GRU regression model and method for realizing intelligent detection of deltamethrin pesticide residues by using same

Abstract

The invention provides a GRU regression model and a method for realizing intelligent detection of deltamethrin pesticide residues by using the model, and belongs to the technical field of pesticide residue detection. According to the method, key information variables in a time sequence spectrum are automatically extracted by adopting a GRU algorithm, deviation caused by artificial feature selection in traditional modeling is avoided, spectrum acquisition is performed on deltamethrin in food by adopting an SERS sensor with a TiO2 / g-C3N4 substrate in combination with a GRU regression model, and a rapid intelligent deltamethrin detection system is constructed in combination with a deep learning GRU algorithm, so that the detection accuracy of deltamethrin in food is improved. According to the method, the Raman signal is subjected to feature extraction and modeling prediction, so that the deltamethrin residual quantity is intelligently identified and quantitatively analyzed, and the method is simple to operate and wide in application prospect.

Description

Technical Field

[0001] The present invention belongs to the technical field of pesticide residue detection, and in particular relates to a GRU regression model and a method for realizing intelligent detection of deltamethrin pesticide residues by using the model. Background Art

[0002] Tea is one of my country's most representative traditional agricultural products, beloved by consumers for its rich nutritional value and health benefits. During tea cultivation, pesticides are often sprayed to control pests and diseases, improve yield, and enhance appearance quality. However, improper pesticide use or incomplete cleaning of residues can lead to a certain level of pesticide residue in tea leaves. Long-term consumption of tea products containing pesticide residues may pose potential risks to human health, particularly chronic diseases such as the liver, spleen, and stomach. During tea cultivation, to control common pests such as aphids, spider mites, and green stink bugs, growers often use highly effective pyrethroid pesticides such as deltamethrin. While these pesticides are effective in controlling pests and ensuring tea yields, they are highly fat-soluble and metabolized slowly, making them prone to residues in finished tea products. Long-term consumption can damage organs such as the nervous system and liver.

[0003] To protect consumer health, current standards clearly stipulate the maximum residue limit of deltamethrin in tea, which is 10 mg / kg. Therefore, accurate detection of deltamethrin residues in tea has become a key factor in ensuring tea quality and safety. Currently, mainstream methods for pesticide residue detection include high-performance liquid chromatography, gas chromatography-mass spectrometry, high-performance liquid chromatography-mass spectrometry, and enzyme-linked immunosorbent assay (ELISA). These methods have good accuracy and stability, but they also have obvious shortcomings: 1) The detection process is complex, requiring professional technicians and precision instruments; 2) The detection cycle is long, making it difficult to meet the needs of rapid on-site testing; 3) The detection cost is high, which is not conducive to large-scale testing. Summary of the Invention

[0004] To solve the above technical problems, the present invention proposes a GRU regression model and a method for intelligent detection of cypermethrin pesticide residues using the model.

[0005] To achieve the above objectives, the present invention provides the following technical solutions:

[0006] One of the technical solutions of the present invention:

[0007] A gated recurrent unit (GRU) regression model for detecting deltamethrin comprises an input layer, a hidden layer, a fully connected layer, and an output layer; the input layer is used to receive deltamethrin SERS spectral data preprocessed by standard normal transformation (SNV); the hidden layer has 1-2 layers, 8-16 neurons, and an activation function is a ReLU function; the fully connected layer has 1 layer, is used to map features extracted by the hidden layer to the output layer, and the output layer outputs one feature to characterize the deltamethrin content.

[0008] Furthermore, the method for constructing the GRU regression model for detecting deltamethrin includes the following steps:

[0009] A deltamethrin standard solution was prepared, and a titanium dioxide-graphite carbon nitride SERS-enhanced substrate solution was mixed with the deltamethrin standard solution. The surface-enhanced Raman spectrum (SERS) of the mixture was collected and preprocessed with SNV to remove background interference and random noise in the spectrum, thereby obtaining a preprocessed SERS spectrum with a high signal-to-noise ratio. The preprocessed SERS spectrum was divided into a training set and a test set to obtain a GRU regression model for detecting deltamethrin.

