Torreya grandis planting soil nutrient monitoring model and method based on big data

Abstract

The invention provides a torreya grandis planting soil nutrient monitoring model and method based on big data. Firstly, a data acquisition terminal platform compatible with various sensors is constructed, and the data acquisition terminal platform can support various sensors, so that comprehensive data acquisition can be carried out in a Chinese torreya planting environment. Then, a big data remote monitoring platform is constructed, and a data analysis module of the big data remote monitoring platform converts the collected data into a readable format so as to facilitate storage and subsequent analysis; the intelligent processing module performs deep analysis and prediction on soil nutrients based on an advanced data analysis algorithm, and helps a user to realize intelligent decision making. And finally, under the support of a big data platform, firstly performing dimension reduction processing on data collected from the detection platform by utilizing a principal component analysis method, and then performing numerical fitting by adopting a multiple linear regression model in combination with a calibration value of a laboratory. Model parameters are optimized through a least square method, and finally real-time monitoring and accurate derivation of the nitrogen, phosphorus and potassium contents of the soil are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent soil detection, and specifically relates to a big data-based nutrient monitoring model and method for Torreya grandis planting soil. Background Technique

[0002] The appropriate balance of nitrogen, phosphorus, and potassium content in the soil of Torreya grandis planting environment has an important impact on the growth, flowering, fruiting, and quality of Torreya grandis. Lack or excess of any one element may lead to poor growth, reduced yield, or decreased fruit quality. Therefore, timely and accurate monitoring of the nutrient content in Torreya grandis planting soil is the key to ensuring the healthy growth of Torreya grandis.

[0003] Traditional soil nutrient detection methods usually adopt chemical analysis methods. This process includes mixing the sample soil with pure water in a certain proportion, and then using chemical reagents for extraction and filtration. Although this method has a certain degree of accuracy, it has the disadvantages of high cost, long detection cycle, and cumbersome operation. In addition, traditional methods are difficult to meet the requirements of Torreya grandis planting for rapid response and real-time monitoring.

[0004] In recent years, many studies have focused on using spectral detection technology for rapid detection of nitrogen, phosphorus, and potassium content in soil. This method can obtain a large amount of data in a short time and has the advantages of rapidity and non-destruction. However, spectral data is usually very complex, and its processing and analysis require high-level algorithms and technologies. The complexity and diversity of the data make traditional analysis methods difficult to apply, increasing the difficulty of data processing. In addition, the high cost of spectral analysis equipment and the need for professional technical personnel also limit its feasibility in wide applications. Summary of the Invention

[0005] The present invention precisely aims at the problems existing in the prior art, such as long detection time and cumbersome data processing operations, and provides a big data-based nutrient monitoring model and method for Torreya grandis planting soil.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A big data-based nutrient monitoring model for Torreya grandis planting soil uses the principal component analysis method to perform dimensionality reduction processing on multi-dimensional soil data and optimize the feature input; after dimensionality reduction, the principal components that have a significant impact on the nitrogen, phosphorus, and potassium content are selected as the model input, and then a multiple linear regression model is applied for parameter estimation; by comparing with the laboratory calibration value, the least squares method is used to optimize the model parameters to minimize the error between the predicted value and the actual value, so as to achieve precise monitoring and real-time derivation of the nitrogen, phosphorus, and potassium content in the soil.

