Trace thiabendazole detection method and system based on terahertz metamaterial resonance enhancement

Abstract

The present application provides a kind of trace thiabendazole detection method and system of terahertz metamaterial resonance enhancement, the method comprises: based on the characteristic peak frequency of single period structure and thiabendazole, the parameter information of the metamaterial sensor is constructed, and modeling simulation is carried out according to parameter information, to obtain the terahertz metamaterial with L type composite double-peak structure;Preparation thiabendazole sample, and utilize terahertz metamaterial to carry out spectrum collection to thiabendazole sample, to obtain corresponding spectral data;Support vector regression model is constructed, and the optimization algorithm of weighted vector average is integrated, and global optimization of hyperparameter is carried out, to construct corresponding optimization model;According to spectral data and optimization model, corresponding detection model is constructed, and nondestructive testing of trace thiabendazole is realized using detection model.The present application not only realizes high sensitive quantitative analysis to trace thiabendazole residue, also verifies the feasibility and superiority of INFO optimization algorithm in the field of spectral detection.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for detecting trace amounts of thiabendazole using terahertz metamaterial resonance enhancement. Background Technology

[0002] Thiabendazole (TBZ) is a broad-spectrum fungicide widely used in agriculture to control plant diseases caused by fungi, particularly in vegetables, fruits, and grains. It primarily inhibits pathogen growth by suppressing the formation of microtubules within fungal cells, thus interfering with cell division and proliferation. Common diseases it controls include downy mildew, gray mold, powdery mildew, and leaf spot, demonstrating significant disease control effects. Furthermore, thiabendazole is often used as a preservative to extend the shelf life of fruits and vegetables and prevent spoilage. However, the issue of thiabendazole residues in food has raised widespread concern, as long-term consumption of food containing these residues may pose health risks. Although its acute toxicity is low, long-term exposure may lead to a range of health problems, including carcinogenicity, endocrine disruption, and damage to organs such as the liver and kidneys. In particular, long-term consumption of food with excessive thiabendazole residues may pose potential risks to the immune system and reproductive health. Therefore, although international food safety standards impose strict limits on the use and residue levels of thiabendazole, continued monitoring of its use is necessary to reduce potential harm to humans and the environment. For example, the Chinese national standard GB 2763-2021, "Maximum Residue Limits for Pesticides in Food," stipulates that the maximum permissible residue concentration of thiamethoxam in apples is 3 μg / ml. In daily life, if thiamethoxam is used improperly or food is not thoroughly cleaned, its residue level is very likely to exceed the limit, thus affecting people's health and causing food safety issues. Therefore, exploring a rapid, accurate, and sensitive method for detecting thiamethoxam residue concentration is of great practical significance.

[0003] Currently, common methods for detecting pesticide residues include gas chromatography, high-performance liquid chromatography, enzyme-linked immunosorbent assay (ELISA), liquid chromatography-mass spectrometry (LC-MS), and hyperspectral imaging. These methods can effectively detect pesticide residues and have high sensitivity, capable of detecting pesticide residues at low concentrations, and can handle different types of samples, including agricultural products and environmental samples. However, these methods also have some limitations. For example, sample pretreatment is complex, and some methods are sensitive to matrix interference, which may affect the accuracy of detection. While some traditional methods are low-cost and simple, their sensitivity is relatively low, making them suitable only for preliminary screening of high-concentration samples. Therefore, there is an urgent need to research a detection technology that is both highly sensitive and rapid and non-destructive. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide a method and system for detecting trace amounts of thiabendazole using terahertz metamaterial resonance enhancement, so as to at least solve the shortcomings of the above-mentioned technology.

[0005] This invention proposes a method for detecting trace amounts of thiabendazole using terahertz metamaterial resonance enhancement, comprising:

[0006] The parameter information of the metamaterial sensor was constructed based on the single periodic structure and the characteristic peak frequency of thiabendazole, and modeling and simulation were performed based on the parameter information to obtain the terahertz metamaterial with an L-shaped composite double peak structure.

[0007] A thiabendazole sample was prepared, and the terahertz metamaterial was used to acquire the spectrum of the thiabendazole sample to obtain the corresponding spectral data.

[0008] A support vector regression model is constructed, and a weighted vector average optimization algorithm is incorporated to perform global hyperparameter optimization in order to construct the corresponding optimized model.

[0009] A corresponding detection model is constructed based on the spectral data and the optimized model, and the detection model is used to achieve non-destructive detection of trace amounts of thiabendazole.

[0010] Furthermore, the steps of constructing parameter information for the metamaterial sensor based on the single periodic structure and the characteristic peak frequency of thiabendazole, and performing modeling and simulation based on the parameter information to obtain the terahertz metamaterial with an L-shaped composite bimodal structure include:

[0011] Obtain the characteristic peak frequency of thiabendazole, and obtain the design peak frequency point of the metamaterial based on the characteristic peak frequency of thiabendazole;

[0012] The quality factor and sensitivity of the metamaterial are calculated by simulation spectrum, and the parameters of the preset database are scanned according to the designed peak frequency point, the quality factor and the sensitivity to obtain the geometric parameters of the metamaterial.

[0013] A single periodic structure is selected as the modeling object, and modeling and simulation are performed based on the geometric parameters to obtain a terahertz metamaterial with an L-shaped composite bimodal structure.

[0014] Furthermore, after the steps of preparing the thiabendazole sample and using the terahertz metamaterial to acquire the spectrum of the thiabendazole sample to obtain the corresponding spectral data, the method further includes:

[0015] The spectral data is subjected to differential detection using a preset statistical algorithm to obtain verified spectral data;

[0016] The verified spectral data is dimensionality reduced using a preset feature extraction algorithm to obtain dimensionality-reduced spectral data.

[0017] Furthermore, the steps involved in constructing a support vector regression model and incorporating a weighted vector average optimization algorithm to perform global hyperparameter optimization, thereby building the corresponding optimized model, include:

[0018] A support vector regression model is constructed, and a set of randomly initialized candidate solution vectors is used as the search starting point. The population parameters of the weighted vector average optimization algorithm are initialized, and each population parameter is combined to train the support vector regression model. Cross-validation is used to evaluate the prediction performance in order to achieve fitness evaluation.