[0010] Furthermore, the concentration of the deltamethrin standard solution is 0.01-1000 μg / mL, and 15 spectra are collected for each concentration.

[0011] Furthermore, the concentration of the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution is 1 mg / mL.

[0012] Furthermore, the volume ratio of the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution to the deltamethrin standard solution is 1:2-3, and the mixing reaction time is 3-5 minutes.

[0013] Furthermore, the parameters for collecting the SERS spectrum of the mixed solution are: excitation wavelength 785nm, integration time 1-3s, collection band 200-2000cm -1 .

[0014] Furthermore, the ratio of the training set to the test set is 2:1.

[0015] The second technical solution of the present invention:

[0016] A method for intelligent detection of deltamethrin pesticide residues using the above-mentioned GRU regression model includes the following steps:

[0017] (1) Preparation of titanium dioxide-graphite carbon nitride SERS-enhanced substrate (TiO2 / g-C3N4 substrate);

[0018] (2) constructing a GRU regression model using the titanium dioxide-graphite carbon nitride SERS enhancement substrate obtained in step (1);

[0019] (3) extracting, centrifuging, and filtering the food containing the deltamethrin pesticide to obtain a food extract sample;

[0020] (4) The food extract solution is mixed with the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution, the SERS spectrum of the mixed solution is collected and pre-processed by SNV, and the pre-processed SERS spectrum is input into the GRU regression model to obtain the content of cypermethrin in the food.

[0021] Furthermore, the preparation method of the titanium dioxide-graphite carbon nitride SERS enhanced substrate comprises the following steps:

[0022] Titanium tetrachloride is used as a titanium source and added dropwise to a NaOH solution with stirring to generate a titanium dioxide precursor. After adjusting the pH, titanium dioxide nanoparticles are obtained through hydrothermal treatment, centrifugal washing, and drying.

[0023] Calcining melamine to obtain graphite-phase carbon nitride;

[0024] The titanium dioxide nanoparticles and graphite carbon nitride are ultrasonically mixed at a mass ratio of 1:0.15, and evaporated in a water bath to obtain the titanium dioxide-graphite carbon nitride SERS enhancement substrate.

[0025] Furthermore, the food extract is obtained by mixing and leaching food containing deltamethrin pesticide with water in a ratio of 1 g:25 mL, and then centrifuging and filtering.

[0026] Furthermore, the volume ratio of the food extract solution to the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution is 1:2-3, and the mixing reaction time is 3-5 minutes;

[0027] The parameters for collecting the SERS spectrum of the mixed solution are the same as those for constructing the GRU regression model.

[0028] Compared with the prior art, the present invention has the following advantages and technical effects:

[0029] (1) The GRU regression model constructed in this invention automatically retains important information in the time series through a gating mechanism, suppresses invalid fluctuations, and possesses excellent nonlinear regression and generalization capabilities. The GRU algorithm is used to automatically extract key information variables from time series spectra, avoiding the bias caused by artificial feature selection in traditional modeling. Compared with traditional machine learning methods, this model reduces the reliance on artificial feature selection while improving the recognition and generalization capabilities of weak Raman signals under complex background interference.

[0030] (2) In the method for intelligent detection of cypermethrin pesticide residues of the present invention, the TiO2 / g-C3N4 substrate used has both good surface adsorption properties and photochemical enhancement capabilities; the Raman signal shows good uniformity and stability under multiple repeated measurements. The TiO2 / g-C3N4 substrate can be stably stored at room temperature and can be used as a new SERS substrate with high sensitivity, good repeatability, and simple preparation, which is suitable for the efficient and rapid detection of cypermethrin in food.