[0008] Further, the soil nutrient monitoring model realizes the detection of nitrogen, phosphorus, and potassium content in the soil based on the principal component analysis method and the multiple linear regression model, and the steps are as follows:

[0009] 1), Use the principal component analysis method to reduce the dimension of the data collected from the detection platform, and identify the principal components that have a significant impact on the nitrogen, phosphorus, and potassium contents. Specifically:

[0010] First, standardize the original data so that the mean of each feature is 0 and the variance is 1;

[0011]

[0012] where, X ij is the j-th feature of the i-th sample, μ j is the mean of the j-th feature, and σ j is the standard deviation of the j-th feature;

[0013] Calculate the covariance matrix through the standardized data:

[0014]

[0015] where, Z is the standardized data matrix, and n is the number of samples;

[0016] Solve the eigenvalues and eigenvectors of the covariance matrix to determine the principal components:

[0017] Cv = λv

[0018] where, λ is the eigenvalue, and v is the corresponding eigenvector;

[0019] Select the first k eigenvectors according to the size of the eigenvalues to form a new feature space;

[0020] Project the original data onto the new feature space to obtain the principal components:

[0021] Y = X·W

[0022] where, Y is the data after dimensionality reduction, and W is the selected eigenvector matrix;

[0023] 2), Use the multiple linear regression model for numerical fitting, and optimize the model parameters through the least squares method to achieve real-time monitoring and accurate derivation of the soil nitrogen, phosphorus, and potassium contents. Specifically:

[0024] Construct a multiple linear regression model:

[0025] Y = β0 + β1X1 + β2X2 + … + β k X k + ∈

[0026] where, Y is the dependent variable (nitrogen, phosphorus, and potassium contents), X1, X2, …, X k are the independent variables (principal components), β0 is the intercept, and β1, β2, …, βk is the regression coefficient, and ∈ is the error term;

[0027] The least squares method is used to estimate the regression coefficient, minimizing the squared difference between the predicted value and the actual value:

[0028]

[0029] where, Ŷ i is the model predicted value;

[0030] By solving the partial derivative of the sum of squared errors with respect to β and setting it equal to zero, the normal equation is obtained:

[0031]

[0032] By solving the normal equation, the optimal regression coefficient β is obtained;

[0033] where, the sum of squared errors function is:

[0034]

[0035] In addition, the present invention also proposes a method for monitoring the soil nutrients of Torreya grandis planting based on big data, and the steps are as follows:

[0036] 1), Construct a data acquisition terminal platform

[0037] Soil pH sensors, soil temperature and humidity sensors, soil conductivity sensors, and salinity sensors are arranged in the Torreya grandis planting area to collect the basic information of the soil in real time, including the original data of soil information such as pH value, temperature, humidity, conductivity, and salinity;

[0038] 2), Construct a big data remote monitoring platform

[0039] The big data remote monitoring platform consists of a data parsing module, an intelligent processing module, and a visualization display module. The data parsing module converts the received original data into a readable format and performs noise filtering and data cleaning; the parsed data will be stored in the database for subsequent access and analysis; the data intelligent processing module will deeply analyze the parsed data, and use the LSTM (Long Short-Term Memory Network) algorithm to evaluate and predict the soil nutrients; automatically generate a decision report based on historical data and real-time data to help users adjust soil management measures in a timely manner;

[0040] 3), Precise monitoring and real-time derivation of the nitrogen, phosphorus, and potassium content in the soil of Torreya grandis planting

[0041] First, the principal component analysis method is used to reduce the dimension of multi-dimensional soil data and optimize the feature input. After dimensionality reduction, the principal components that have a significant impact on the nitrogen, phosphorus, and potassium contents are selected as the model input. Then, a multiple linear regression model is applied for parameter estimation. By comparing with the laboratory calibration values, the least squares method is used to optimize the model parameters to minimize the error between the predicted value and the actual value, thereby achieving precise monitoring and real-time derivation of the nitrogen, phosphorus, and potassium contents in the soil.

[0042] Furthermore, in step 1), a variety of sensors arranged in the Torreya grandis planting area have waterproof and high-temperature resistance characteristics, and the key components are coated with fireproof materials. Each sensor transmits data to the data acquisition terminal platform through a compatible interface. The circuit board of the data acquisition terminal platform uses a moisture-proof coating to improve moisture resistance. The device shell is designed as a sealed structure, which can effectively prevent moisture intrusion, and is equipped with a dedicated heat dissipation device to prevent failures caused by temperature rise.

