Pesticide residue control analysis method for lycium barbarum planting land
By establishing a time series data set of pesticide application and residues on wolfberry planting land, a pesticide application-residue correlation model is constructed, the trend of cumulative residues of pesticides is predicted and rotational use strategies is formulated, which solves the shortcomings of pesticide residue risk control in the existing technology, and effectively manages the pesticide residue levels in wolfberry planting land.
Patent Information
- Application Number
- CN202411493666.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-24
- Publication Date
- 2025-03-11
AI Technical Summary
The prior art lacks effective strategies to control the risk of pesticide residues on the land of wolfberry planting.
By obtaining pesticide application record data and residue monitoring data of wolfberry planting land, a time series data set of pesticide application and residue amount was established, pretreatment and standardization was performed, and pesticide application-residue correlation model was constructed based on time series analysis methods, predicting the trend of pesticide accumulation residues, and formulating the time interval and order of pesticide rotation use.
Effectively control the risk of pesticide residues, ensure that the residue levels continue to be within an acceptable range, and provide scientific basis and technical support for the sustainable planting of crops such as wolfberry.
Smart Images
Figure CN119598275A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of wolfberry planting, and in particular relates to a method for controlling and analyzing pesticide residues in wolfberry planting land. Background Art
[0002] Pesticide residues in soil have a memory effect. Through long-term monitoring of the dynamic changes of pesticide residues in wolfberry planting land, the relationship between historical pesticide application and current residue levels is analyzed, a pesticide application-residue correlation model is constructed, the trend of pesticide cumulative residues is predicted, and the ecological risks of continuous pesticide use are evaluated. Based on this, the pesticide rotation strategy is optimized to effectively control the risk of pesticide residues. However, existing technologies lack corresponding strategies to control the risk of pesticide residues. Summary of the invention
[0003] In order to solve the above technical problems, the present invention proposes a method for controlling and analyzing pesticide residues in wolfberry planting land to solve the problems existing in the above prior art.
[0004] To achieve the above object, the present invention provides a method for controlling and analyzing pesticide residues in wolfberry planting land, comprising the following steps:
[0005] Obtain pesticide application record data and pesticide residue monitoring data for wolfberry planting land, and establish a time series data set of pesticide application and residue;
[0006] Preprocessing the time series data to obtain preprocessed data, and performing data alignment and standardization processing on the preprocessed data to obtain normalized time series data;
[0007] Based on the time series analysis method and the normalized time series data, a pesticide application-residue association model is established;
[0008] The pesticide application-residue correlation model is used to predict the pesticide cumulative residue trend when different pesticides are continuously applied, and the pesticide cumulative residue trend prediction result is obtained;
[0009] According to the prediction results of the cumulative pesticide residue trend, the time interval and sequence of pesticide rotation are formulated, and residue monitoring is carried out to obtain monitoring data, and the pesticide application-residue association model is dynamically updated according to the monitoring data.
[0010] Preferably, the pesticide application record data includes application date, pesticide name and application dosage;
[0011] The pesticide residue monitoring data includes the detection date and the pesticide residue concentration.
[0012] Preferably, the method for obtaining preprocessed data comprises:
[0013] The time series data is smoothed by using a moving average method to obtain smoothed time series data; the trend term and the period term of the smoothed time series data are fitted by a least square method, and the trend term is removed from the original data to obtain preprocessed data.
[0014] Preferably, the method for obtaining normalized time series data comprises:
[0015] The preprocessed data is standardized by using the Z-score standardization method, and the data is converted into standard normal distribution data with a mean of 0 and a variance of 1. The time series of the standard normal distribution data is aligned by a dynamic time warping algorithm to obtain normalized time series data.
[0016] Preferably, the method for establishing a pesticide application-residue association model comprises:
[0017] An initial prediction model is constructed based on a long short-term memory neural network model, and the pesticide application amount is used as an input sequence and the residue amount is used as an output sequence to train the initial prediction model to obtain a pesticide application-residue amount association model.
