Soil nutrient prediction system and method based on radioactive elements
By using a soil nutrient prediction model based on a multi-layer neural network and a large language model, and by conducting on-site detection using soil radioactive element information, the problem of high cost and slow speed in existing soil nutrient detection technologies has been solved, enabling rapid and accurate soil nutrient prediction and improvement decisions.
Patent Information
- Application Number
- CN202511514311.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-22
- Publication Date
- 2026-02-10
AI Technical Summary
Existing methods for soil nutrient testing require laboratory analysis, which is costly, slow, and the samples are prone to deterioration during transit, making it difficult to achieve rapid testing of deep soil properties in large-scale farmland.
By establishing a soil nutrient prediction model based on a multi-layer neural network, utilizing soil radioactive element information for on-site detection, and combining a large language model and knowledge base, rapid prediction of soil nutrients and decision-making on remediation can be achieved.
It enables rapid and accurate detection of soil nutrients, reduces costs, provides decision-making solutions for soil improvement, and improves detection efficiency and accuracy.
Smart Images
Figure CN121503758A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent agricultural technology, and in particular to a method and system for predicting soil nutrients based on radioactive elements. Background Technology
[0002] Currently, existing soil nutrient testing methods mainly rely on laboratory chemical testing. This involves collecting fresh soil samples and preparing them through a process of air drying, grinding, sieving, and packaging. The collected samples are then sent to the laboratory where soil nutrients are measured using chemical reagents. Examples of methods include using sodium bicarbonate solution for shaking extraction to measure available phosphorus, ammonium acetate solution for centrifugation extraction to measure available potassium, and potassium dichromate oxidation-external heating method to measure organic matter.
[0003] A single sample full-item test requires ≥72 hours (including sample preparation and analysis), the cost of a single sample test is ¥300-500 (including manpower and consumables), the equipment investment is ¥1 million+ (for a complete laboratory setup) and requires professional laboratory technicians to operate precision instruments.
[0004] In practical applications, existing technologies have the following problems and shortcomings:
[0005] Commonly used soil testing methods include atomic absorption spectrometry (AAS), atomic fluorescence spectrometry (AFS), X-ray fluorescence spectrometry (XRF), ion chromatography (IC), electrochemical analysis methods, chemical methods, and laser-induced breakdown spectroscopy (LIBS). All of these methods require pretreatment of the soil, such as drying, grinding, and sieving, resulting in poor timeliness, high testing costs, and difficulty in achieving rapid testing of soil properties in the deep underground layers of large-scale farmland.
[0006] With the continuous deepening and development of soil quality research, radioactive nuclides in the soil can effectively reflect changes in soil fertility and structure through the release of gamma rays in the quantitative evaluation of soil erosion intensity. These radioactive nuclides are relatively stably and uniformly distributed on the soil surface through atmospheric deposition, and their slow radioactive decay causes physical and mechanical transport of soil particles. Therefore, radioactive nuclides can, to a certain extent, characterize the degree of soil erosion and quality degradation.
[0007] In summary, the aforementioned defects in the existing technology are due to the high cost and slow speed of laboratory testing, and the fact that soil samples are prone to deterioration during transport to the laboratory, affecting the test results. Through analysis and research on soil radioactive elements and nutrient elements obtained from testing, the inventors discovered that this defect can be solved by establishing and training an artificial intelligence model. That is, without sending samples to the laboratory for testing, the information on soil radioactive elements measured on-site can be directly converted into soil nutrient content information through a computer model program.
[0008] Therefore, using soil radioactive element information for rapid soil nutrient detection is of great practical significance. There is an urgent need to research a novel method and system for predicting soil nutrients based on radioactive elements. Summary of the Invention
[0009] To address the problem in existing technologies that cannot directly infer soil nutrient content from soil radioactive element information, a method and system for predicting soil nutrients based on radioactive elements is proposed.
[0010] In a first aspect, embodiments of this application provide a soil nutrient prediction system based on radioactive elements, the system comprising:
[0011] Soil nutrient prediction model construction module: Based on different soil types, it is used to construct and train various soil nutrient prediction models based on multi-layer neural networks. The soil nutrient prediction model is used to predict the content of various nutrients in the soil based on the content of various radioactive elements in the input soil.
[0012] Soil Data Query Agent Module: This module takes a query containing radioactive elements in the soil as input to a large language model, and then accesses the database and returns the query results based on the parsed database statement.