[0043] Furthermore, in step 2), the data acquisition terminal platform accesses the received data to the Internet through the Ethernet access method. The big data remote monitoring platform converts the received raw data into a readable format and performs noise filtering and data cleaning. In view of the multi-source heterogeneity and dynamic characteristics of the received data, by extracting, fusing, and integrating the collected soil information data of Torreya grandis planting, a unified data unit data set that supports different calculation model processing is formed. The intelligent processing module will deeply analyze the parsed data and use the LSTM (Long Short-Term Memory Network) algorithm to evaluate and predict the soil nutrients.

[0044] The present invention develops a data acquisition terminal platform that is compatible with a variety of sensors, integrates diverse data transmission methods to meet the requirements of different environments, constructs a big data remote monitoring platform, and analyzes and visually displays the data collected by the sensors through data analysis and intelligent processing modules. Utilizing the relationship between parameters such as nitrogen, phosphorus, and potassium contents and conductivity, first, the principal component analysis method is used to reduce the dimension of the data collected by the big data monitoring platform and optimize the input features. Combining with the laboratory calibration values, a multiple linear regression model is adopted and the least squares method is used to optimize the model parameters to achieve real-time monitoring and precise derivation of soil nutrients. Compared with the prior art, the beneficial effects of the present invention are as follows:

[0045] 1), Through the data acquisition terminal platform that is compatible with a variety of sensors, it is possible to collect information such as the pH value, temperature, humidity, conductivity, and salinity of the soil in real time, thereby realizing the instant monitoring of soil nutrients and meeting the requirements of Torreya grandis planting for rapid response.

[0046] 2) The data acquisition terminal platform has good scalability, can be compatible with sensors of different types, quantities, and interfaces, adapts to the needs of diverse data acquisition in practical applications, and enhances the flexibility of the system. Through noise filtering and cleaning of the acquired data, a unified data unit dataset is formed, simplifying the subsequent data processing process, improving the data analysis efficiency, and reducing the complexity of manual intervention.

[0047] 3) In the big data remote monitoring platform, after the data parsing module and the intelligent processing module process the monitoring data through intelligent algorithms, they predict and analyze the growth of Torreya grandis, the change of soil fertility, and the optimal production situation of crops, and make intelligent decision judgments.

[0048] 4) Combining the advantages of the big data platform, in the soil nitrogen, phosphorus, and potassium content detection model based on the principal component analysis method and the multiple linear regression model, the principal component analysis method is used for dimensionality reduction processing, and at the same time, the influencing factors of different soil types on the nitrogen, phosphorus, and potassium contents are identified. Through learning and comparison with the laboratory calibration values, the model can adaptively adjust parameters to adapt to different soil types and states.

[0049] 5) By combining the principal component analysis method with the multiple linear regression model, the feature input is optimized, significantly improving the monitoring accuracy of the soil nitrogen, phosphorus, and potassium contents. The model optimizes the parameters through the least squares method, effectively reducing the prediction error and realizing the accurate derivation of the nutrient content. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is the flowchart of the steps of the method of the present invention.

[0051] Figure 2 is the comparison diagram of the prediction errors of the total nitrogen content detected by using the KNN regression model and the method of the present invention respectively in the case of the present invention.

[0052] Figure 3 is the comparison diagram of the prediction errors of the total phosphorus content detected by using the KNN regression model and the method of the present invention respectively in the case of the present invention.

[0053] Figure 4 is the comparison diagram of the prediction errors of the total potassium content detected by using the KNN regression model and the method of the present invention respectively in the case of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0054] To enable those skilled in the technical field to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the scope of protection of this application.

