A preparation method and system of a weeding composition based on application analysis

Abstract

The present invention relates to the technical field of the preparation of weeding compositions, and particularly relates to a method and a system for preparing a weeding composition based on application analysis. The method includes: obtaining plant information and application scenario information; the environment-agent-plant interaction network retrieves single-component agent information related to the plant information and the application scenario information from the agent effect database, and constructs a network of the interaction among plant characteristics, environmental characteristics, and agent characteristics, wherein the single-component agent information is diuron and carfentrazone-ethyl; analyzing the ratio of diuron and carfentrazone-ethyl through the environment-agent-plant interaction network to obtain a preparation plan for the weeding composition; the plan evaluation model evaluates the preparation plan for the weeding composition based on historical experimental data, and optimizes the preparation plan for the weeding composition according to the evaluation result. Through the present invention, the problem of the relatively high cost of the ratio of diuron and carfentrazone-ethyl based on actual application is effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of the preparation of weeding compositions, and particularly to a method and a system for preparing a weeding composition based on application analysis. Background Art

[0002] Diuron is a systemic and broad-spectrum herbicide that can be absorbed by the roots and leaves of plants, mainly through the roots. After the roots of weeds absorb the drug, it is transmitted to the above-ground leaves and spreads around along the leaf veins, inhibiting the Hill reaction in photosynthesis. This drug requires light to kill plants, causing the affected weeds to fade from the leaf tips and edges until the whole leaf withers, unable to produce nutrients and starving to death. However, at high doses, it can be used as a non-selective herbicide; Carfentrazone-ethyl is a foliar treatment herbicide of the triazolinone type, suitable for controlling broad-leaved weeds and sedges in various gramineous crop fields. Carfentrazone-ethyl is a contact-type selective herbicide that is absorbed by plant leaves within 15 minutes after spraying. It is not affected by rain. Symptoms of poisoning appear in weeds 3 - 4 hours later, and they die within 2 - 4 days. Currently, there are already multiple preparations containing carfentrazone-ethyl registered for use in wheat fields, rice fields, sugarcane fields, etc. However, as a broad-spectrum herbicide, if used alone for a long time and in a large area, it may cause target weeds to develop drug resistance and damage non-target plants.

[0003] If diuron and carfentrazone-ethyl are used in combination, the above problems can be solved to a certain extent. However, different ratios need to be made according to different environments and situations, including factors such as the types of target weeds, growth stages, environmental conditions, and crop types. Otherwise, problems such as weed resistance, crop damage, and ecological damage will occur. Currently, the combined use of diuron and carfentrazone-ethyl for individual crops and application scenarios requires repeated experiments and mostly relies on human judgment, and the costs in all aspects of the process are relatively high.

[0004] The information disclosed in this background art section is only intended to deepen the understanding of the overall background art of the present disclosure, and should not be regarded as an admission or any form of implication that this information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0005] The present invention provides a method and a system for preparing a weeding composition based on application analysis, which can effectively solve the problems in the background art.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for preparing a weeding composition based on application analysis, the method comprising:

[0008] Obtaining plant information and application scenario information, where the plant information includes target plant information and non-target plant information;

[0009] Construct an environment-chemical-plant interaction network based on a chemical effect database. The environment-chemical-plant interaction network retrieves single-component chemical information related to the plant information and application scenario information from the chemical effect database, and constructs a network of interactions among plant characteristics, environmental characteristics, and chemical characteristics. Among them, the single-component chemical information is diuron and carfentrazone-ethyl;

[0010] Analyze the ratio of diuron and carfentrazone-ethyl through the environment-chemical-plant interaction network, and obtain a preparation plan for a herbicidal composition according to the analysis results;

[0011] Construct a scheme evaluation model. The scheme evaluation model evaluates the preparation plan of the herbicidal composition based on historical experimental data, and optimizes the preparation plan of the herbicidal composition according to the evaluation results.