[0018] Preferably, the method for formulating the time interval and sequence of pesticide rotation use includes:
[0019] Through the toxicological model, the chronic toxic effects of the pesticide cumulative residue trend prediction results on different organisms are calculated, and a pesticide environmental fate model is constructed based on the physicochemical properties of the pesticide, the application method and the characteristics of the environmental medium. The pesticide environmental fate model is used to simulate the migration and transformation process of the pesticide in the medium and predict the environmental fate endpoint of the pesticide. The time interval and sequence for the rotation of pesticides are formulated based on the chronic toxic effects and the environmental fate endpoint.
[0020] Preferably, the pesticide environmental fate model is a distribution model based on the fugacity model.
[0021] Preferably, the method for dynamically updating the pesticide application-residue association model comprises:
[0022] The real-time monitoring data is input into the trained pesticide application-residue association model for prediction, and the difference between the model prediction result and the measured value is calculated to obtain the prediction error. If the prediction error exceeds the preset threshold, the model is retrained using the real-time monitoring data and the model parameters are dynamically updated.
[0023] Compared with the prior art, the present invention has the following advantages and technical effects:
[0024] The present invention discloses a method for formulating a pesticide rotation strategy based on time series analysis. The method obtains long-term pesticide application and residue monitoring data of wolfberry planting land, establishes a pesticide application-residue correlation model, predicts the trend of pesticide cumulative residue, and thus formulates a pesticide rotation strategy. The method can effectively control the risk of pesticide residues, ensure that the pesticide residue level remains within an acceptable range, and provide scientific basis and technical support for the sustainable planting of wolfberry and other crops. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0026] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION
[0027] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0028] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0029] Embodiment 1
[0030] like Figure 1 As shown, this embodiment provides a method for analyzing pesticide residue control in wolfberry planting land, comprising the following steps:
[0031] Obtain pesticide application record data and pesticide residue monitoring data for wolfberry planting land, and establish a time series data set of pesticide application and residue;
[0032] Preprocess the time series data to obtain preprocessed data, align and standardize the preprocessed data to obtain normalized time series data;
[0033] Based on time series analysis methods and normalized time series data, a pesticide application-residue correlation model was established;
[0034] Through the pesticide application-residue correlation model, the cumulative pesticide residue trend when different pesticides are continuously applied is predicted, and the prediction results of the cumulative pesticide residue trend are obtained;
[0035] According to the prediction results of pesticide cumulative residue trend, the time interval and sequence of pesticide rotation are formulated, and residue monitoring is carried out to obtain monitoring data. The pesticide application-residue correlation model is dynamically updated based on the monitoring data.
[0036] The specific implementation steps are as follows:
[0037] Step S101, obtaining long-term pesticide application record data of wolfberry planting land and pesticide residue monitoring data of the corresponding period, and establishing a time series data set of pesticide application and residue.
[0038] Obtain the pesticide application record data of wolfberry planting land, including application date, pesticide name, application dosage and other information, to form a pesticide application time series data set. Obtain wolfberry pesticide residue monitoring data corresponding to the pesticide application record time, including detection date, pesticide residue concentration, etc., to form a pesticide residue time series data set.
[0039] Step S102, preprocessing the pesticide application and residue time series data, removing outliers and noise data, and performing data alignment and standardization to obtain normalized time series data.
[0040] According to the time series data of pesticide application and residue, the moving average method is used to smooth the original data, remove the noise and outliers in the data, and obtain the smoothed time series data. For the smoothed time series data, the trend term and cycle term of the data are fitted by the least squares method, and the trend term is removed from the original data to obtain the detrended time series data. The Z-score standardization method is used to standardize the detrended time series data, converting the data into standard normal distribution data with a mean of 0 and a variance of 1, eliminating the dimensional differences between different data. According to the standardized time series data, the dynamic time warping algorithm is used to align the time series of different lengths and sampling frequencies so that they have the same time scale and number of sampling points, which is convenient for subsequent data analysis and modeling.
[0041] In this embodiment, the aligned time series data can also be decomposed into approximate coefficients and detail coefficients of different scales by wavelet decomposition, and the abnormal degree of the data is judged according to the statistical characteristics of the wavelet coefficients, and the abnormal data points are removed. The wavelet coefficients after the abnormal points are removed are thresholded and reconstructed to obtain the denoised time series data, and are denormalized to the scale of the original data.