[0013] Soil nutrient prediction intelligent agent module: It is used to read the query results of the database, select a suitable soil nutrient prediction model through the large model, call the soil nutrient prediction model, and realize the prediction and output of soil nutrient content information based on soil radioactive element information, and integrate and output a land health check report.
[0014] In this embodiment of the invention, the above-mentioned soil nutrient prediction system based on radioactive elements further includes:
[0015] Soil remediation decision-making intelligent agent module: Based on the input land health check report, it accesses the land knowledge base and uses a large model to match the most relevant remediation documents to the input land conditions, retrieves relevant information, and generates a land remediation decision plan through the large model.
[0016] In this embodiment of the invention, the soil nutrient prediction model construction module performs the following steps:
[0017] A loss function is defined to measure the deviation between the model's predicted values and the actual values, so that the deviation between the predicted values and the actual values of the soil nutrient prediction model on the training data is minimized, and the optimal parameters for model training are obtained.
[0018] The gradient descent algorithm is used for model training, and the gradient of the loss function of the prediction model with respect to different parameters is calculated.
[0019] During training, forward propagation is used to calculate and retain the relevant intermediate values of each layer, and backpropagation uses the chain rule to calculate the gradient of the loss function with respect to the parameters of each layer. Regularization and other methods are used to prevent overfitting of the parameters.
[0020] In this embodiment of the invention, the soil data query intelligent agent module described above performs the following steps:
[0021] The input natural language query statement is transformed into a database query statement through a large language model;
[0022] Based on the information in the database query statement, access the database and return the query results;
[0023] The query results are output and stored in the form of a database file.
[0024] In this embodiment of the invention, the soil nutrient prediction intelligent agent module performs the following steps:
[0025] Based on the query results, the land data is parsed and matched with the corresponding soil nutrient prediction model to predict soil nutrients.
[0026] The soil nutrient prediction model performs inference to obtain nutrient prediction data for the land;
[0027] The soil nutrient prediction data and location information data are integrated to output a land health check report.
[0028] In this embodiment of the invention, the soil remediation decision-making intelligent agent module described above performs the following steps:
[0029] Analyze the input land health check report, access all documents in the land knowledge base, and use a large model to match the governance documents related to the input land situation;
[0030] Retrieve relevant information from governance documents and generate governance decision-making schemes for land types using a large language model.
[0031] Secondly, embodiments of this application provide a method for predicting soil nutrients based on radioactive elements, employing the aforementioned soil nutrient prediction system based on radioactive elements, and the method includes:
[0032] The steps for building a soil nutrient prediction model are as follows: Based on different soil types, various soil nutrient prediction models based on multi-layer neural networks are built and trained. The soil nutrient prediction models are used to predict the content of various nutrient elements in the soil based on the content of various radioactive elements in the input soil.
[0033] Soil data query steps: Input the query statement containing soil radioactive elements into the large language model, access the database based on the parsed database statement, and return the query results;
[0034] Soil nutrient prediction steps: Read the query results from the database, select a suitable soil nutrient prediction model through the large model, call the soil nutrient prediction model, and output the soil nutrient content information based on the soil radioactive element information, and integrate and output the land health check report.
[0035] In this embodiment of the invention, the above-mentioned method for predicting soil nutrients based on radioactive elements further includes:
[0036] Soil remediation decision-making steps: Based on the input land health report, access the land knowledge base, and use a large model to match the most relevant remediation documents to the input land conditions, retrieve relevant information, and generate a land remediation decision plan through the large model.
[0037] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a method for predicting soil nutrients based on radioactive elements.
[0038] Fourthly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the method for predicting soil nutrients based on radioactive elements as described above.
[0039] Compared with existing technologies, it has the following outstanding advantages:
[0040] 1) The method and system of this invention propose a soil nutrient prediction model; the current soil nutrient content can be predicted by the radioactive element information of the soil.