[0019] Individuals are ranked according to their fitness, and their positions are iteratively updated by introducing an information vector mechanism, thereby selecting different combinations of parameters;

[0020] The parameter combination with the lowest fitness is selected as the optimal parameter combination, and the support vector regression model is output and trained to construct the corresponding optimized model.

[0021] This invention also proposes a terahertz metamaterial resonance-enhanced trace thiabendazole detection system, comprising:

[0022] The modeling and simulation module is used to construct the parameter information of the metamaterial sensor based on a single periodic structure and the characteristic peak frequency of thiabendazole, and to perform modeling and simulation based on the parameter information to obtain a terahertz metamaterial with an L-shaped composite double-peak structure.

[0023] A spectral acquisition module is used to prepare thiabendazole samples and to acquire the spectra of the thiabendazole samples using the terahertz metamaterial to obtain the corresponding spectral data.

[0024] The model optimization module is used to build a support vector regression model and incorporates a weighted vector average optimization algorithm to perform global hyperparameter optimization in order to build the corresponding optimized model.

[0025] The non-destructive testing module is used to construct a corresponding detection model based on the spectral data and the optimized model, and to use the detection model to perform non-destructive testing on trace amounts of thiabendazole.

[0026] Furthermore, the modeling and simulation module is specifically used for:

[0027] Obtain the characteristic peak frequency of thiabendazole, and obtain the design peak frequency point of the metamaterial based on the characteristic peak frequency of thiabendazole;

[0028] The quality factor and sensitivity of the metamaterial are calculated by simulation spectrum, and the parameters of the preset database are scanned according to the designed peak frequency point, the quality factor and the sensitivity to obtain the geometric parameters of the metamaterial.

[0029] A single periodic structure is selected as the modeling object, and modeling and simulation are performed based on the geometric parameters to obtain a terahertz metamaterial with an L-shaped composite bimodal structure.

[0030] Furthermore, the system also includes:

[0031] The differential detection module is used to perform differential detection on the spectral data using a preset statistical algorithm to obtain verified spectral data.

[0032] The data dimensionality reduction module is used to perform dimensionality reduction processing on the verified spectral data using a preset feature extraction algorithm to obtain dimensionality-reduced spectral data.

[0033] Furthermore, the model optimization module is specifically used for:

[0034] A support vector regression model is constructed, and a set of randomly initialized candidate solution vectors is used as the search starting point. The population parameters of the weighted vector average optimization algorithm are initialized, and each population parameter is combined to train the support vector regression model. Cross-validation is used to evaluate the prediction performance in order to achieve fitness evaluation.

[0035] Individuals are ranked according to their fitness, and their positions are iteratively updated by introducing an information vector mechanism, thereby selecting different combinations of parameters;

[0036] The parameter combination with the lowest fitness is selected as the optimal parameter combination, and the support vector regression model is output and trained to construct the corresponding optimized model.

[0037] The present invention also proposes a storage medium storing a computer program that, when executed by a processor, implements the above-described method for detecting trace thiabendazole using terahertz metamaterial resonance enhancement.

[0038] The present invention also proposes a computer, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described terahertz metamaterial resonance-enhanced trace thiabendazole detection method.

[0039] The present invention relates to a terahertz metamaterial resonance-enhanced method and system for detecting trace thiabendazole. A terahertz metamaterial sensor based on an "L"-shaped composite double-peak structure is designed and constructed to achieve dual resonance absorption enhancement. This sensor is then applied to the non-destructive detection of thiabendazole residues. Based on thiabendazole transmission spectral data acquired by the terahertz metamaterial sensing platform, a high-performance quantitative modeling method integrating the INFO optimization algorithm and support vector regression (SVR) is proposed. This method not only achieves highly sensitive quantitative analysis of trace thiabendazole residues but also verifies the feasibility and superiority of the INFO optimization algorithm in the field of spectral detection. Attached Figure Description

[0040] Figure 1 This is a flowchart of the terahertz metamaterial resonance-enhanced trace thiabendazole detection method in the first embodiment of the present invention.

[0041] Figure 2 This is a calculation diagram of the quality factor Q value parameter of the metamaterial in the first embodiment of the present invention;

[0042] Figure 3 This is a calculation diagram of the sensitivity S-parameter of the metamaterial in the first embodiment of the present invention;

[0043] Figure 4 This is a diagram showing the dimensional parameters of the metamaterial sensor in the first embodiment of the present invention;

[0044] Figure 5 This is a three-dimensional diagram of the metamaterial sensor in the first embodiment of the present invention;

[0045] Figure 6 This is a hotspot distribution diagram of the metamaterial sensor in the first embodiment of the present invention in the 1.79 THz frequency band;

[0046] Figure 7 This is a hotspot distribution diagram of the metamaterial sensor in the first embodiment of the present invention in the 2.30 THz frequency band;

[0047] Figure 8 The first embodiment of the present invention shows the feature variable selection process and results based on the IRIV algorithm, where (a) shows the trend of the number of IRIV feature variable subsets and (b) shows the distribution of feature variables.

[0048] Figure 9 This is the distribution of VIP values ​​corresponding to each spectral variable in the feature variable screening results based on the VIP algorithm in the first embodiment of the present invention;

[0049] Figure 10 This is the distribution of feature variables in the feature variable screening results based on the VIP algorithm in the first embodiment of the present invention;

[0050] Figure 11 The first embodiment of the present invention shows the feature variable screening process and results based on the VISSA algorithm, where (a) is the RMSECV comparison and (b) is the feature variable distribution.

[0051] Figure 12 This is a continuous response surface plot of the INFO-optimized SVR parameters in the first embodiment of the present invention;

[0052] Figure 13This is a fitting graph of the actual concentration values ​​and model predictions of the VISSA-INFOSVR model in the first embodiment of the present invention;

[0053] Figure 14 The evaluation results of the INFO-SVR model, SVR model, ACO model, and GA-SVR model in the first embodiment of the present invention are shown in the figure.

[0054] Figure 15 This is a structural block diagram of the terahertz metamaterial resonance-enhanced trace thiabendazole detection system in the second embodiment of the present invention;

[0055] Figure 16 This is a structural block diagram of the computer in the third embodiment of the present invention.