[0042] Step S103, using a time series analysis method, such as an autoregressive moving average model or a long short-term memory neural network model, to establish a pesticide application-residue association model to characterize the quantitative relationship between the pesticide application amount and the residue level.
[0043] For the paired pesticide application amount and residue level time series data, the autoregressive moving average model can be used for modeling. Through model parameter estimation and optimization, the autoregressive moving average model that describes the quantitative relationship between pesticide application amount and residue level can be obtained. It is also possible to consider using the long short-term memory neural network model for modeling. By constructing a neural network structure containing long short-term memory units, and using the pesticide application amount as the input sequence and the residue level as the output sequence, the long short-term memory neural network model is trained. The pesticide application-residue association model can be obtained.
[0044] Step S104, using the established association model, predict the trend of pesticide accumulation residues under different pesticide continuous application scenarios, and evaluate the ecological risk of long-term continuous application of pesticides.
[0045] Predict the residue level in the future. Calculate the chronic toxic effects of accumulated pesticide residues on different organisms through toxicological models, and evaluate the impact of long-term pesticide residues on the ecosystem. Consider the physical and chemical properties of pesticides, application methods, environmental medium characteristics and other factors, build a pesticide environmental fate model, simulate the migration and transformation process of pesticides in soil, water and other media, and predict the environmental fate endpoint of pesticides. Based on the prediction results of the pesticide residue association model and the environmental fate model, form an ecological risk assessment report for the long-term application of pesticides, and provide a decision-making basis for the scientific management of pesticides and ecological protection.
[0046] Step S105, according to the prediction result of the trend of pesticide cumulative residue, determine the time interval and sequence of pesticide rotation, formulate a pesticide rotation strategy, and control the risk of pesticide residue.
[0047] According to the prediction results of pesticide residue trend, determine when each pesticide reaches the peak residue and the duration of the peak, determine the minimum time interval between two consecutive uses of the same pesticide, and avoid the risk of pesticide residue accumulation. Through the peak time of pesticide residue and the safe interval time, preliminarily determine the time node for pesticide rotation and form a rotation schedule. On the basis of the pesticide rotation schedule, further optimize the order of pesticide rotation according to factors such as the control targets, dosage, and duration of efficacy of different pesticides to avoid continuous use of pesticides with similar control targets or high residue risks. Integrate the optimized time interval and sequence of pesticide rotation into the final pesticide rotation strategy, including detailed information such as the time point, dosage, and number of times of use of each pesticide.
[0048] Step S106, applying the pesticide rotation strategy to wolfberry planting practice, regularly monitoring the pesticide residue level in the soil, and feeding the monitoring data back to the correlation model, dynamically updating the model parameters, and improving the prediction accuracy.
[0049] During the wolfberry growing season, according to the pesticide rotation strategy, soil samples are collected regularly to test pesticide residue levels and obtain actual monitoring data. The actual monitoring data is input into the trained pesticide residue prediction model, the difference between the model prediction results and the measured values is evaluated, and the prediction error is calculated. If the prediction error exceeds the preset threshold, the model is retrained using the newly collected monitoring data, and the model parameters are dynamically updated to continuously improve the accuracy of pesticide residue prediction and guide subsequent pesticide rotation decisions.
[0050] Step S107, comprehensively evaluate the implementation effect of the pesticide rotation strategy, determine whether the pesticide residue level is effectively controlled, and further optimize the rotation strategy if necessary to ensure that the pesticide residue risk remains within an acceptable range.
[0051] Apply the optimized pesticide rotation strategy to the pesticide use management system to guide and control farmers' pesticide use. Continuously collect pesticide residue detection data, dynamically update the pesticide residue risk assessment model, and evaluate the pesticide residue risk in real time. When the pesticide residue risk exceeds the acceptable range, timely warn and initiate the re-optimization of the pesticide rotation strategy to ensure that the pesticide residue risk remains controllable.