[0041] 2) The method and system of this invention propose an intelligent soil information analysis and decision-making system; it can automatically calculate and analyze soil nutrients based on the selected plot information and the soil nutrient prediction model, and provide a decision-making scheme for soil improvement. Attached Figure Description
[0042] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0043] Figure 1 This is a schematic diagram of the soil nutrient prediction system based on radioactive elements according to the present invention;
[0044] Figure 2This is a schematic diagram of the calling process between intelligent agents according to an embodiment of the present invention;
[0045] Figure 3 This is a schematic diagram of a fully connected neural network structure according to an embodiment of the present invention;
[0046] Figure 4 This is a schematic diagram of forward propagation in an embodiment of the present invention;
[0047] Figure 5 This is a schematic diagram of back propagation in an embodiment of the present invention;
[0048] Figures 6a-6c This is a schematic diagram of the model evaluation results in an embodiment of the present invention;
[0049] Figure 7 This is a schematic diagram of the soil data query intelligent agent module in an embodiment of the present invention;
[0050] Figure 8 This is a schematic diagram of the soil nutrient prediction intelligent agent module in an embodiment of the present invention;
[0051] Figure 9 This is a schematic diagram of the soil remediation decision-making intelligent agent module according to an embodiment of the present invention;
[0052] Figure 10 This is a flowchart of a soil nutrient prediction method based on radioactive elements, as described in an embodiment of the present invention.
[0053] Figure 11 This is a schematic diagram of the computer hardware of the present invention. Detailed Implementation
[0054] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.
[0055] It should also be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.
[0056] It should also be understood that, in various embodiments of the present invention, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
[0057] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.
[0058] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0059] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0060] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0061] To make the above-mentioned features and effects of the present invention clearer and easier to understand, specific embodiments are described below in conjunction with the accompanying drawings. This specification discloses one or more embodiments incorporating the features of the present invention. The disclosed embodiments are merely illustrative. The scope of protection of the present invention is not limited to the disclosed embodiments, but is defined by the appended claims.
[0062] The following are system embodiments corresponding to the above method embodiments. This embodiment can be implemented in conjunction with the above embodiments. The relevant technical details mentioned in the above embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above embodiments.
[0063] This invention aims to propose a method for predicting soil nutrients based on radioactive elements. The invention constructs and trains a deep learning model based on a multi-layer neural network. The input is the content of 18 radioactive elements (uranium, thorium, potassium, etc.) in the soil, and the output is the model's predicted content of 9 nutrient elements (nitrogen, phosphorus, potassium, etc.) in the soil. The accuracy of the nutrient content predicted by the model compared to laboratory measurements can reach 80%-90%.
[0064] This invention designs and develops a multi-agent collaborative system based on a large model. The system mainly includes a data query agent, a model calling agent, and a text retrieval agent. The agents call each other sequentially and cooperate with each other. It can analyze soil conditions using a soil nutrient prediction model and provide reasonable land management solutions with the help of a soil knowledge base.
[0065] The system of this application embodiment will be described in detail below with reference to specific embodiments:
[0066] Example 1
[0067] like Figure 1 As shown in the embodiments of this application, a soil nutrient prediction system based on radioactive elements is proposed. This soil nutrient prediction system is a multi-agent collaborative system based on large model decision-making. The system mainly includes a soil nutrient prediction model construction module 101, a soil data query agent module 102, a soil nutrient prediction agent module 103, and a soil remediation decision-making agent module 104. The calling process between agents is as follows: Figure 2 As shown:
[0068] Soil nutrient prediction model construction module 101: Based on different soil types, it is used to construct and train various soil nutrient prediction models based on multi-layer neural networks. The soil nutrient prediction model is used to predict the content of various nutrient elements in the soil based on the content of various radioactive elements in the input soil.
[0069] Soil data query intelligent agent module 102: It is used to input the query statement containing soil radioactive elements into the large language model, access the database according to the generated database statement, and return the query results;
[0070] Soil nutrient prediction intelligent agent module 103: It is used to read the query results of the database, select a suitable soil nutrient prediction model through the large model, call the soil nutrient prediction model, and realize the prediction and output of soil nutrient content information based on soil radioactive element information, and integrate and output a land health check report.
[0071] In this embodiment of the invention, the above-mentioned soil nutrient prediction system based on radioactive elements further includes:
[0072] Soil remediation decision-making intelligent agent module 104: Based on the input land health check report, it accesses the land knowledge base and uses a large model to match the most relevant remediation documents to the input land conditions, retrieves relevant information, and generates a land remediation decision-making scheme through the large model.
[0073] In this embodiment of the invention, the soil nutrient prediction model construction module 101 performs the following steps:
[0074] A loss function is defined to measure the deviation between the model's predicted values and the actual values, so that the deviation between the predicted values and the actual values of the soil nutrient prediction model on the training data is minimized, and the optimal parameters for model training are obtained.