[0056] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0057] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0058] Unless otherwise defined, all technical and scientific terms used in this embodiment have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this embodiment is for the purpose of describing particular embodiments only and is not intended to limit the invention. The term "and / or" as used in this embodiment includes any and all combinations of one or more of the associated listed items.

[0059] Example 1

[0060] Please see Figure 1 The figure shows a method for detecting trace thiabendazole using terahertz metamaterial resonance enhancement according to the first embodiment of the present invention. The method specifically includes steps S101 to S104:

[0061] S101, the parameter information of the metamaterial sensor is constructed based on the single periodic structure and the characteristic peak frequency of thiabendazole, and modeling and simulation are performed according to the parameter information to obtain the terahertz metamaterial with an L-shaped composite double peak structure;

[0062] Furthermore, step S101 specifically includes steps S1011 to S1013:

[0063] S1011, Obtain the characteristic peak frequency of thiabendazole, and obtain the design peak frequency point of the metamaterial based on the characteristic peak frequency of thiabendazole;

[0064] S1012, calculate the quality factor and sensitivity of the metamaterial through simulation spectrum, and perform parameter scanning on the preset database according to the designed peak frequency point, the quality factor and the sensitivity to obtain the geometric parameters of the metamaterial;

[0065] S1013, Select a single periodic structure as the modeling object, and perform modeling and simulation based on the geometric parameters to obtain a terahertz metamaterial with an L-shaped composite bimodal structure.

[0066] This embodiment employs metamaterial sensor technology combined with terahertz spectroscopy to quantitatively detect trace amounts of thiabendazole residues at different concentrations. Specifically, it includes:

[0067] (1) Based on electromagnetic theory, an "L"-shaped composite metamaterial sensor is proposed;

[0068] (2) Conduct spectral response characteristics analysis of thiabendazole solutions of different concentrations; and use statistical methods to verify the significant differences in the spectra obtained from the experiment;

[0069] (3) A novel regression model based on the INFO optimization algorithm is proposed, and its performance and prediction accuracy are analyzed and compared with those of traditional optimization algorithm regression models to verify its effectiveness and superiority.

[0070] (4) The adaptability of different feature extraction methods in the regression model was analyzed, the improvement effect of spectral data of different quality on model performance was compared, and finally a new regression model with high accuracy and strong robustness was proposed.

[0071] (5) Use statistical methods to verify the statistical reliability of the model's predicted values.

[0072] In practice, the TAS7500 terahertz time-domain spectrometer (THz-TDS) manufactured by Advantest Corporation of Japan was used as the spectral acquisition device. This instrument utilizes femtosecond laser-excited photoconductive antenna technology to construct a terahertz pulse source. By controlling the time delay between the reference and probe beams using delay lines, it achieves precise acquisition of the time-domain terahertz signal and obtains the sample's frequency response information in the terahertz band through Fourier transform. The TAS7500 system features a wide frequency coverage (0.1–4.0 THz), high time resolution (<100 fs), and a good signal-to-noise ratio (dynamic range >70 dB). It supports measurements in both transmission and reflection modes and is suitable for non-destructive detection and quantitative analysis of various materials.

[0073] Specifically, thiabendazole has characteristic peak frequencies of 0.9, 1.26, 1.70, 1.93, and 2.40 THz in the terahertz band. To improve the sensitivity of thiabendazole detection by terahertz spectroscopy through resonance effect, the characteristic peak frequencies inherent in the material itself were selected as the peak frequencies for the metamaterial design (i.e., selected from 0.9, 1.26, 1.70, 1.93, and 2.40 THz).

[0074] Furthermore, in software simulations, if the frequencies of two characteristic peaks are close together, during actual measurement using an instrument after the metamaterial is fabricated, insufficient instrument resolution or power may cause the two peaks appearing in the simulation to merge into one. To avoid this phenomenon and ensure that the metamaterial can achieve the requirement of "improved sensitivity," characteristic peaks with a frequency difference greater than 0.5 THz are selected as a design premise. In this embodiment, the 1.70 and 2.40 THz peak positions of thiabendazole itself are selected for design.

[0075] In this embodiment, for the designed metamaterial structure, the Q-factor and sensitivity (S) of the metamaterial are calculated through simulation spectrum. Generally, the larger the Q-factor and S, the better the performance of the designed metamaterial. In this embodiment, these two indicators are also used to confirm the final structural parameters. In the field of metamaterials, the Q-factor (Quality Factor) is an important parameter for measuring resonance performance, defined as the resonance frequency (…). ) and resonance peak bandwidth ( The Q value reflects the selectivity of the structure for a specific frequency and its ability to store and dissipate energy in the system. A higher Q value indicates a sharper resonance peak, higher frequency resolution, and lower system energy loss. In terahertz metamaterial devices, a high Q value usually means a stronger local electric field enhancement effect and higher detection sensitivity, making it particularly suitable for high-precision sensing scenarios such as weak signal sensing and trace substance identification. Therefore, optimizing structural parameters to obtain a higher Q value is one of the key design goals for improving the performance of terahertz metamaterial sensors. The calculation formula is:

[0076] ;

[0077] Sensitivity (S) is a key indicator that measures a structure's ability to respond to changes in external physical or chemical parameters. It is often used to describe a structure's performance in sensing changes in refractive index, concentration, or thickness. Sensitivity is typically defined as the resonant frequency shift (S / S). ) and the physical quantity that caused the change (such as the change in refractive index, The ratio of 1 / 2 GHz to 1 RIU is expressed in units of GHz / RIU. Higher sensitivity indicates a more significant response of the structure to small perturbations. The calculation formula is:

[0078]

[0079] Furthermore, after determining the design characteristic peaks at 1.70 and 2.40 THz, the structure was defined as an "L"-shaped structure through parameter scanning based on experience and databases. Further, more detailed and refined parameter scanning was conducted on the existing basis, including the calculation of the Q value as follows... Figure 2 As shown. f 1. f The FWHM values ​​of 2 are 50 GHz and 242 GHz, respectively. According to the formula for calculating the Q value, the Q values ​​are 35.5 and 9.5.

[0080] The calculation of sensitivity S is as follows: Figure 3 As shown in the figure. It can be seen that when the refractive index increases from 1.0 to 2.5, f 1. f 2. The frequencies were shifted by 53.4 GHz and 240 GHz, respectively. The calculated sensitivities were 35.6 GHz / RIU and 160 GHz / RIU, respectively.