[0012] Furthermore, constructing an environment-chemical-plant interaction network based on a chemical effect database includes:

[0013] Retrieve the single-component chemical related to the application scenario information from the chemical effect database according to the application analysis;

[0014] Establish a plant-chemical interaction relationship. Use the plant information, diuron, and carfentrazone-ethyl as nodes of the network, and establish strong and weak network connections according to the sensitivity and toxicity responses of the single-component chemical to the target plant information and non-target plant information in the plant information;

[0015] Set the edge weights of the plant-chemical network. The edge weights represent the absorption and metabolism capabilities of plants for single-component chemicals;

[0016] Add the application scenario information to the plant-chemical network, add the attributes of network nodes according to the application scenario information, and add the attributes of edge weights according to the influence of the application scenario information on the plant-chemical network.

[0017] Furthermore, it also includes establishing the chemical effect database, including:

[0018] Collect the chemical information of diuron and carfentrazone-ethyl, and conduct first-classification management of the single-component chemicals according to the adaptability of the single-component chemicals to the application scenario information;

[0019] Collect plant information, including target plant information and non-target plant information, and conduct second-classification management of the single-component chemicals according to the sensitivity of the single-component chemicals to the plant information;

[0020] Establish a first-classification management table and a second-classification management table according to the first-classification management and the second-classification management respectively, and set chemical labels for the single-component chemicals;

[0021] Create an index on the pharmaceutical label fields of the first classification management table and the second classification management table, perform cross-search through combined query, and create a stored procedure and related functions to simplify the query process.

[0022] Furthermore, analyze the application of the herbicide composition through the environment-herbicide-plant interaction network, and obtain a preparation plan for the herbicide composition according to the analysis results, including:

[0023] Use the method of complex network theory to identify the structural characteristics of the environment-herbicide-plant interaction network, including the degree distribution of nodes, clustering coefficient, and small-world property of the network;

[0024] According to the structural characteristics, analyze the centrality of the environment-herbicide-plant interaction network, set the node sensitivity, and select the ratio and dosage of diuron and carfentrazone-ethyl according to the node sensitivity to form a preparation plan for the herbicide composition.

[0025] Furthermore, analyze the centrality of the environment-herbicide-plant interaction network, including:

[0026] Calculate the degree of each node, and judge the degree centrality of the node according to the calculation result of the degree. The degree of the node is the number of edges connected to the node;

[0027] Calculate the degree of each node as a mediator, and judge the betweenness centrality of the node according to the degree of the mediator. The degree of the mediator is the frequency of passing through the node on the path connecting other nodes;

[0028] Calculate the closeness centrality of the node. The closeness centrality is the distance between each node and other nodes, indicating the rapid influence of the node on the network;

[0029] Calculate the eigenvector centrality. The eigenvector centrality is the weight of the node connected to other nodes and the centrality of the other nodes.

[0030] Furthermore, construct a scheme evaluation model, including:

[0031] Collect historical data information on the preparation of herbicide compositions, and extract the application scenario information, target plant information, and non-target plant information corresponding to the historical preparation plan for herbicide compositions according to the application analysis. The historical data information on the preparation of herbicide compositions includes the performance of diuron and carfentrazone-ethyl on target plants, non-target plants, and application scenarios respectively, as well as the performance of the ratio of diuron and carfentrazone-ethyl on target plants, non-target plants, and application scenarios;

[0032] Clean the historical herbicidal composition preparation data information, and integrate the drug information, application scenario information, target plant information, and non-target plant information of the cleaned historical herbicidal composition preparation data information into a historical sample set by implementation project;

[0033] Select the deep learning method to learn the historical sample set and obtain the relevance of the elements in the historical sample set;

[0034] Establish an input port and an output port according to the learning result. Countless input ports input the application scenario information, target plant information, and non-target plant information, and the output port outputs the herbicidal composition preparation plan.

[0035] Furthermore, train and verify the plan evaluation model, including:

[0036] Construct a training set and a verification set according to the historical herbicidal composition preparation data information;

[0037] Train the plan evaluation model through the training set and verify the trained result through the verification set;

[0038] Optimize the plan evaluation model according to the training and verification results.

[0039] Furthermore, regularly update the historical herbicidal composition preparation data information in the plan evaluation model, and re-learn the updated historical herbicidal composition preparation data information through the deep learning method.