[0075] The gradient descent algorithm is used for model training, and the gradient of the loss function of the prediction model with respect to different parameters is calculated.
[0076] During training, forward propagation is used to calculate and retain the relevant intermediate values of each layer, and backpropagation uses the chain rule to calculate the gradient of the loss function with respect to the parameters of each layer. Regularization and other methods are used to prevent overfitting of the parameters.
[0077] Specifically, in a specific embodiment of the present invention, the training of the above-mentioned prediction model includes:
[0078] 1) Neural network model structure
[0079] like Figure 3 As shown, an Artificial Neural Network (ANN) is a mathematical or computational model that mimics the information processing of a biological nervous system. It consists of a large number of nodes (or neurons) and the connections between them. Each node represents a simple processing unit capable of receiving input signals, processing them, and outputting a signal. The strength of the connections between these nodes, i.e., the weights, can be adjusted through a learning process, enabling the network to perform specific tasks such as classification, regression, and clustering.
[0080] The current problem type is a typical multiple regression problem, and it is trained using a multilayer fully connected neural network model and related algorithms as shown in the figure below.
[0081] 2) Neural Network Model Training Methods
[0082] Model training involves finding suitable parameters that minimize the deviation between the model's predicted values and the actual values on the training data. First, a loss function is defined to measure the deviation between the model's predicted values and the actual values. The loss function used in this model training is the mean squared error (MSE).
[0083] The model training uses the gradient descent algorithm to calculate the gradient of the model's loss function with respect to different parameters (the gradient is the direction in which the function changes the fastest, and it can be obtained by differentiation).
[0084] like Figure 4 and Figure 5 As shown, in actual training, the neural network model is trained using gradient descent by combining forward propagation (calculating and retaining the relevant intermediate values of each layer) and backpropagation (using the chain rule to calculate the gradient of the loss function with respect to the parameters of each layer). Simultaneously, dropout and regularization are used during training to prevent overfitting.
[0085] 3) Model Performance Evaluation
[0086] Cross-validation of the deep learning model was performed using a small dataset of 350 sets of data (k=7, i.e., 300 sets of training data and 50 sets of test data). The evaluation results are as follows: Figures 6a-6c As shown.
[0087] In this embodiment of the invention, the soil data query intelligent agent module 102 performs the following steps:
[0088] The input natural language query statement is transformed into a database query statement through a large language model;
[0089] Based on the information in the database query statement, access the database and return the query results;
[0090] The query results are output and stored in the form of a database file.
[0091] Specifically, in a specific embodiment of the present invention, the soil data query intelligent agent module 102 has the ability to connect to and query the database and perform data query and processing. It can convert the input natural language query into an SQL database query statement through a large model, and then call relevant tools according to the database information to access the database and return the query results. At the same time, it can choose to output and store the query results in the form of a file.
[0092] like Figure 7 As shown, suppose the example database stores information on time, farm, plot, temperature, type, and soil radioactivity measured by gamma spectroscopy. The figure below illustrates how a data query agent can directly query and return all information for plots 10 to 20 of the Northeast Demonstration Farm from June 25th to 27th.
[0093] In this embodiment of the invention, the soil nutrient prediction intelligent agent module 103 performs the following steps:
[0094] Based on the query results, the land data is parsed and matched with the corresponding soil nutrient prediction model to predict soil nutrients.
[0095] The soil nutrient prediction model performs inference to obtain nutrient prediction data for the land;
[0096] The soil nutrient prediction data and location information data are integrated to output a land health check report.
[0097] Specifically, in this embodiment of the invention, the soil nutrient prediction agent module 103 has the ability to access file directories and perform file analysis, call model code, and integrate information. When working, the agent first reads the soil data file that needs to be parsed, then calls tools to access all types of soil nutrient prediction models in the model directory and selects a suitable soil model from the large model list. It then calls the model to convert the radioactive element information in the data into soil nutrient content information, and finally integrates all the data into complete information about the land plot.
[0098] like Figure 8 As shown, taking the output of the aforementioned data query agent as an example, assuming the model library contains multiple pre-trained soil nutrient prediction models, the agent will first parse the land data and match the Northeast saline-alkali land model as the most suitable model for soil nutrient prediction. Then, it will call the tool to execute the model code for inference, obtaining the land nutrient prediction data. Finally, it will integrate the soil nutrient data with other location information and output the final land health check report. The following figure illustrates the agent's workflow and results.