[0081] Based on the above analysis, the structural parameters of the metamaterial are determined as follows: periodicity. P 1:70 μm; rectangular structure edge length L 1:40μm; Long side length of L-shaped structure L 2:28 μm; short side length of L-shaped structure L 3:18 μm; Spacing between the long sides of different L-shaped structures L 4:14 μm; Spacing between the short sides of different L-shaped structures L 5:4 μm; the width of the rectangular structure edge W 1:4 μm; L-shaped structure width W At a thickness of 2:2 μm, the Q value and S of the metamaterial are optimal (see [reference]). Figure 4 The designed metamaterial sensor was structurally modeled using Lumerical FDTD Solutions software. Considering that metamaterials are typically composed of periodically repeating structural units, their overall electromagnetic response characteristics can be approximated by simulating the unit structures, thus significantly improving simulation efficiency and computational accuracy. Therefore, this embodiment selects a single periodic structure as the modeling object for simulation analysis. Please refer to... Figure 5 The image shown is a three-dimensional diagram of the metamaterial sensor structure. The substrate material is silicon with a refractive index of 3.335, and the surface metal layer is gold, which is used to enhance its resonant response characteristics in the terahertz band.

[0082] After the metamaterial sensor design was completed, FDTD Solutions was used to perform full-band terahertz simulation of the "L"-shaped composite bimodal terahertz metamaterial sensor. Please refer to [link / reference]. Figures 6 to 7 The figure shows the spatial distribution of the electric field intensity in the near-field region of an "L"-shaped composite bimodal metamaterial sensor simulated based on the FDTD algorithm. In the 1.79 THz frequency band (e.g., ... Figure 6 As shown in the diagram, the electric field energy is mainly concentrated on the left and right sides of the outer side of the structure, forming two symmetrical high-intensity hotspot regions, constituting the main "hotspot" distribution characteristics. Simultaneously, a certain intensity of electric field concentration can also be observed in the central region of the structure, causing the overall electric field intensity distribution to exhibit an approximately "quadripolar" pattern. The electric field intensity decreases radially from the inside out, with a clear color gradient, reflecting the local enhancement characteristics of the electric field. This distribution pattern indicates that the structure has a selective enhancement capability for terahertz waves with specific polarization directions in this frequency band and possesses a strong spatial field concentration effect.

[0083] Furthermore, when the frequency increases to 2.30 THz (e.g. Figure 7 As shown in the figure, the electric field distribution pattern is reconstructed, with hotspots concentrated on both sides of the structure's center, exhibiting a relatively symmetrical "dumbbell-shaped" enhancement feature. At this point, the hotspot area is significantly smaller than at 1.79 THz, but the local electric field concentration is significantly increased, resulting in more vivid color contrast and reflecting a stronger electric field intensity gradient. This frequency-dependent localization of the electric field response reflects the structure's frequency-domain dispersion characteristics and multi-frequency resonance capability. This frequency-dependent distribution characteristic indicates that the sensor has switchable field enhancement modes at different operating frequencies.

[0084] S102, prepare a thiabendazole sample, and use the terahertz metamaterial to collect the spectrum of the thiabendazole sample to obtain the corresponding spectral data;

[0085] In the specific implementation, thiabendazole standard solution (initial concentration 100 μg / ml) provided by the Aladdin Reagent Platform was selected as the reference substance. To construct a trace detection system, 21 concentration gradient samples were prepared using a stepwise dilution method according to the requirements of the national standard GB 2763-2021. Ultrapure deionized water was used to quantitatively dilute the stock solution using a high-precision pipette. Each sample was vortexed at 3000 rpm for 3 minutes to ensure uniform distribution of solute molecules in the solvent. The final series of concentration solutions were transferred to brown volumetric flasks and sealed for storage. Specific concentration distribution parameters are detailed in Appendix Table 1. This standardized preparation process minimizes experimental errors and provides a reliable sample basis for subsequent spectroscopic detection.

[0086] Table 1. Concentration gradient of thiabendazole solution

[0087]

[0088] Spectroscopic acquisition was performed on trace amounts of thiabendazole solution residue. The experiment required thorough preheating using an air compressor equipped with an air drying module to eliminate environmental interference. Before detection, dry air was continuously circulated into the spectrometer's optical cavity, and closed-loop control was implemented using temperature and humidity sensors to ensure that the relative humidity within the cavity was ≤ 10% and the temperature remained stable within the range of 25±0.5℃. A concentration-increasing injection strategy was employed for sample processing. Thiabendazole solutions of different concentration gradients were quantitatively added at 20 μl / time using a microsyringe and sequentially dropped onto the surface of the metamaterial sensing chip. Given the strong absorption characteristics of liquid water for THz waves, the sample-loaded sensing chip was dried in a 50℃ vacuum drying oven for 30 minutes to ensure that solute molecules formed a uniform thin film on the chip surface. The dried metamaterial was then installed in the transmission detection module of the THz-TDS system, and allowed to stand for 2 minutes after each injection to allow the system to reach thermal equilibrium. To improve data reliability, a composite sampling mode of 5 points × 10 repeated measurements was adopted, and a total of 1050 raw spectral data were obtained from 21 groups of concentration samples.

[0089] In some alternative embodiments, after step S102, the method further includes:

[0090] The spectral data is subjected to differential detection using a preset statistical algorithm to obtain verified spectral data;

[0091] The verified spectral data is dimensionality reduced using a preset feature extraction algorithm to obtain dimensionality-reduced spectral data.

[0092] In practice, the F-test, t-test, and P-value algorithms are used to detect differences in spectral data. The F-test is a statistical method used to compare whether there are significant differences between the variances of two or more populations. It is commonly used to test the homogeneity of sample variances and the significance analysis of multiple group means. Its applications are wide-ranging, such as testing the overall significance of a model in regression analysis or judging treatment differences among multiple groups. When used to test whether the variances of two populations are equal, the formula for its statistic is:

[0093] ;

[0094] in and The variance of two independent samples (usually the larger variance is taken as the numerator);

[0095] Furthermore, the t-test is a parametric test used to examine whether the difference between the means of two samples is significant. Depending on the data type, the t-test includes one-sample t-test, paired-sample t-test, and independent-sample t-test. In this study, the independent-sample t-test is primarily used to compare whether the difference between the predicted and true values ​​of the target analyte concentration is significant. The formula for calculating the independent-sample t-test is:

[0096]

[0097] in, and The mean of the two samples is denoted as . and For variance, and The sample size is represented by the t-test. This provides reliable statistical support for spectral modeling and analysis, ensuring high reliability and stability of the quantitative results.