[0040] A herbicidal composition preparation system based on application analysis, the system includes:

[0041] An information acquisition module that acquires plant information and application scenario information, and the plant information includes target plant information and non-target plant information;

[0042] An interaction network generation module that constructs an environment-chemical-plant interaction network based on a chemical effect database. The environment-chemical-plant interaction network retrieves single-component chemical information related to the plant information and application scenario information from the chemical effect database, and constructs a network of interactions among plant characteristics, environmental characteristics, and chemical characteristics, where the single-component chemical information is diuron and carfentrazone-ethyl;

[0043] A preparation plan generation module that analyzes the ratio of diuron and carfentrazone-ethyl through the environment-chemical-plant interaction network and obtains a herbicidal composition preparation plan according to the analysis result;

[0044] The preparation plan evaluation module constructs a plan evaluation model. The plan evaluation model evaluates the herbicide composition preparation plan based on historical experimental data and optimizes the herbicide preparation plan according to the evaluation results.

[0045] Further, the interaction network generation module includes:

[0046] The information retrieval unit retrieves the single-component agent related to the application scenario information from the agent effect database according to the application analysis.

[0047] The interaction relationship establishment unit establishes the plant-agent interaction relationship, takes the plant information, diuron, and carfentrazone-ethyl as the nodes of the network, and establishes strong and weak network connections according to the sensitivity and toxicity responses of the single-component agent to the target plant information and non-target plant information in the plant information.

[0048] The edge weight setting unit sets the edge weights of the plant-agent network, and the edge weights represent the absorption and metabolism capabilities of the plant to the single-component agent.

[0049] The scenario information adding unit adds the application scenario information to the plant-agent network, adds the attributes of the network nodes according to the application scenario information, and adds the attributes of the edge weights according to the influence of the application scenario information on the plant-agent network.

[0050] Through the technical solution of the present invention, the following technical effects can be achieved:

[0051] Effectively solves the problem that the cost of the ratio of diuron and carfentrazone-ethyl based on actual application is relatively high, and generates different ratios of diuron and carfentrazone-ethyl for specific application analysis in a data-driven and intelligent manner, so that the composition fully adapts to the actual application scenario and the target and non-target plants.

[0052] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the description. And in order to make the above and other purposes, features, and advantages of the present application more obvious and understandable, the specific embodiments of the present application are specifically given below. Description of the Drawings

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0054] Figure 1Schematic flow chart of a method for preparing a herbicidal composition based on application analysis;

[0055] Figure 2 Schematic flow chart of constructing an environment - agent - plant interaction network;

[0056] Figure 3 Schematic flow chart of establishing an agent effect database;

[0057] Figure 4 Schematic flow chart of constructing a scheme evaluation model. Detailed implementation manners

[0058] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this invention belongs. The terms used in the description of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0060] Embodiment 1

[0061] As Figure 1 shown, the present application provides a method for preparing a herbicidal composition based on application analysis, and the method includes:

[0062] A method for preparing a herbicidal composition based on application analysis, the method includes:

[0063] S10: Obtain plant information and application scenario information, where the plant information includes target plant information and non - target plant information;

[0064] S20: Based on the agent effect database, construct an environment - agent - plant interaction network. The environment - agent - plant interaction network retrieves single - component agent information related to the plant information and application scenario information from the agent effect database, and constructs a network of interactions between plant characteristics, environmental characteristics, and agent characteristics. Among them, the single - component agent information is diuron and carfentrazone - ethyl;

[0065] S30: Analyze the ratio of diuron and carfentrazone - ethyl through the environment - agent - plant interaction network, and obtain a preparation scheme for the herbicidal composition according to the analysis results;

[0066] S40: Construct a scheme evaluation model. The scheme evaluation model evaluates the preparation scheme of the herbicidal composition based on historical experimental data, and optimizes the preparation scheme of the herbicidal composition according to the evaluation results.