[0031] (3) The present invention adopts a TiO2 / g-C3N4-based SERS sensor combined with a GRU regression model, which integrates the Raman spectrum signal amplification of cypermethrin and the quantitative regression model. It can accurately predict the cypermethrin content in food, is simple to operate, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings, which constitute part of the present invention, are provided to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are provided to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0033] Figure 1 This is a flow chart of the present invention using the GRU regression model to achieve intelligent detection of deltamethrin pesticide residues;

[0034] Figure 2 Schematic diagram of the model structure of the GRU regression model constructed in Example 1. DETAILED DESCRIPTION

[0035] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as limiting the present invention, but rather as a more detailed description of certain aspects, features, and embodiments of the present invention.

[0036] It should be understood that the terms described herein are intended only to describe particular embodiments and are not intended to limit the present invention. In addition, for numerical ranges herein, it should be understood that each intermediate value between the upper and lower limits of the range is also specifically disclosed. Each smaller range between any intermediate value within a stated value or stated range and any other stated value or intermediate value within the stated range is also encompassed by the present invention. The upper and lower limits of these smaller ranges may be independently included or excluded within the scope.

[0037] Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art. Although only preferred methods and materials are described herein, any methods and materials similar or equivalent to those described herein may also be used in the practice or testing of the present invention. All documents mentioned in this specification are incorporated by reference to disclose and describe the methods and / or materials associated with the documents. In the event of any conflict with any incorporated document, the contents of this specification shall prevail.

[0038] It will be apparent to those skilled in the art that various modifications and variations may be made to the specific embodiments described herein without departing from the scope or spirit of the invention. Other embodiments will be apparent to those skilled in the art from the description of the invention. The description and examples are intended to be exemplary only.

[0039] The words “include,” “including,” “have,” “contain,” etc. used in this document are open-ended terms, meaning including but not limited to.

[0040] An embodiment of the present invention proposes a GRU regression model for detecting cypermethrin, including an input layer, a hidden layer, a fully connected layer, and an output layer; the input layer is used to receive cypermethrin SERS spectral data preprocessed by SNV; the number of hidden layers is 1-2, the number of neurons in the hidden layer is 8-16, and the activation function is a ReLU function; the number of fully connected layers is 1, which is used to map the features extracted by the hidden layer to the output layer, and the output layer outputs 1 feature to characterize the cypermethrin content.

[0041] In a preferred embodiment of the present invention, the method for constructing the GRU regression model for detecting deltamethrin comprises the following steps:

[0042] A deltamethrin standard solution was prepared, and a titanium dioxide-graphite carbon nitride SERS enhancement substrate solution was mixed with the deltamethrin standard solution. The SERS spectrum of the mixture was collected and preprocessed with SNV to remove background interference and random noise in the spectrum to obtain a preprocessed SERS spectrum with a high signal-to-noise ratio. The preprocessed SERS spectrum was divided into a training set and a test set to obtain a GRU regression model for detecting deltamethrin.

[0043] In a preferred embodiment of the present invention, the concentration of the deltamethrin standard solution is 0.01-1000 μg / mL, and 15 spectra are collected for each concentration.

[0044] In a preferred embodiment of the present invention, the concentration of the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution is 1 mg / mL.

[0045] In a preferred embodiment of the present invention, the volume ratio of the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution to the deltamethrin standard solution is 1:2-3, and the mixing reaction time is 3-5 minutes.

[0046] In a preferred embodiment of the present invention, the parameters for collecting the SERS spectrum of the mixed solution are: excitation wavelength 785 nm, integration time 1-3 s, collection band 200-2000 cm -1 .

[0047] In a preferred embodiment of the present invention, the ratio of the training set to the test set is 2:1.