[0098] Specifically, the p-value is a statistic used in hypothesis testing to measure the significance of observed results. It is defined as the probability of observing the current sample statistic value or a more extreme result given that the null hypothesis is true. A smaller p-value indicates that the observed result is less likely to occur under the null hypothesis, thus supporting the rejection of the null hypothesis. Generally, a p-value less than the significance level (e.g., 0.05) indicates a significant difference in the data; a p-value greater than the significance level (e.g., 0.05) indicates no significant difference in the data. The formulas for calculating p-value in F-test and t-test are as follows:

[0099] ;

[0100] ;

[0101] in, and For the sample size, , and Let be the degrees of freedom of the sample, and F be a random variable following an F-distribution. It is a random variable T greater than t statistic The probability, The random variable T represents a variable with degrees of freedom of 1 The t-distribution, These are the observed values ​​of the calculated F-statistic.

[0102] Furthermore, the validated spectral data were dimensionality-reduced using the IRIV, VIP, and VISSA algorithms, respectively. The IRIV algorithm, an iterative variable selection method based on a wrapper strategy, is widely used for optimizing the selection of high-dimensional features in quantitative spectral detection. This algorithm categorizes original variables into strong-information, weak-information, non-informational, and interference variables by constructing sub-models and analyzing variable importance. During iteration, it retains only features that positively contribute to model performance. Variable importance is typically determined by comparing the model performance before and after removing a variable, for example, using the relative cross-validation error rate of change.

[0103] ;

[0104] in, To remove the model error after eliminating the j-th variable, This represents the model error, which includes all variables.

[0105] Please see Figure 8 The diagram illustrates the feature variable selection process and results based on the IRIV algorithm. To ensure the stability and representativeness of the selection results, this embodiment sets the iteration termination condition as the variable subset tends to stabilize and the error no longer decreases significantly. During multiple iterations, IRIV continuously eliminates irrelevant variables that contribute little to the error, ultimately selecting 24 informative variables from the original spectral data, accounting for 6.09% of the total variables, greatly reducing the dimensionality of the data. Figure 8 (a) shows the trend of the number of variables in each iteration. Figure 8 (b) in the diagram shows the distribution of variables after variable selection.

[0106] Furthermore, the VIP algorithm is a feature variable evaluation method based on partial least squares regression (PLSR), used to measure the contribution of each independent variable to the response variable during modeling. It is widely used in feature band selection in quantitative spectral analysis. The VIP value is calculated by comprehensively considering the projected weights of each variable on each latent variable and its contribution to the explained variance. The calculation formula is as follows:

[0107] ;

[0108] in, The number of independent variables. The number of latent variables extracted. This represents the relationship between the a-th latent variable and the response variable. Explanation of variance For the total variance, For variables The weights on the a-th latent variable;

[0109] Please see Figures 9 to 10 The figure shows the feature variable selection process and results based on the VIP algorithm. The VIP value measures the contribution of each variable to the PLS model by assessing its projection importance. Variables with a VIP value greater than 1 are generally considered to have higher information value. In this study, a threshold of 1 was set, the number of PLS ​​principal components was 10, and 141 variables with significant VIP values ​​were selected as effective feature inputs. Figure 9 The distribution of VIP values ​​corresponding to each spectral variable is shown; Figure 10 The distribution of variables after variable selection is shown. The results indicate that variables with high VIP values ​​are mainly concentrated in the frequency bands of 1.04 ~ 1.19 THz, 1.38 ~ 1.85 THz, and 2.77 ~ 3.20 THz, suggesting that the spectral characteristics in this region are highly correlated with changes in thiabendazole concentration.

[0110] The VISSA algorithm is a feature selection algorithm based on the idea of ​​progressively compressing the variable space. The basic steps of VISSA are: randomly selecting multiple subsets of variables from the initial variable set, constructing prediction models for each subset, obtaining their root mean square error (RMSECV) through cross-validation, and selecting the subset with the better performance as the candidate variable set for the next iteration, continuously narrowing the variable space until the variable space stabilizes or reaches a predetermined minimum number of variables. Its goal is to minimize the prediction error.

[0111] ;

[0112] in, It is the first i The predicted value for each sample, This is the actual value. This represents the number of samples.

[0113] Please see Figure 11 The diagram shows the feature variable selection process and results based on the VISSA algorithm. VISSA iteratively compresses the variable space, gradually eliminating redundant features that contribute little to modeling, thereby improving the model's robustness and generalization ability. In this embodiment, the initial variable subset is set as full-band spectral data. Under the set iteration step size and selection strategy, 120 significant variables were ultimately retained. Figure 11 (a) shows the comparison of RMSECV of the model before and after variable selection using the VISSA algorithm; Figure 11Figure (b) shows the distribution characteristics of the variables after VISSA filtering on the frequency axis. The significant decrease in RMSECV indicates that the VISSA algorithm effectively improves the model's predictive performance in cross-validation after removing redundant variables, enhances the model's generalization ability and stability, and has the potential to suppress overfitting. The RMSECV is minimum at 0.26987 when the number of variables is 120.

[0114] S103, construct a support vector regression model and incorporate a weighted vector average optimization algorithm to perform global hyperparameter optimization in order to construct the corresponding optimized model;

[0115] Furthermore, step S103 specifically includes steps S1031 to S1033:

[0116] S1031, Construct a support vector regression model, and use a set of randomly initialized candidate solution vectors as the search starting point. Initialize the population parameters of the weighted vector average optimization algorithm, combine each population parameter to train the support vector regression model, and use cross-validation to evaluate the prediction performance in order to achieve fitness evaluation.

[0117] S1032, sort the individuals according to the fitness, and iteratively update the individual positions by introducing an information vector mechanism, thereby selecting different combinations of parameters;

[0118] S1033, Select the parameter combination with the minimum fitness as the optimal parameter combination, output and train the support vector regression model to construct the corresponding optimized model.