[0055] Embodiment 1

[0056] Please refer to Figure 1 As shown, the present invention proposes a torreya grandis planting soil nutrient monitoring model and method based on big data, including a data acquisition terminal platform compatible with multiple sensors, a big data remote monitoring platform, and a soil nitrogen, phosphorus, and potassium content detection model based on the principal component analysis method and the multiple linear regression model. The data acquisition terminal platform compatible with multiple sensors is realized by a soil pH value sensor, a soil temperature and humidity sensor, a soil conductivity sensor, and a salinity sensor, and the original data of soil information is obtained by using the above various sensors. The big data remote monitoring platform is realized by a data analysis module, an intelligent processing module, and a visualization display module, which converts the collected data into a readable format, and based on advanced data analysis algorithms, deeply analyzes and predicts the soil nutrients, and makes the data presentation clearer through the way of chart display, facilitating the quick judgment of the soil state. The soil nitrogen, phosphorus, and potassium content detection model based on the principal component analysis method and the multiple linear regression model uses the principal component analysis method to perform dimensionality reduction processing on multi-dimensional soil data and optimize the feature input. After dimensionality reduction, the principal components that have a significant impact on the nitrogen, phosphorus, and potassium content are selected as the model input, and then the multiple linear regression model is applied for parameter estimation. By comparing with the laboratory calibration value, the least squares method is used to optimize the model parameters to minimize the error between the predicted value and the actual value, so as to achieve the accurate monitoring and real-time derivation of the soil nitrogen, phosphorus, and potassium content.

[0057] This system can collect the pH value, temperature, humidity, and conductivity of the soil in real time, meeting the requirements of torreya grandis planting for rapid response. At the same time, the system has good scalability, adapts to different types and quantities of sensors, and enhances flexibility. By combining the principal component analysis method and the multiple linear regression model, optimizing the feature input, significantly improving the monitoring accuracy of the nitrogen, phosphorus, and potassium content in the torreya grandis planting area, reducing the prediction error, and achieving accurate derivation.

[0058] In Torreya grandis planting areas, soil pH sensors, soil temperature and humidity sensors, soil electrical conductivity sensors, and salinity sensors with waterproof and high-temperature resistance characteristics and key components coated with fireproof materials are arranged to collect basic soil information in real time. The sensors transmit data to the data acquisition terminal platform through an RS-485 hub in the data acquisition terminal platform. The platform takes a control box containing an STC12C5A60S2 chip as the core, mainly responsible for information acquisition and transmission.

[0059] The basic soil information collected is connected to the Internet through Ethernet access. The data analysis module in the big data remote monitoring platform converts the received raw data into a readable format and performs noise filtering and data cleaning. The soil information data of Torreya grandis planting soil after noise filtering and data cleaning is extracted, fused, and integrated to form a unified data unit dataset that supports different calculation model processing. The data intelligent processing module will deeply analyze the parsed data and use the LSTM (Long Short-Term Memory Network) algorithm to evaluate and predict soil nutrients. Generate a decision report automatically according to historical data and real-time data to help users adjust soil management measures in a timely manner.

[0060] Based on the principal component analysis method and the multiple linear regression model, the detection of soil nitrogen, phosphorus, and potassium content is realized, and the steps are as follows:

[0061] 1). Use the principal component analysis method to perform dimensionality reduction processing on the data collected from the detection platform, and identify the principal components that have a significant impact on the nitrogen, phosphorus, and potassium content. Specifically:

[0062] First, standardize the original data so that the mean of each feature is 0 and the variance is 1;

[0063]

[0064] Among them, X ij is the jth feature of the ith sample, μ j is the mean of the jth feature, and σ j is the standard deviation of the jth feature;

[0065] Calculate the covariance matrix through the standardized data:

[0066]

[0067] Among them, Z is the standardized data matrix, and n is the number of samples;

[0068] Solve the eigenvalues and eigenvectors of the covariance matrix to determine the principal components:

[0069] Cv = λv

[0070] Among them, λ is the eigenvalue and v is the corresponding eigenvector;

[0071] Select the first k eigenvectors according to the magnitude of the eigenvalues to form a new feature space;

[0072] Project the original data onto the new feature space to obtain the principal components:

[0073] Y = X·W

[0074] Among them, Y is the data after dimensionality reduction, and W is the matrix of selected eigenvectors;

[0075] 2) Use a multiple linear regression model for numerical fitting, optimize the model parameters through the least squares method, and achieve real-time monitoring and accurate derivation of soil nitrogen, phosphorus, and potassium contents. Specifically:

[0076] Construct a multiple linear regression model:

[0077] Y = β0 + β1X1 + β2X2 + … + β k X k + ∈

[0078] Among them, Y is the dependent variable (nitrogen, phosphorus, and potassium contents), X1, X2, …, X k are the independent variables (principal components), β0 is the intercept, β1, β2, …, β k are the regression coefficients, and ∈ is the error term;

[0079] Use the least squares method to estimate the regression coefficients and minimize the sum of squared differences between the predicted values and the actual values:

[0080]

[0081] Among them, Ŷ i is the model predicted value;

[0082] By solving the partial derivative of the sum of squared errors with respect to β and setting it equal to zero, the normal equation is obtained:

[0083]

[0084] By solving the normal equation, the optimal regression coefficient β is obtained;

[0085] Among them, the sum of squared errors function is:

[0086]

[0087] Example 2

[0088] Tests and comparative examples

[0089] For the monitoring method proposed by the present invention, in a torreya grandis planting park, multiple soil sampling points are set, ensuring that the distance between sampling points is maintained between 25 and 40 meters, and a total of 120 soil samples are collected. After each sampling, the soil samples are taken back to the laboratory for chemical analysis to determine their nitrogen, phosphorus, and potassium contents, and the basic information such as pH value, humidity, temperature, and conductivity of the corresponding sampling points is recorded according to the torreya grandis planting big data platform. 40 samples are randomly selected from the 120 collected soil samples as the test set, and the remaining 80 samples are used for model training. For the method of the present invention, before training, the training samples are first reduced in dimension using PCA, and then a multiple linear regression model is used for prediction.

[0090] A comparative experiment is carried out for the KNN regression model. The specific implementation includes:

[0091] The nitrogen, phosphorus, and potassium contents of the soil samples obtained by chemical analysis methods are also used as calibration values. However, the basic information such as pH value, humidity, temperature, and conductivity of the corresponding sampling points is obtained manually using various detection instruments. The Z-score method is used to identify outliers in the obtained data and correct them. Specifically:

[0092] Calculate the mean and standard deviation for each feature such as pH value, humidity, temperature, and conductivity:

[0093]

[0094]

[0095] For each data point, calculate its Z-score and determine whether it exceeds the preset threshold of 3:

[0096]

[0097] Check the Z-score of each data point. If |Z| > 3, then regard this data point as an outlier and replace this outlier with the mean value.

[0098] The corrected data is standardized, and then 40 samples are randomly selected as the test set, and the remaining 80 samples are used for model training.

[0099] Optimize the K value through the cross-validation method. The specific operation of using the KNN regression model for soil nutrient detection is as follows:

[0100] Construct a KNN model:

[0101]

[0102] Among them, is the predicted value, y i is the target value of the K nearest neighbors, and the K value is selected using the cross-validation method.

[0103] Set the range of the K value to be from 1 to 10 and divide the training set into 5 folds. For each K value, train the model on 4 folds, perform validation on the remaining 1 fold, and record its root mean square error. Calculate the average RMSE of 5 validations for each K value, and finally select the K value 5 with the minimum average RMSE as the parameter of the model.

[0104] Use the optimized KNN model to predict the contents of nitrogen, phosphorus, and potassium in the soil. Use the root mean square error (RMSE) to evaluate the performance of the model in soil nutrient monitoring, and conduct a comparative analysis with the prediction results of the method of the present invention.