[0048] Exemplarily, the method for constructing a GRU regression model in an embodiment of the present invention includes the following steps:

[0049] First, deltamethrin standard solutions with concentrations of 0.01, 0.1, 1, 10, 100, and 1000 μg / mL were prepared. Then, a TiO2 / g-C3N4 substrate solution (solvent: water, concentration: 1 mg / mL) was added to the deltamethrin standard solution and mixed. The volume ratio of the TiO2 / g-C3N4 substrate solution to the deltamethrin standard solution was 1:2.5. After the mixing reaction lasted for 5 minutes, a SERS spectrum was collected. The portable Raman spectrometer was set to an excitation wavelength of 785 nm, an integration time of 3 s, and a collection band of 200-2000 cm -1 , 15 spectra were collected for each concentration, for a total of 90 spectra; the spectra were subjected to SNV preprocessing using MatlabR2022b software to obtain a SERS spectrum of deltamethrin with a high signal-to-noise ratio, which was the preprocessed SERS spectrum;

[0050] Based on the above-mentioned SERS spectra after SNV preprocessing, they are divided into training set and prediction set in a ratio of 2:1 and input into the GRU regression model. The established GRU model includes input layer, hidden layer, fully connected layer and output layer; the input layer of the GRU model receives SERS spectral data, performs multi-dimensional feature mining through the hidden layer, extracts features of the input spectral information and outputs feature vectors into the fully connected layer, and uses the fully connected layer to map these extracted features to the output layer; the number of hidden layers is 2, the number of neurons in the hidden layer is 8; the activation function is the ReLU function; the number of fully connected layers is set to 1 layer to output 1 feature.

[0051] The embodiment of the present invention also proposes a method for realizing intelligent detection of deltamethrin pesticide residues using the above-mentioned GRU regression model (see flowchart). Figure 1 ), including the following steps:

[0052] (1) Preparation of titanium dioxide-graphite carbon nitride SERS-enhanced substrate (TiO2 / g-C3N4 substrate);

[0053] (2) constructing a GRU regression model using the TiO2 / g-C3N4 substrate obtained in step (1);

[0054] (3) extracting, centrifuging, and filtering the food containing the deltamethrin pesticide to obtain a food extract sample;

[0055] (4) The food extract sample was mixed with the TiO2 / g-C3N4 substrate solution, the SERS spectrum of the mixture was collected and preprocessed by SNV, and the pretreated SERS spectrum was input into the GRU regression model to obtain the content of cypermethrin in the food.

[0056] In a preferred embodiment of the present invention, the method for preparing the TiO2 / g-C3N4 substrate comprises the following steps:

[0057] Titanium tetrachloride is used as a titanium source and added dropwise to a NaOH solution with stirring to generate a titanium dioxide precursor. After adjusting the pH, the precursor is hydrothermally treated, centrifuged, washed, and dried to obtain titanium dioxide (TiO2) nanoparticles.

[0058] Calcination of melamine to obtain graphite carbon nitride (g-C3N4);

[0059] Titanium dioxide nanoparticles and graphite-phase carbon nitride were ultrasonically mixed in a mass ratio of 1:0.15 and evaporated in a water bath to obtain a TiO2 / g-C3N4 substrate.

[0060] In a preferred embodiment of the present invention, the food extract is obtained by mixing food containing deltamethrin pesticide with water at a ratio of 1 g: 25 mL, and then centrifuging and filtering.

[0061] In a preferred embodiment of the present invention, the volume ratio of the food extract solution to the TiO2 / g-C3N4 substrate solution is 1:2-3, and the mixing reaction time is 3-5 minutes;

[0062] The parameters for collecting the SERS spectrum of the mixed solution are the same as those for constructing the GRU regression model.

[0063] The room temperature in the embodiments of the present invention refers to "25±3°C".

[0064] The technical solution of the present invention is further illustrated by the following examples.

[0065] It should be noted that the technical means not described in detail in the embodiments of the present invention, such as the specific operation of using MatlabR2022b software to perform SNV preprocessing on the spectrum and the specific operation of SERS spectrum acquisition, are all existing technologies and can be implemented using conventional technical means in the field.