[0119] In practical implementation, a Support Vector Regression (SVR) model is constructed, which is an extension of the Support Vector Machine (SVM) in regression problems. Its core idea is to tolerate errors. Within a certain range, find the optimal regression function (penalty coefficient C and kernel function parameters). This enables the model to generalize well to the training data. In quantitative detection, SVR can be used to construct a predictive model between the concentration of a target component and its corresponding spectral features, which has significant advantages, especially when the number of samples is limited and the dimensionality of variables is high.

[0120] Specifically, a weighted vector average optimization algorithm (in this embodiment, the INFO algorithm is selected) is incorporated into the support vector regression model. The INFO algorithm searches for the optimal combination of SVR hyperparameters [C,γ] in a predefined search space to minimize the target regression error (such as RMSE), thereby obtaining the SVR regression model with the best accuracy.

[0121] Its basic logic consists of four steps:

[0122] Population initialization: Let the population size be N, and each individual be... Initialized as:

[0123] ;

[0124] in , For the lower / upper bound of the variable, It is a uniform random vector.

[0125] Using a set of randomly initialized candidate solution vectors (i.e., multiple combinations of [C,γ]) as the starting point for the search, individuals are randomly selected from the current population. , , The vector is constructed using the following weighted mean formula:

[0126] ;

[0127] in, .

[0128] Vector-based guided update: Generate a search vector by combining the current individual with the optimal individual.

[0129] ;

[0130] in, A random individual in the population. Let r be the search vector. , This is the optimal individual position vector in the population;

[0131] Direction vector perturbation update: Further enhances local search capabilities by adjusting and updating individual positions through perturbations that guide the direction.

[0132] ;

[0133] in, This is the vector obtained through "direction vector perturbation update". A random individual in the population. It is a uniform random vector.

[0134] Fitness evaluation: Each parameter combination is used to train the SVR model, and its predictive performance is evaluated using cross-validation. RMSECV is commonly used as the fitness metric to evaluate whether the current parameter combination is optimal; generally, a smaller RMSECV is better. For all candidate vectors... , , With the current individual Perform a fitness comparison and select the optimal one to update the current solution:

[0135] ;

[0136] The objective function (fitness function) is defined as the RMSE or cross-validation error of the SVR model on the training set:

[0137] ;

[0138] Based on this objective function, INFO continuously optimizes the parameters [C, To minimize RMSECV:

[0139] .

[0140] Information vector sorting and updating: INFO sorts individuals based on fitness and iteratively updates the position of individuals by introducing an information vector mechanism (referencing the best individual, local individuals, group trends, etc.) to select different combinations of parameters and prevent getting trapped in local optima.

[0141] Output the optimal solution and train the final model: After the search is complete, select the parameter combination with the smallest RMSECV as the optimal parameter combination, that is, to achieve the maximum number of iterations or meet the error accuracy, and output the optimal individual. The SVR model is output and trained to form the final INFO-SVR prediction model after intelligent optimization and parameter tuning.

[0142] To allow for a more intuitive view of the parameter optimization process during data processing in the INFO-SVR model, the optimization process is visualized. Specifically, the parameters obtained through INFO-SVR optimization are C=7.9818 and γ=0.0197. The specific parameters obtained are as follows... Figure 12 As shown ( Figure 12 This is a logarithmic graph, meaning that the C and γ axes are actually the calculated values ​​of log10 C and log10 γ.

[0143] according to Figure 12 As can be seen, the three-dimensional graph consists of different combinations of C and γ and their corresponding RMSECV values. The actual process of parameter optimization is to select different combinations of C and γ that minimize the RMSECV of SVR; this combination of C and γ is considered the optimal parameter by INFO.

[0144] Please see Figure 13 The figure shows the fitting plot of the actual concentration values ​​and model predictions of the INFO-SVR model after VISSA processing.

[0145] S104, construct a corresponding detection model based on the spectral data and the optimized model, and use the detection model to achieve non-destructive detection of trace amounts of thiabendazole.

[0146] In practical implementation, a corresponding detection model is constructed using spectral data and an optimized model. This model is then used to achieve non-destructive detection of trace amounts of thiabendazole. To verify the effectiveness and superiority of the constructed INFO-SVR model in the quantitative detection of thiabendazole using terahertz spectroscopy, this embodiment systematically compares and analyzes it with the traditional Support Vector Regression (SVR) model and SVR models optimized using Ant Colony Optimization (ACO) and Genetic Algorithm (GA). All models are trained and predicted on the same training and test sets, using the coefficient of determination (R²). 2 Evaluation metrics such as root mean square error (RMSE) and mean absolute error (MAE) were used to comprehensively assess the model's performance in terms of fitting accuracy and generalization ability, ensuring the objectivity and fairness of the comparison results. The model results are shown in Table 2. Figure 14 As shown.

[0147] Table 2 Evaluation of the Prediction Performance of Each Model

[0148]

[0149] Experimental results show that the INFO-SVR model exhibits superior modeling performance across all evaluation metrics. Compared to the original SVR model without optimization algorithms, INFO-SVR significantly reduces RMSE on the test set, and R... 2 The values ​​are significantly improved, demonstrating superior prediction accuracy and stability. Compared to the commonly used ACO-SVR and GA-SVR models, INFO-SVR still exhibits superior performance, indicating that the INFO algorithm has higher search efficiency and stronger global optimization capabilities in optimizing SVR hyperparameters.

[0150] In this embodiment, conditional judgments such as rand < 0.5 are frequently used in the INFO algorithm to achieve probabilistic selection among multiple candidate update strategies. By randomly switching individuals between different strategies, INFO achieves a dynamic balance between global search and local exploitation, simulating probabilistic behavioral decision-making processes in nature. This makes the optimization path more flexible and adaptable, effectively reducing the risk of getting trapped in local optima. The INFO algorithm constructs update strategies with vector mean as the core, and through weighted guidance and elite retention mechanisms, it effectively enhances the convergence direction and global exploration capability of the search process. Especially when dealing with complex datasets such as terahertz spectra, which are high-dimensional, nonlinear, and have redundant features, the INFO algorithm exhibits stronger resistance to trapping and robustness, and can more accurately mine potential feature-response relationships. INFO itself has a simple structure and few parameters, avoiding the parameter tuning costs and generalization risks in multi-parameter models, and providing more efficient and stable optimization support for SVR models.