[0099] In this embodiment of the invention, the soil remediation decision-making intelligent agent module 104 performs the following steps:
[0100] Analyze the input land health check report, access all documents in the land knowledge base, and use a large model to match the governance documents related to the input land situation;
[0101] Retrieve relevant information from governance documents and generate governance decision-making schemes for land types using a large language model.
[0102] Specifically, in a specific embodiment of the present invention, the soil remediation decision-making intelligent agent module 104 has the function of accessing the file directory and performing file analysis, retrieving information related to the input question in the knowledge base and generating answers using a large model, i.e., general retrieval augmented generation (RAG) capability, which will not be elaborated here.
[0103] like Figure 9 As shown, taking the land health check report output by the soil nutrient prediction agent module 103 as an example, assuming the knowledge base already contains multiple documents related to soil remediation, the agent will first analyze the input land health check report, then call tools to access all documents in the land knowledge base, and use a large model to match the remediation documents most relevant to the input land situation, retrieve relevant information from them, and finally output a remediation decision plan for the land through a large model generation method. The following figure illustrates the working process and results of this agent.
[0104] Specifically, in this embodiment of the invention, the Python version required for the soil nutrient prediction system based on radioactive elements to run is 3.11, and the corresponding Python packages and versions are as follows:
[0105] torch == 2.7.0
[0106] torchvision == 0.21.0
[0107] transformers == 4.52.3
[0108] langchain == 0.3.25
[0109] langchain-tools == 0.1.34
[0110] llama-index == 0.12.37
[0111] llama-index-embeddings-huggingface == 0.5.4
[0112] llama-index-retrievers-bm25 == 0.5.2
[0113] jieba == 0.42.1
[0114] typing == 3.7.4.3
[0115] dashscope == 1.23.3
[0116] dotenv == 0.9.9
[0117] As described above, the method of the present invention can be implemented well.
[0118] Compared with the prior art, the present invention has the following outstanding advantages and beneficial effects:
[0119] The soil nutrient prediction model proposed in this invention can predict the current soil nutrient content based on the radioactive element information of the soil. The intelligent soil information analysis and decision-making system proposed in this invention can automatically calculate and analyze soil nutrients based on the selected plot information and provide a decision-making scheme for soil improvement.
[0120] Example 2
[0121] like Figure 10 As shown, this application provides a method for predicting soil nutrients based on radioactive elements, employing the aforementioned soil nutrient prediction system based on radioactive elements. The method includes:
[0122] Soil nutrient prediction model construction step 201: According to different soil types, various soil nutrient prediction models based on multi-layer neural networks are constructed and trained. The soil nutrient prediction models are used to predict the content of various nutrient elements in the soil based on the content of various radioactive elements in the input soil.
[0123] Soil data query step 202: Input the query statement containing soil radioactive elements into the large language model, access the database according to the generated database statement, and return the query results;
[0124] Soil nutrient prediction step 203: Read the query results from the database, select a suitable soil nutrient prediction model through the large model, call the soil nutrient prediction model, and realize the prediction and output of soil nutrient content information based on soil radioactive element information, and integrate and output the land health check report.
[0125] In this embodiment of the invention, the above-mentioned method for predicting soil nutrients based on radioactive elements further includes:
[0126] Soil remediation decision-making step 204: Based on the input land health check report, access the land knowledge base, and use a large model to match the remediation documents most relevant to the input land conditions, retrieve relevant information, and generate a land remediation decision-making plan through the large model.
[0127] Example 3
[0128] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for predicting soil nutrients based on radioactive elements.
[0129] Example 4
[0130] This application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the soil nutrient prediction method based on radioactive elements as described above.
[0131] In addition, combined Figure 1 The method for predicting soil nutrients based on radioactive elements described in this application can be implemented by electronic devices, such as computer devices. Figure 11 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of this application.
[0132] In some embodiments, the computer device may further include a communication interface 83 and a bus 80. For example, Figure 11 As shown, the processor 81, memory 82, and communication interface 83 are connected through bus 80 and complete communication with each other.
[0133] Specifically, the processor 81 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.
[0134] The memory 82 can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program instructions executed by the processor 81.