[0067] Specifically, first, obtain the plant information and application scenario information of the current preparation task. The plant information may include the attribute information of the target plant information and non-target plant information. Through the plant attributes, it is possible to better target the sensitivity or absorption degree of the target and non-target to the medicament. The application scenario information takes into account the environment, soil, climate, etc. of the application scenario, which can produce a good weeding reaction with the application composition. Secondly, when configuring the weeding composition, the impact on the application environment can also be fully considered to achieve green environmental protection and not damage the ecology. The medicament effect database constructs a one-to-one effect relationship between the single-component medicament and the plant, and between the single-component medicament and the environment and soil. Based on the medicament effect database, the single-component medicament information interacts with the plant information and application scenario information to construct a comprehensive network model. Considering factors such as the toxicity, concentration, and application method of the medicament, according to the network analysis results, determine the best herbicide combination plan, including the type, ratio, concentration, etc. of the medicament. Finally, establish a plan evaluation model using historical data information and existing knowledge, so as to adjust the medicament composition preparation plan obtained from the interaction network analysis to a certain extent through the feedback of the empirical ratio.

[0068] Through the technical solution of the present invention, the problem of the relatively high ratio cost of diuron and carfentrazone-ethyl based on actual application is effectively solved, and through digital and intelligent methods, different ratios of diuron and carfentrazone-ethyl are analyzed for specific applications, so that the composition is fully adapted to the actual application scenario and the target plant and non-target plant.

[0069] Furthermore, as Figure 2 shown, construct an environment-medicament-plant interaction network based on the medicament effect database, including:

[0070] S211: Retrieve the single-component medicament related to the application scenario information from the medicament effect database according to the application analysis;

[0071] S212: Establish a plant-medicament interaction relationship, take the plant information, diuron and carfentrazone-ethyl as the nodes of the network, and establish strong and weak network connections according to the sensitivity and toxicity reactions of the single-component medicament to the target plant information and non-target plant information in the plant information;

[0072] S213: Set the edge weight of the plant-medicament network, and the edge weight represents the absorption and metabolism ability of the plant to the single-component medicament;

[0073] S214: Add the application scenario information to the plant-medicament network, and add the attributes of the network nodes according to the application scenario information, and add the attributes of the edge weight according to the influence of the application scenario information on the plant-medicament network.

[0074] Based on the above embodiments, an environment-pesticide-plant interaction network is constructed. Growth characteristics, physiological characteristics, and leaf surface structures of target plants and non-target plants are obtained from the pesticide effect database, and data information such as the composition, concentration, and toxicity of pesticides is acquired. The data information retrieved from the database is used to construct the attributes of the nodes and edges of the network. Among them, plant characteristics and pesticide characteristics are used as the attributes of the nodes, and the toxicity degree of pesticides to plants or the absorption and reaction ability of plants to pesticides are used as the attributes of the edges. Then, the application scenario information is added to the established plant-pesticide interaction relationship. The setting method of the node and edge weights of the network is the same as the above method. In the process of establishing the environment-pesticide-plant interaction network, the method adopted in this embodiment is to first establish the plant-pesticide interaction relationship, and then add the application scenario information to the established network. Since the effect of pesticides on plants is a direct and substantial relationship, their interaction relationship will directly affect the growth, development, and health of plants. Therefore, establishing the relationship network of pesticides and plants first can more accurately simulate the action modes and effects of different pesticides on plants, providing a more accurate basis for adding environmental factors later. Environmental factors refer to the location factors of plants, such as geographical climate, temperature, soil, etc. On the basis of establishing the pesticide-plant relationship, environmental factors are added to form a comprehensive and systematic environment-pesticide-plant interaction network. This method comprehensively considers pesticide characteristics, plant characteristics, and environmental factors, making the evaluation more comprehensive and accurate.

[0075] Furthermore, as Figure 3 shown, it also includes establishing a pesticide effect database, including:

[0076] S221: Collect the pesticide information of diuron and carfentrazone-ethyl, and conduct the first classification management of single-component pesticides according to the adaptability of single-component pesticides to application scenario information;

[0077] S222: Collect plant information, including target plant information and non-target plant information, and conduct the second classification management of single-component pesticides according to the sensitivity of single-component pesticides to plant information;

[0078] S223: Establish the first classification management table and the second classification management table respectively according to the first classification management and the second classification management, and set pesticide labels for the corresponding single-component pesticides;

[0079] S224: Establish an index on the pesticide label fields of the first classification management table and the second classification management table, perform cross-search through joint query, and establish stored procedures and related functions to simplify the query process.