[0066] Example 1

[0067] A method for intelligent detection of deltamethrin pesticide residues using a GRU regression model includes the following steps:

[0068] (1) Preparation of TiO2 / g-C3N4 substrate

[0069] S1. Weigh 20 mL of titanium tetrachloride as a titanium source and slowly add it dropwise to 60 mL of NaOH solution. Stir and react at 240 rpm at room temperature for 2 h to form a TiO2 precursor. Then, adjust the pH of the solution to 6 using NaOH solution. After standing and homogenizing, perform a hydrothermal reaction at 180°C for 6 h. After the reaction is completed, cool to room temperature, centrifuge and wash the resulting product at 8000 rpm for 10 min. Repeat the washing three times, and dry the precipitate to obtain TiO2 nanoparticles.

[0070] S2. 13 g of melamine was weighed into a 50 mL crucible wrapped in two layers of tin foil. The mixture was heated to 550°C at a rate of 1.5°C / min in air and held in a muffle furnace for 4 h. Upon cooling, a yellow powder, g-C3N4, was obtained.

[0071] S3. The prepared TiO2 nanoparticles and g-C3N4 were ultrasonically mixed at a mass ratio of 1:0.15 for 20 minutes. After mixing evenly, the mixture was placed in an 80°C water bath and evaporated to obtain a TiO2 / g-C3N4 substrate.

[0072] (2) Collection of Deltamethrin SERS Spectra (Raman Spectra)

[0073] First, deltamethrin standard solutions with concentrations of 0.01, 0.1, 1, 10, 100, and 1000 μg / mL were prepared; then, the TiO2 / g-C3N4 substrate solution (solvent: water, concentration: 1 mg / mL) prepared in step (1) was added to the deltamethrin standard solution and mixed, wherein the volume ratio of the TiO2 / g-C3N4 substrate solution to the deltamethrin standard solution was 1:2.5; after the mixing reaction lasted for 5 minutes, a SERS spectrum was collected, wherein the portable Raman spectrometer was set to an excitation wavelength of 785 nm, an integration time of 3 s, and a collection band of 200-2000 cm -1 , 15 spectra were collected for each concentration, for a total of 90 spectra; the spectra were subjected to SNV preprocessing using MatlabR2022b software to obtain a SERS spectrum of deltamethrin with a high signal-to-noise ratio, which was the preprocessed SERS spectrum;

[0074] (3) Construction of GRU regression model

[0075] Based on the SERS spectrum after SNV preprocessing in step (2), the ratio of 2:1 is divided into a training set and a prediction set, which are input into the GRU regression model. The established GRU model includes an input layer, a hidden layer, a fully connected layer and an output layer. The input layer of the GRU model receives the SERS spectrum data, performs multi-dimensional feature mining through the hidden layer, extracts features of the input spectrum information and outputs feature vectors into the fully connected layer, and uses the fully connected layer to map these extracted features to the output layer. The number of hidden layers is 2, and the number of neurons in the hidden layer is 8. The activation function is the ReLU function. The number of fully connected layers is set to 1 to output 1 feature. The model structure of the GRU regression model is as follows: Figure 2 As shown;

[0076] (4) Tea sample pretreatment

[0077] 2.0 g of tea sample was accurately weighed and placed in a 50 mL centrifuge tube. 3.0 mL of pre-prepared deltamethrin standard solution (0.1 μg / mL, to construct a contaminated sample) and 27.0 mL of deionized water were added. After thorough mixing, the sample was treated in an ultrasonic cleaner for 20 min. The mixture was then homogenized using a high-speed homogenizer for 1 min. After homogenization, the sample was centrifuged at 5000 rpm for 10 min, and the supernatant was collected. Finally, the supernatant was filtered through an organic filter membrane with a pore size of 0.22 μm to obtain the tea extract for subsequent detection and analysis.

[0078] (5) Detection of pesticide residues in tea

[0079] The prepared tea extract was mixed with a TiO2 / g-C3N4 substrate solution (prepared as in step (2)) at a volume ratio of 1:2.5 and subjected to a thorough oscillation reaction for 5 minutes to allow the target molecule to fully adsorb onto the SERS substrate surface. Subsequently, a portable Raman spectrometer was used to collect SERS spectra of the mixed solution under the same parameters as in step (2). The obtained raw spectral data was preprocessed using SNV and input into a GRU regression model to obtain quantitative detection results for the concentration of cypermethrin in tea leaves. The average value of the detection result was 0.108 μg / mL. In addition, there was no significant difference compared with the traditional detection method, high-performance liquid chromatography (HPLC), (p < 0.05).