[0151] Based on the above analysis, the INFO-SVR model demonstrates superior performance compared to the other three models, fully validating the feasibility and advantages of applying the INFO algorithm to SVR modeling, and proving the potential and practical value of this method in quantitative analysis of terahertz spectroscopy. Its excellent modeling capabilities not only help improve the accuracy of pesticide residue detection but also provide new ideas for optimizing terahertz spectral data processing methods.

[0152] In summary, the terahertz metamaterial resonance-enhanced trace thiabendazole detection method described in the above embodiments of the present invention designs and constructs a terahertz metamaterial sensor based on an "L"-shaped composite double-peak structure to achieve dual resonance absorption enhancement. This sensor is then applied to the non-destructive detection of thiabendazole residues. Based on the thiabendazole transmission spectrum data acquired by the terahertz metamaterial sensing platform, a high-performance quantitative modeling method integrating the INFO optimization algorithm and support vector regression (SVR) is proposed. This not only achieves highly sensitive quantitative analysis of trace thiabendazole residues but also verifies the feasibility and superiority of the INFO optimization algorithm in the field of spectral detection.

[0153] Example 2

[0154] Another aspect of this invention proposes a terahertz metamaterial resonance-enhanced trace thiabendazole detection system, please refer to [link / reference needed]. Figure 15 The figure shows a terahertz metamaterial resonance-enhanced trace thiabendazole detection system according to a second embodiment of the present invention. The system includes:

[0155] The modeling and simulation module 11 is used to construct the parameter information of the metamaterial sensor based on the single periodic structure and the characteristic peak frequency of thiabendazole, and to perform modeling and simulation based on the parameter information to obtain the terahertz metamaterial with an L-shaped composite double peak structure.

[0156] Furthermore, the modeling and simulation module 11 is specifically used for:

[0157] Obtain the characteristic peak frequency of thiabendazole, and obtain the design peak frequency point of the metamaterial based on the characteristic peak frequency of thiabendazole;

[0158] The quality factor and sensitivity of the metamaterial are calculated by simulation spectrum, and the parameters of the preset database are scanned according to the designed peak frequency point, the quality factor and the sensitivity to obtain the geometric parameters of the metamaterial.

[0159] A single periodic structure is selected as the modeling object, and modeling and simulation are performed based on the geometric parameters to obtain a terahertz metamaterial with an L-shaped composite bimodal structure.

[0160] The spectral acquisition module 12 is used to prepare thiabendazole samples and to acquire the spectra of the thiabendazole samples using the terahertz metamaterial to obtain the corresponding spectral data.

[0161] Model optimization module 13 is used to construct a support vector regression model and incorporate a weighted vector average optimization algorithm to perform global hyperparameter optimization in order to construct the corresponding optimized model.

[0162] Furthermore, the model optimization module 13 is specifically used for:

[0163] A support vector regression model is constructed, and a set of randomly initialized candidate solution vectors is used as the search starting point. The population parameters of the weighted vector average optimization algorithm are initialized, and each population parameter is combined to train the support vector regression model. Cross-validation is used to evaluate the prediction performance in order to achieve fitness evaluation.

[0164] Individuals are ranked according to their fitness, and their positions are iteratively updated by introducing an information vector mechanism, thereby selecting different combinations of parameters;

[0165] The parameter combination with the lowest fitness is selected as the optimal parameter combination, and the support vector regression model is output and trained to construct the corresponding optimized model.

[0166] The non-destructive testing module 14 is used to construct a corresponding detection model based on the spectral data and the optimized model, and to use the detection model to achieve non-destructive testing of trace amounts of thiabendazole.

[0167] Furthermore, the system also includes:

[0168] The differential detection module is used to perform differential detection on the spectral data using a preset statistical algorithm to obtain verified spectral data.

[0169] The data dimensionality reduction module is used to perform dimensionality reduction processing on the verified spectral data using a preset feature extraction algorithm to obtain dimensionality-reduced spectral data.

[0170] The functions or operation steps implemented by the above modules and units are largely the same as those in the above method embodiments, and will not be repeated here.

[0171] The terahertz metamaterial resonance-enhanced trace thiabendazole detection system provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0172] Example 3

[0173] This invention also proposes a computer, please refer to [link / reference]. Figure 16 The computer shown in the third embodiment of the present invention includes a memory 10, a processor 20, and a computer program 30 stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, it implements the above-described terahertz metamaterial resonance-enhanced trace thiabendazole detection method.

[0174] The memory 10 includes at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 10 can be an internal storage unit of a computer, such as the computer's hard disk. In other embodiments, the memory 10 can be an external storage device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Furthermore, the memory 10 can include both internal and external storage units of the computer. The memory 10 can be used not only to store application software and various types of data installed on the computer, but also to temporarily store data that has been output or will be output.

[0175] In some embodiments, the processor 20 may be an electronic control unit (ECU), a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip, used to run program code stored in the memory 10 or process data, such as executing access restriction programs.

[0176] It should be pointed out that, Figure 16The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0177] This invention also proposes a storage medium storing a computer program that, when executed by a processor, implements the terahertz metamaterial resonance-enhanced trace thiabendazole detection method described above.