[0080] On the basis of the above embodiments, during the construction of the medicament effect database, information on single-component medicaments was collected, and their adaptability was evaluated according to the application scenario information. The medicaments were classified and managed for the first time according to their adaptability. In the same way, information on target plants and non-target plants was collected, and their sensitivity was evaluated. The medicaments were classified and managed for the second time according to the sensitivity of the plants. Then, a medicament classification management table and medicament labels were established, and the information on adaptability and sensitivity was marked on the medicaments respectively. Finally, an index was established on the medicament label fields to achieve the functions of combined query and cross-retrieval. At the same time, stored procedures and related functions were established to simplify the query process. Through the implementation of these steps, a complete medicament effect database was established, which facilitated the subsequent construction of the environment-medicament-plant interaction network and the rapid retrieval and query of data. This network can conduct comprehensive analysis based on medicament characteristics, plant information, and application scenario information, providing strong support for formulating appropriate herbicide combination plans.

[0081] Furthermore, the application of the herbicide composition is analyzed through the environment-medicament-plant interaction network, and a preparation plan for the herbicide composition is obtained according to the analysis results, including:

[0082] Using the method of complex network theory to identify the structural characteristics of the environment-medicament-plant interaction network, including the degree distribution of nodes, clustering coefficient, and small-world property of the network;

[0083] According to the structural characteristics, analyze the centrality of the environment-medicament-plant interaction network, set the node sensitivity, and select the ratio and dosage of diuron and carfentrazone-ethyl according to the node sensitivity to form a preparation plan for the herbicide composition.

[0084] As a preference of the above embodiments, the herbicide composition is analyzed through the environment-agent-plant interaction network. First, the structural characteristics of the network are identified by using the method of complex network theory. Specifically, when determining the dataset of the environment-agent-plant interaction network, relevant information on agents, plants, and the environment is collected and an agent effect database is constructed. Then, the network is actually defined, such as the relationship between agent plants and scenarios under the network, and the structural characteristics of the network are analyzed. Among them, the node degree distribution is the number of edges connected to the node, the clustering coefficient is the degree of mutual connection between the neighbors of the node, the small world of the network is to detect the high clustering of the network but with a short average path length. Specifically, most nodes can be connected to each other through a small number of edges, so the average path length of the network is relatively short, which means that the information spreads relatively fast in the network, and the nodes in the network tend to form clusters, that is, the neighbors of the nodes are also connected to each other, which results in many dense subgroups in the network, that is, the clustering coefficient is relatively high. In addition to the short-distance connections between nodes, there are also a small number of long-distance connections, which cross the local structure in the network and make the network show global characteristics. When analyzing the centrality of the network, the centrality of each node is calculated to determine the importance and sensitivity of the node. According to the sensitivity setting of the node, a suitable single-component agent is selected to form a preparation plan for the herbicidal composition. Finally, when formulating the herbicidal composition plan, the proportioning plan can be formulated according to the characteristics of the agent and the network analysis results, and the plan is optimized to improve the herbicidal effect and reduce the damage to non-target plants. This method can improve the accuracy of the herbicidal composition preparation plan, reduce costs and resource waste, thereby improving work efficiency and accuracy.

[0085] Furthermore, the analysis of the centrality of the environment-agent-plant interaction network includes:

[0086] Calculate the degree of each node, and judge the degree centrality of the node according to the calculation result of the degree. The degree of the node is the number of edges connected to the node;

[0087] Calculate the degree of each node as a mediator, and judge the betweenness centrality of the node according to the degree of the mediator. The degree of the mediator is the frequency of passing through the node on the path connecting other nodes;

[0088] Calculate the closeness centrality of the node. The closeness centrality is the distance between each node and other nodes, indicating the rapid influence of the node on the network;

[0089] Calculate the eigenvector centrality. The eigenvector centrality is the weight of the node connected to other nodes and the centrality of other nodes.

