is used to perform PCA dimensionality reduction on the feature peak matrix. Send to classifier; The classifier is used to classify the feature peak matrix The quality grade of talc samples was predicted.

8. The talc grade classification device according to claim 7, characterized in that: The spectral data screening in the data preprocessing module adopts the descending minimum fluctuation screening method. The specific method is as follows: for the preprocessed talc spectral data, the spectral characteristic peak with a wavelength of 396.254 nm is selected, and the peak intensity is arranged in descending order. The first 5 data are selected to form a data set, and the relative standard deviation of the data set is calculated. One new data is added in sequence to form a new data set and the relative standard deviation is calculated. The data set with the smallest relative standard deviation is selected and used as the spectral data in the stable stage.

9. The talc grade classification device according to claim 7, characterized in that: The classifier is a support vector machine.

10. The talc grade classification device according to claim 1, characterized in that: The automatic sorting module includes a sorting photoelectric sensor, a programmable logic controller and a sorting actuator, wherein: The sorting photoelectric sensor is used to generate a sorting trigger signal when the talc sample passes through and send it to the programmable logic controller; After receiving the sorting trigger signal, the programmable logic controller is used to calculate the time when the talc sample arrives at the sorting actuator according to the conveyor belt speed and the distance between the photoelectric sensor and the sorting actuator, and generate a sorting instruction based on the quality grade and send it to the sorting actuator; The sorting execution mechanism is used to complete the sorting action according to the sorting instruction.

Images