[0052] The specific implementation is as follows:
[0053] First, in the data collection stage, the type, amount (such as 5 liters of pesticide per hectare), and application time of each application of pesticides are recorded in detail. At the same time, environmental data such as soil pH, organic matter content (such as organic matter content of 5%), temperature (such as average temperature of 22°C), and rainfall (such as monthly rainfall of 100 mm) are collected. In the data preprocessing stage, data cleaning technology is used to remove outliers, such as data with abnormal soil pH values lower than 3 or higher than 9 are considered abnormal and excluded. Principal component analysis (PCA) is used for feature selection to reduce model complexity and improve computational efficiency. The data is normalized to ensure that data of different dimensions can contribute fairly to the model training process. The random forest algorithm is used to analyze the pesticide residue data, and the national standard threshold for pesticide residues is set to 0.1 mg / kg. If the detection value exceeds this threshold, it is judged that the pesticide residue exceeds the standard. For the excessive pesticide residue data, combined with the pesticide rotation data, the Apriori association rule mining algorithm is used, with the minimum support set to 0.5 and the minimum confidence set to 8, to mine the association rules between the pesticide rotation strategy and the pesticide residue level. Based on the mined association rules, the C5 decision tree algorithm is used to generate optimization suggestions for the pesticide rotation strategy, such as extending the rotation interval of the two pesticides to more than 15 days, reducing the frequency of use of high-risk pesticides to no more than 2 times a month, etc. The optimized pesticide rotation strategy is uploaded to the pesticide use management system. The system automatically pushes optimization suggestions based on the farmers' pesticide use records, and warns and controls the use of excessive risk pesticides. The system automatically updates the pesticide residue detection data every day, and uses the sliding window method to dynamically update the pesticide residue risk assessment model based on the data of the last 30 days. When the pesticide residue risk prediction value exceeds 0.08 mg / kg, the warning mechanism is triggered, and the re-optimization process of the pesticide rotation strategy is automatically started. Through iterative optimization, the pesticide residue risk is stabilized within a controllable range.
[0054] In this embodiment, in order to establish a correlation model between pesticide application amount and residue level, the pesticide application amount data and corresponding residue level monitoring data for 36 months from January 2020 to December 2022 were first collected. The pesticide application amount data and the residue level data were paired according to the month to form 36 groups of paired time series data. For the paired data, the autoregressive moving average model ARMA (p, q) was used for modeling, where p is the order of the autoregressive term and q is the order of the moving average term. The optimal p and q values were determined to be 2 and 1, respectively, using the AIC criterion, and the ARMA (2, 1) model was obtained. The model parameters were estimated using the maximum likelihood estimation method, and the autoregressive coefficients were 6 and 3, and the moving average coefficient was 8. The estimated parameters were substituted into the ARMA (2, 1) model to fit the relationship between pesticide application amount and residue level, and the goodness of fit R was found. 2is 85, indicating that the model has a good fitting effect. In order to further improve the prediction accuracy, the long short-term memory neural network LSTM model can be used for modeling. A neural network containing 2 LSTM hidden layers is constructed, each hidden layer contains 64 neurons. The pesticide application amount is used as the input sequence and the residue level is used as the output sequence. The model is trained using the Adam optimization algorithm with a training cycle of 100 epochs and a batch size of 32. The root mean square error RMSE of the trained LSTM model on the test set is 08, which is better than 12 of the ARMA model. Therefore, the LSTM model is selected as the final prediction model. Using the trained LSTM model, the pesticide application amount data from January to June 2023 are input, and the pesticide residue levels in the corresponding months are predicted to be 32, 28, 35, 31, 29 and 33 mg / kg respectively. The predicted values are compared with the actual monitoring values, and the average relative error is calculated to be 2%, which meets the accuracy requirements of pesticide residue prediction. Finally, using this LSTM model, the pesticide application amounts are 8, 0, and 2 kg / hm 2 When the pesticide residue levels were 36, 42, and 49 mg / kg, it was predicted, providing a quantitative basis for the rational control of pesticide application.