[0090] As a preference of the above embodiments, regarding degree centrality, the degree of a chemical agent node represents its ability to affect other plants, and the degree of a plant node represents the number of chemical agents that affect it; regarding betweenness centrality, for a chemical agent node, a chemical agent with a high betweenness centrality affects the key paths connecting other plants, while for a plant node, a plant with a high betweenness centrality plays a key role in the process of chemical agent transmission; regarding closeness centrality, for a chemical agent node, a chemical agent with a high closeness centrality affects the surrounding plants faster, and for a plant node, a plant with a high closeness centrality is more easily affected by the surrounding chemical agents; regarding eigenvector centrality, for a chemical agent node, a chemical agent with a high eigenvector centrality is a chemical agent with a greater influence in the network, and for a plant node, a plant with a high eigenvector centrality is more easily affected or plays a greater role in affecting other plants; among them, for eigenvector centrality calculation, the interaction network can be first transformed into an adjacency matrix; then the adjacency matrix is eigen-decomposed to obtain eigenvectors and corresponding eigenvalues; then the eigenvectors are normalized to ensure that the sum of all their elements is 1; and the i-th element in the normalized eigenvector is divided by the corresponding eigenvalue. The centrality indexes calculated through the analysis process are used to find the chemical agents and plants with higher centrality, and the ratio of the chemical agent composition is adjusted or more suitable chemical agents are selected according to the analysis results to obtain a preparation plan for the herbicidal composition. By analyzing centrality, the importance and influence degree of each node in the environment-chemical agent-plant interaction network can be understood more deeply. Also, according to the centrality analysis results, the ratio of the herbicide can be optimized, more suitable chemical agents can be selected, the herbicidal effect can be improved, and at the same time, by understanding which plants are more easily affected by the chemical agents, the impact on non-target plants can be reduced and the ecological environment can be protected.

[0091] Furthermore, as Figure 4 shown, a construction plan evaluation model is built, including:

[0092] S41: Collect historical data information on the preparation of herbicidal compositions, and extract the application scenario information, target plant information, and non-target plant information corresponding to the historical herbicidal composition preparation plan according to application analysis. The historical data information on the preparation of herbicidal compositions includes the performance of diuron and carfentrazone-ethyl on target plants, non-target plants, and application scenarios, as well as the performance of the ratio of diuron and carfentrazone-ethyl on target plants, non-target plants, and application scenarios;

[0093] S42: Clean the historical data information on the preparation of herbicidal compositions, and integrate the chemical agent information, application scenario information, target plant information, and non-target plant information of the cleaned historical data information on the preparation of herbicidal compositions into a historical sample set based on implementation items;

[0094] S43: Select the deep learning method to learn the historical sample set to obtain the correlation of the elements in the historical sample set;

[0095] S44: Establish input ports and output ports according to the learning results. Countless input ports input application scenario information, target plant information, and non-target plant information, and the output port outputs the preparation plan of the weeding composition.

[0096] In this embodiment, through the collection and integration of historical weeding composition preparation data, a scheme evaluation model for deep learning is constructed. Specifically, the data of historical weeding composition preparation is collected, and the corresponding application scenario information, target plant information, and non-target plant information are extracted from it. Then, the collected data is cleaned and integrated to ensure the quality and accuracy of the data. A suitable deep learning model is selected, and the historical sample set is used as the training data to train the model. The correlation between the agents, application scenarios, target plants, and non-target plants in the historical data is learned. Input ports are set in the model to receive application scenario information, target plant information, and non-target plant information, and output ports are set to output the preparation plan of the weeding composition generated according to the input information. Through this scheme evaluation model, personalized preparation plans of weeding compositions can be quickly generated according to different historical application scenarios, target plants, and non-target plant information, reducing the trial-and-error cost and time, improving the efficiency and accuracy of scheme formulation, and learning from the historical data information to obtain the weeding preparation plan and evaluating and adjusting the weeding preparation plan obtained from the interaction network analysis.

[0097] Furthermore, training and validating the scheme evaluation model includes:

[0098] Construct a training set and a validation set according to the historical weeding composition preparation data information;

[0099] Train the scheme evaluation model through the training set and validate the trained result through the validation set;

[0100] Optimize the scheme evaluation model according to the training and validation results.

[0101] Furthermore, regularly update the historical weeding composition preparation data information in the scheme evaluation model, and relearn the updated historical weeding composition preparation data information through the deep learning method.