[0105] Figures 2-4 They are respectively the comparison graphs of the prediction errors of the nitrogen, phosphorus, and potassium contents of the soil samples to be measured in the target domain Torreya grandis plantation park. Figures 2-4 The dotted line in it shows the prediction error of the KNN regression model, and the solid line is the error curve of the method of the present invention. It can be seen from the figure that the method of the present invention has higher accuracy in predicting soil nutrient parameters than the KNN regression model.

[0106] Table 1 Comparison of the root mean square error (RMSE) values between the method of the present invention and the KNN regression model.

[0107]

[0108] It can be seen from Table 1 that in terms of the three indicators of nitrogen, phosphorus, and potassium, the root mean square error values of the present invention are significantly lower than those of the KNN regression model, which indicates that the model established by the method of the present invention can more accurately describe soil nutrient information.

[0109] Compared with the method of using the KNN regression model for soil nutrient monitoring, the advantages of the monitoring method of the present invention are as follows:

[0110] In terms of data collection, compared with using various detection instruments manually with the KNN regression model to obtain basic information such as pH value, humidity, temperature, and conductivity of corresponding sampling points, the present invention can collect information such as the PH value, temperature, humidity, and conductivity of the soil in real time through a data collection terminal compatible with multiple sensors, meet the requirements of Torreya grandis planting for rapid response, and realize the instant monitoring of soil nutrients. Moreover, the data collection terminal platform proposed by the present invention has good scalability, can be compatible with different types, quantities, and interfaces of sensors, adapt to the requirements of diverse data collection in practical applications, and enhance the flexibility of the system.

[0111] In terms of data processing efficiency, the present invention filters and cleans the noise of the collected data to form unified data units, simplifies the subsequent data processing process, improves the data analysis efficiency, and reduces the complexity of manual intervention. However, the method of using the KNN regression model for soil nutrient monitoring has high requirements for data preprocessing. If the data quality is poor, the performance of the model will be significantly affected, and the processing process is cumbersome.

[0112] In terms of the accuracy of soil nutrient monitoring, the method of using the KNN regression model for soil nutrient monitoring is sensitive to noise and outliers. The prediction accuracy is affected by the selection of the K value and the quality of neighboring samples, and there is a lack of built-in feature selection or optimization mechanism. The present invention combines the principal component analysis method with the multiple linear regression model to optimize the feature input, significantly improving the monitoring accuracy of soil nitrogen, phosphorus, and potassium contents. By optimizing the parameters through the least squares method, the prediction error is effectively reduced, and the accurate derivation of nutrient contents is achieved.

Claims

1. A soil nutrient monitoring model for Torreya grandis cultivation based on big data, characterized in that: The principal component analysis method was used to reduce the dimensionality of multidimensional soil data and optimize the feature input. After dimensionality reduction, the principal components that have a significant impact on the nitrogen, phosphorus and potassium content were selected as model inputs, and then the multivariate linear regression model was used for parameter estimation. By comparing with the laboratory calibration values, the least squares method was used to optimize the model parameters to minimize the error between the predicted value and the actual value, thereby achieving accurate monitoring and real-time derivation of soil nitrogen, phosphorus and potassium content.