[0080] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A GRU regression model for detecting deltamethrin, characterized in that: Includes input layer, hidden layer, fully connected layer and output layer; The input layer is used to receive the SERS spectrum data of deltamethrin preprocessed by SNV; The number of hidden layers is 1-2, the number of neurons in the hidden layer is 8-16, and the activation function is the ReLU function; The fully connected layer has one layer, which is used to map the features extracted by the hidden layer to the output layer, and the output layer outputs one feature to characterize the content of deltamethrin.

2. The GRU regression model for detecting deltamethrin according to claim 1, wherein The construction method includes the following steps: A deltamethrin standard solution was prepared, and a titanium dioxide-graphite carbon nitride SERS enhanced substrate solution was mixed with the deltamethrin standard solution. The SERS spectra of the mixture were collected and preprocessed with SNV. The preprocessed SERS spectra were divided into training and test sets to obtain a GRU regression model for detecting deltamethrin.

3. The GRU regression model for detecting deltamethrin according to claim 2, wherein The concentration of the deltamethrin standard solution is 0.01-1000 μg / mL.

4. The GRU regression model for detecting deltamethrin according to claim 2, wherein The volume ratio of the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution to the deltamethrin standard solution is 1:2-3, and the mixing reaction time is 3-5 minutes.

5. The GRU regression model for detecting deltamethrin according to claim 2, wherein The parameters for collecting the SERS spectrum of the mixed solution are: excitation wavelength 785nm, integration time 1-3s, collection band 200-2000cm -1 .

6. The GRU regression model for detecting deltamethrin according to claim 2, wherein The ratio of the training set to the test set is 2:

1.

7. A method for intelligent detection of deltamethrin pesticide residues using the GRU regression model according to any one of claims 1 to 6, characterized in that: The following steps are involved: (1) Preparation of titanium dioxide-graphite carbon nitride SERS-enhanced substrate; (2) constructing a GRU regression model using the titanium dioxide-graphite carbon nitride SERS enhancement substrate obtained in step (1); (3) extracting, centrifuging, and filtering the food containing the deltamethrin pesticide to obtain a food extract sample; (4) The food extract solution is mixed with the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution, the SERS spectrum of the mixed solution is collected and pre-processed by SNV, and the pre-processed SERS spectrum is input into the GRU regression model to obtain the content of cypermethrin in the food.

8. The method for realizing intelligent detection of deltamethrin pesticide residues using the GRU regression model according to claim 7, wherein: The preparation method of the titanium dioxide-graphite carbon nitride SERS enhanced substrate comprises the following steps: Titanium tetrachloride is used as a titanium source and added dropwise to a NaOH solution with stirring to generate a titanium dioxide precursor. After adjusting the pH, titanium dioxide nanoparticles are obtained through hydrothermal treatment, centrifugal washing, and drying. Calcining melamine to obtain graphite-phase carbon nitride; The titanium dioxide nanoparticles and graphite carbon nitride are ultrasonically mixed at a mass ratio of 1:0.15, and evaporated in a water bath to obtain the titanium dioxide-graphite carbon nitride SERS enhancement substrate.

9. The method for realizing intelligent detection of deltamethrin pesticide residues using the GRU regression model according to claim 7, wherein: The food extract is obtained by mixing deltamethrin-containing food with water in a ratio of 1 g: 25 mL, and then centrifuging and filtering.

10. The method for realizing intelligent detection of deltamethrin pesticide residues using the GRU regression model according to claim 7, wherein: The volume ratio of the food extract solution to the titanium dioxide-graphite carbon nitride SERS enhancement substrate solution is 1:2-3, and the mixing reaction time is 3-5 minutes; The parameters for collecting the SERS spectrum of the mixed solution are the same as those for constructing the GRU regression model.

Images