[0178] Those skilled in the art will understand that the logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0179] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0180] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0181] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0182] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for detecting trace amounts of thiabendazole using terahertz metamaterial resonance enhancement, characterized in that, include: The parameter information of the metamaterial sensor was constructed based on the single periodic structure and the characteristic peak frequency of thiabendazole, and modeling and simulation were performed based on the parameter information to obtain the terahertz metamaterial with an L-shaped composite double peak structure. A thiabendazole sample was prepared, and the terahertz metamaterial was used to acquire the spectrum of the thiabendazole sample to obtain the corresponding spectral data. A support vector regression model is constructed, and a weighted vector average optimization algorithm is incorporated to perform global hyperparameter optimization in order to construct the corresponding optimized model. A corresponding detection model is constructed based on the spectral data and the optimized model, and the detection model is used to achieve non-destructive detection of trace amounts of thiabendazole. The steps involved in constructing parameter information for the metamaterial sensor based on a single periodic structure and the characteristic peak frequency of thiabendazole, and then performing modeling and simulation based on this parameter information to obtain a terahertz metamaterial with an L-shaped composite bimodal structure, include: The characteristic peak frequency of thiabendazole is obtained, and the design peak frequency points of the metamaterial are obtained based on the characteristic peak frequency of thiabendazole, wherein the peak frequency points are 0.9, 1.26, 1.70, 1.93, and 2.40 THz, respectively. The quality factor and sensitivity of the metamaterial are calculated by simulation spectrum, and the parameters of the preset database are scanned according to the designed peak frequency point, the quality factor and the sensitivity to obtain the geometric parameters of the metamaterial. A single periodic structure is selected as the modeling object, and modeling and simulation are performed based on the geometric parameters to obtain a terahertz metamaterial with an L-shaped composite bimodal structure, wherein the structural parameters of the terahertz metamaterial are periodic. P 1 is 70 μm; the rectangular edge of the structure is long. L 1 is 40μm; the long side of the L-shaped structure is 40μm. L 2 is 28μm; the short side length of the L-shaped structure is... L 3 is 18μm; the spacing between the long sides of different L-shaped structures L 4 is 14μm; the spacing between the short sides of different L-shaped structures L 5 is 4μm; the rectangular edge of the structure is wide W 1 is 4μm; L-shaped structure width W 2 represents 2μm.

2. The method for detecting trace amounts of thiabendazole using terahertz metamaterial resonance enhancement according to claim 1, characterized in that, After the steps of preparing a thiabendazole sample and using the terahertz metamaterial to acquire the spectrum of the thiabendazole sample to obtain the corresponding spectral data, the method further includes: The spectral data is subjected to differential detection using a preset statistical algorithm to obtain verified spectral data; The verified spectral data is dimensionality reduced using a preset feature extraction algorithm to obtain dimensionality-reduced spectral data.

3. The method for detecting trace amounts of thiabendazole using terahertz metamaterial resonance enhancement according to claim 1, characterized in that, The steps to construct a support vector regression model, incorporate a weighted vector average optimization algorithm, and perform global hyperparameter optimization to build the corresponding optimized model include: A support vector regression model is constructed, and a set of randomly initialized candidate solution vectors is used as the search starting point. The population parameters of the weighted vector average optimization algorithm are initialized, and each population parameter is combined to train the support vector regression model. Cross-validation is used to evaluate the prediction performance in order to achieve fitness evaluation. Individuals are ranked according to their fitness, and their positions are iteratively updated by introducing an information vector mechanism, thereby selecting different combinations of parameters; The parameter combination with the lowest fitness is selected as the optimal parameter combination, and the support vector regression model is output and trained to construct the corresponding optimized model.

4. A terahertz metamaterial resonance-enhanced trace thiabendazole detection system, characterized in that, include: The modeling and simulation module is used to construct the parameter information of the metamaterial sensor based on a single periodic structure and the characteristic peak frequency of thiabendazole, and to perform modeling and simulation based on the parameter information to obtain a terahertz metamaterial with an L-shaped composite double-peak structure. A spectral acquisition module is used to prepare thiabendazole samples and to acquire the spectra of the thiabendazole samples using the terahertz metamaterial to obtain the corresponding spectral data. The model optimization module is used to build a support vector regression model and incorporates a weighted vector average optimization algorithm to perform global hyperparameter optimization in order to build the corresponding optimized model. The non-destructive testing module is used to construct a corresponding detection model based on the spectral data and the optimized model, and to use the detection model to achieve non-destructive testing of trace amounts of thiabendazole. Specifically, the modeling and simulation module is used for: The characteristic peak frequency of thiabendazole is obtained, and the design peak frequency points of the metamaterial are obtained based on the characteristic peak frequency of thiabendazole, wherein the peak frequency points are 0.9, 1.26, 1.70, 1.93, and 2.40 THz, respectively. The quality factor and sensitivity of the metamaterial are calculated by simulation spectrum, and the parameters of the preset database are scanned according to the designed peak frequency point, the quality factor and the sensitivity to obtain the geometric parameters of the metamaterial. A single periodic structure is selected as the modeling object, and modeling and simulation are performed based on the geometric parameters to obtain a terahertz metamaterial with an L-shaped composite bimodal structure, wherein the structural parameters of the terahertz metamaterial are periodic. P 1 is 70 μm; the rectangular edge of the structure is long. L 1 is 40μm; the long side of the L-shaped structure is 40μm. L 2 is 28μm; the short side length of the L-shaped structure is... L 3 is 18μm; the spacing between the long sides of different L-shaped structures L 4 is 14μm; the spacing between the short sides of different L-shaped structures L 5 is 4μm; the rectangular edge of the structure is wide W 1 is 4μm; L-shaped structure width W 2 represents 2μm.

5. The terahertz metamaterial resonance-enhanced trace thiabendazole detection system according to claim 4, characterized in that, The system also includes: The differential detection module is used to perform differential detection on the spectral data using a preset statistical algorithm to obtain verified spectral data. The data dimensionality reduction module is used to perform dimensionality reduction processing on the verified spectral data using a preset feature extraction algorithm to obtain dimensionality-reduced spectral data.

6. The terahertz metamaterial resonance-enhanced trace thiabendazole detection system according to claim 4, characterized in that, The model optimization module is specifically used for: A support vector regression model is constructed, and a set of randomly initialized candidate solution vectors is used as the search starting point. The population parameters of the weighted vector average optimization algorithm are initialized, and each population parameter is combined to train the support vector regression model. Cross-validation is used to evaluate the prediction performance in order to achieve fitness evaluation. Individuals are ranked according to their fitness, and their positions are iteratively updated by introducing an information vector mechanism, thereby selecting different combinations of parameters; The parameter combination with the lowest fitness is selected as the optimal parameter combination, and the support vector regression model is output and trained to construct the corresponding optimized model.

7. A storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the terahertz metamaterial resonance-enhanced method for detecting trace thiabendazole as described in any one of claims 1 to 3.

8. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the terahertz metamaterial resonance-enhanced trace thiabendazole detection method as described in any one of claims 1 to 3.

Images