2. The soil nutrient monitoring model for Torreya grandis cultivation according to claim 1, characterized in that: The soil nutrient monitoring model is based on principal component analysis and multiple linear regression model to detect soil nitrogen, phosphorus and potassium content. The steps are as follows: 1) Use principal component analysis to reduce the dimension of the data collected from the detection platform and identify the main components that have a significant impact on the nitrogen, phosphorus and potassium content, specifically: First, the original data is standardized so that the mean of each feature is 0 and the variance is 1; Among them, X ij is the jth feature of the i-th sample, μ j is the mean of the jth feature, σ j is the standard deviation of the jth feature; Calculate the covariance matrix from the standardized data: Among them, Z is the standardized data matrix, n is the number of samples; Solve for the eigenvalues ​​and eigenvectors of the covariance matrix to determine the principal components: Cv=λv Among them, λ is the eigenvalue and v is the corresponding eigenvector; Select the first k eigenvectors according to the size of the eigenvalue to form a new feature space; Project the original data into the new feature space and obtain the principal components: Y=X·W Among them, Y is the data after dimension reduction, and W is the selected eigenvector matrix; 2) Use the multivariate linear regression model for numerical fitting, optimize the model parameters by the least squares method, and realize the real-time monitoring and accurate derivation of soil nitrogen, phosphorus and potassium content, specifically: Construct a multiple linear regression model: Y=β0+β1X1+β2X2+…+β k X k +∈ Among them, Y is the dependent variable (nitrogen, phosphorus, potassium content), X1, X2, ..., X k is the independent variable (principal component), β0 is the intercept, β1, β2,…, β k is the regression coefficient, ∈ is the error term; The regression coefficients are estimated using the least squares method, minimizing the squared difference between the predicted and actual values: Among them, Y^ i is the model prediction value; By solving the partial derivative of the error sum of squares with respect to β and setting it equal to zero, we get the canonical equation: By solving the regular equation, the optimal regression coefficient β is obtained; The error square sum function is:

3. A method for monitoring soil nutrients in Torreya grandis planting based on the monitoring model described in claim 2, characterized in that: Here are the steps: 1) Build a data collection terminal platform Soil pH sensors, soil temperature and humidity sensors, soil conductivity sensors, and salt sensors are placed in the Torreya grandis planting area to collect basic soil information in real time, including raw soil information data of pH value, temperature, humidity, conductivity, and salt; 2) Build a big data remote monitoring platform The big data remote monitoring platform consists of a data analysis module, an intelligent processing module, and a visualization display module. The data analysis module converts the received raw data into a readable format and performs noise filtering and data cleaning. The parsed data will be stored in the database for subsequent access and analysis. The data intelligent processing module will conduct in-depth analysis of the parsed data and use the LSTM (Long Short-Term Memory Network) algorithm to evaluate and predict soil nutrients. Decision reports will be automatically generated based on historical data and real-time data to help users adjust soil management measures in a timely manner. 3) Accurate monitoring and real-time derivation of nitrogen, phosphorus and potassium content in the soil of Torreya grandis Firstly, principal component analysis was used to reduce the dimensionality of multidimensional soil data and optimize the feature input. After dimensionality reduction, the principal components that have a significant impact on the nitrogen, phosphorus and potassium content were selected as model inputs, and then the multivariate linear regression model was used for parameter estimation. By comparing with the laboratory calibration values, the least squares method was used to optimize the model parameters to minimize the error between the predicted value and the actual value, thereby realizing accurate monitoring and real-time derivation of soil nitrogen, phosphorus and potassium content.

4. The method for monitoring soil nutrients in Torreya grandis planting according to claim 3, characterized in that: The various sensors arranged in the Torreya grandis planting area in step 1) have waterproof and high temperature resistance characteristics, and key components are covered with fireproof materials; each sensor transmits data to the data acquisition terminal platform through a compatible interface; The circuit board of the data acquisition terminal platform adopts moisture-proof coating to improve moisture resistance; the device casing is designed as a sealed structure, which can effectively prevent moisture intrusion, and is equipped with a dedicated heat dissipation device to prevent malfunctions caused by increased temperature.

5. The method for monitoring soil nutrients in Torreya grandis planting according to claim 3, characterized in that: In step 2), the data acquisition terminal platform connects the received data to the Internet via Ethernet; the big data remote monitoring platform converts the received raw data into a readable format, and performs noise filtering and data cleaning; in view of the multi-source heterogeneity and dynamic characteristics of the received data, the collected Torreya grandis planting soil information data is extracted, fused and integrated to form a unified data unit data set that supports different computing models; the intelligent processing module conducts an in-depth analysis of the parsed data and uses the LSTM (long short-term memory network) algorithm to evaluate and predict soil nutrients.

Images