[0135] The processor 81 reads and executes computer program instructions stored in the memory 82 to implement any of the soil nutrient prediction methods based on radioactive elements in the above embodiments.
[0136] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0137] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. A soil nutrient prediction system based on radioactive elements, characterized in that, The system includes: Soil nutrient prediction model construction module: Based on different soil types, it is used to construct and train various soil nutrient prediction models based on multi-layer neural networks. The soil nutrient prediction model is used to predict the content of various nutrient elements in the soil based on the content of various radioactive elements in the input soil. Soil Data Query Agent Module: This module takes a query containing radioactive elements in the soil as input to a large language model, and then accesses the database and returns the query results based on the parsed database statement. Soil nutrient prediction intelligent agent module: used to read the query results of the database, select a suitable soil nutrient prediction model through the large model, call the soil nutrient prediction model, realize the prediction and output of soil nutrient content information based on the soil radioactive element information, and integrate and output a land health check report.
2. The soil nutrient prediction system based on radioactive elements according to claim 1, characterized in that, The system also includes: Soil remediation decision-making intelligent agent module: Based on the input land health check report, access the land knowledge base, and use a large model to match the most relevant remediation documents to the input land conditions, retrieve relevant information, and generate a land remediation decision plan through the large model.
3. The soil nutrient prediction system based on radioactive elements according to claim 1 or 2, characterized in that, The soil nutrient prediction model construction module includes the following steps: A loss function is defined to measure the deviation between the model's predicted values and the actual values, so that the deviation between the predicted values and the actual values of the soil nutrient prediction model on the training data is minimized, and the optimal parameters for model training are obtained. The gradient descent algorithm is used for model training, and the gradient of the loss function of the prediction model with respect to different parameters is calculated. During training, forward propagation is used to calculate and retain the relevant intermediate values of each layer, and backpropagation uses the chain rule to calculate the gradient of the loss function with respect to the parameters of each layer. Regularization and other methods are used to prevent overfitting of the parameters.
4. The soil nutrient prediction system based on radioactive elements according to claim 1 or 2, characterized in that, The soil data query intelligent agent module performs the following steps: The input natural language query statement is transformed into a database query statement through a large language model; Based on the information in the database query statement, access the database and return the query results; The query results are output and stored in the form of a database file.
5. The soil nutrient prediction system based on radioactive elements according to claim 1 or 2, characterized in that, The soil nutrient prediction intelligent agent module performs the following steps: Based on the query results, the land data is parsed and matched with the soil nutrient prediction model corresponding to the soil model to predict soil nutrients. The soil nutrient prediction model performs inference to obtain nutrient prediction data for the land; The soil nutrient prediction data and location information data are integrated to output a land health check report.
6. The soil nutrient prediction system based on radioactive elements according to claim 2, characterized in that, The soil remediation decision-making intelligent agent module performs the following steps: The input land health check report is analyzed, all documents in the land knowledge base are accessed, and a large model is used to match the governance documents related to the input land situation. Relevant information is retrieved from the governance documents, and a governance decision scheme for the land type is generated using a large language model.
7. A method for predicting soil nutrients based on radioactive elements, employing the soil nutrient prediction system based on radioactive elements as described in any one of claims 1-6, characterized in that, The method includes: Soil nutrient prediction model construction steps: According to different soil types, various soil nutrient prediction models based on multilayer neural networks are constructed and trained. The soil nutrient prediction models are used to predict the content of various nutrient elements in the soil based on the content of various radioactive elements in the input soil. Soil data query steps: Input the query statement containing soil radioactive elements into the large language model, access the database based on the parsed database statement, and return the query results; Soil nutrient prediction steps: Read the query results of the database, select a suitable soil nutrient prediction model through the large model, call the soil nutrient prediction model, and realize the prediction and output of soil nutrient content information based on the soil radioactive element information, and integrate and output a land health check report.
8. The method for predicting soil nutrients based on radioactive elements according to claim 7, characterized in that, The method further includes: Soil remediation decision-making steps: Based on the input land health report, access the land knowledge base, use a large model to match the most relevant remediation documents to the input land conditions, retrieve relevant information, and generate a land remediation decision plan through the large model.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method for predicting soil nutrients based on radioactive elements as described in any one of claims 7-8.
10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the soil nutrient prediction method based on radioactive elements as described in any one of claims 7-8.