[0055] In this embodiment, the specific method for generating optimization suggestions for pesticide rotation strategy is: when obtaining the residual detection data of different pesticides on crops, the pesticide residue is detected by gas chromatography-mass spectrometry (GC-MS). Taking 40 commonly used pesticides as an example, samples are collected every 7 days for 12 consecutive weeks to obtain a data set of pesticide residues changing over time. Using the statsmodels library in Python, the time series data of each pesticide residue are fitted with an ARIMA model, and the optimal model parameters are selected by indicators such as AIC and BIC. For example, for pesticide A, the ARIMA (1, 1, 2) model is fitted, and the residues predicted for the next 4 weeks are 32 mg / kg, 45 mg / kg, 51 mg / kg and 49 mg / kg respectively. According to the prediction results, pesticide A reaches a residual peak of 51 mg / kg in the 10th week, and the peak lasts for 2 weeks, so the minimum interval between two consecutive uses of pesticide A is 4 weeks. Repeat the above steps to obtain the peak time and safe interval time of 40 pesticides, and form a preliminary rotation schedule.
[0056] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for controlling and analyzing pesticide residues in wolfberry planting land, characterized in that: The following steps are involved: Obtain pesticide application record data and pesticide residue monitoring data for wolfberry planting land, and establish a time series data set of pesticide application and residue; Preprocessing the time series data to obtain preprocessed data, and performing data alignment and standardization processing on the preprocessed data to obtain normalized time series data; Based on the time series analysis method and the normalized time series data, a pesticide application-residue association model is established; The pesticide application-residue correlation model is used to predict the pesticide cumulative residue trend when different pesticides are continuously applied, and the pesticide cumulative residue trend prediction result is obtained; According to the prediction results of the cumulative pesticide residue trend, the time interval and sequence of pesticide rotation are formulated, and residue monitoring is carried out to obtain monitoring data, and the pesticide application-residue association model is dynamically updated according to the monitoring data.
2. The method for controlling and analyzing pesticide residues in wolfberry planting land according to claim 1, characterized in that: The pesticide application record data includes application date, pesticide name and application dosage; The pesticide residue monitoring data includes the detection date and the pesticide residue concentration.
3. The method for controlling and analyzing pesticide residues in wolfberry planting land according to claim 1, characterized in that: The method for obtaining preprocessed data comprises: The time series data is smoothed by using a moving average method to obtain smoothed time series data; the trend term and the period term of the smoothed time series data are fitted by a least square method, and the trend term is removed from the original data to obtain preprocessed data.
4. The method for controlling and analyzing pesticide residues in wolfberry planting land according to claim 1, characterized in that: The method for obtaining normalized time series data comprises: The preprocessed data is standardized by using the Z-score standardization method, and the data is converted into standard normal distribution data with a mean of 0 and a variance of 1. The time series of the standard normal distribution data is aligned by a dynamic time warping algorithm to obtain normalized time series data.
5. The method for controlling and analyzing pesticide residues in wolfberry planting land according to claim 1, characterized in that: The method for establishing the pesticide application-residue association model comprises: An initial prediction model is constructed based on a long short-term memory neural network model, and the pesticide application amount is used as an input sequence and the residue amount is used as an output sequence to train the initial prediction model to obtain a pesticide application-residue amount association model.
6. The method for controlling and analyzing pesticide residues in wolfberry planting land according to claim 1, characterized in that: The method for formulating the time interval and sequence of pesticide rotation use includes: Through the toxicological model, the chronic toxic effects of the pesticide cumulative residue trend prediction results on different organisms are calculated, and a pesticide environmental fate model is constructed based on the physicochemical properties of the pesticide, the application method and the characteristics of the environmental medium. The pesticide environmental fate model is used to simulate the migration and transformation process of the pesticide in the medium, and the environmental fate endpoint of the pesticide is predicted. The time interval and sequence for the rotation of pesticides are formulated based on the chronic toxic effects and the environmental fate endpoint.
7. The method for controlling and analyzing pesticide residues in wolfberry planting land according to claim 6, characterized in that: The pesticide environmental fate model is a distribution model based on the fugacity model.
8. The method for controlling and analyzing pesticide residues in wolfberry planting land according to claim 1, characterized in that: The method for dynamically updating the pesticide application-residue association model includes: The real-time monitoring data is input into the trained pesticide application-residue association model for prediction, and the difference between the model prediction result and the measured value is calculated to obtain the prediction error. If the prediction error exceeds the preset threshold, the model is retrained using the real-time monitoring data and the model parameters are dynamically updated.