[0102] The above embodiments include training and validating a solution evaluation model, where the validation set can use multiple different sets of content, including repeatedly validating the solution evaluation model based on the differences in application scenario information and plant information; in addition, updating the historical data information on the preparation of herbicidal compositions and re-learning using deep learning methods within a certain period. As time goes by, historical experience gradually loses a certain timeliness, and the combined preparation of medicaments will cause a tilt in effects and weights in some detailed factors. This requires real-time updating of the content of deep learning and redefining the model to prevent the evaluation of the model from losing accuracy.

[0103] Embodiment 2

[0104] Based on the same inventive concept as a method for preparing a herbicidal composition based on application analysis in the foregoing embodiments, the present invention further provides a system for preparing a herbicidal composition based on application analysis, the system including:

[0105] An information acquisition module for acquiring plant information and application scenario information, where the plant information includes target plant information and non-target plant information;

[0106] An interaction network generation module for constructing an environment-agent-plant interaction network based on a medicament effect database. The environment-agent-plant interaction network retrieves single-component medicament information related to the plant information and application scenario information from the medicament effect database, and constructs a network of interactions between plant characteristics, environmental characteristics, and medicament characteristics. Among them, the single-component medicament information is diuron and carfentrazone-ethyl;

[0107] A preparation plan generation module for analyzing the ratios of diuron and carfentrazone-ethyl through the environment-agent-plant interaction network, and obtaining a preparation plan for the herbicidal composition according to the analysis results;

[0108] A preparation plan evaluation module for constructing a plan evaluation model, where the plan evaluation model evaluates the preparation plan for the herbicidal composition based on historical experimental data, and optimizes the preparation plan for the herbicide according to the evaluation results.

[0109] The above adjustment system in the present invention can effectively implement the method for preparing a herbicidal composition based on application analysis, and the technical effects that can be achieved are as described in the foregoing embodiments, which will not be elaborated here.

[0110] Furthermore, the interaction network generation module includes:

[0111] An information retrieval unit for retrieving single-component medicaments related to the application scenario information from the medicament effect database according to the application analysis;

[0112] An interaction relationship establishment unit establishes a plant-pesticide interaction relationship, taking plant information, diuron, and carfentrazone-ethyl as nodes of the network, and establishing strong and weak network connections according to the sensitivity and toxicity responses of the single-component pesticide to the target plant information and non-target plant information in the plant information respectively;

[0113] An edge weight setting unit sets the edge weights of the plant-pesticide network, and the edge weights represent the absorption and metabolism capabilities of plants for single-component pesticides;

[0114] A scenario information adding unit adds application scenario information to the plant-pesticide network, adds attributes of network nodes according to the application scenario information, and adds attributes of edge weights according to the influence of the application scenario information on the plant-pesticide network.

[0115] Similarly, for the above optimization solutions of the system, the corresponding optimization effects of the methods in Embodiment 1 can also be respectively achieved, and details are not described here again.

[0116] Although the present application has been described in conjunction with specific features and their embodiments, it is obvious that various modifications and combinations can be made without departing from the spirit and scope of the present application. Accordingly, this specification and the drawings are merely exemplary illustrations of the present application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A method for preparing a herbicidal composition based on application analysis, characterized in that: The method comprises: Acquire plant information and application scenario information, wherein the plant information includes target plant information and non-target plant information; An environment-pesticide-plant interaction network is constructed based on a pesticide effect database, wherein the environment-pesticide-plant interaction network retrieves single-component pesticide information related to the plant information and application scenario information from the pesticide effect database, and constructs a network of interactions among plant characteristics, environmental characteristics, and pesticide characteristics, wherein the single-component pesticide information is diuron and mesotrione; The ratio of diuron and foramsulfuron is analyzed through the environment-pharmaceutical-plant interaction network, and a preparation scheme of the herbicidal composition is obtained according to the analysis results; Constructing a scheme evaluation model, wherein the scheme evaluation model evaluates the preparation scheme of the herbicidal composition based on historical experimental data, and optimizes the preparation scheme of the herbicide according to the evaluation result; Establishing the drug effect database includes: Collecting drug information of diuron and foramsulfuron, and performing first classification management on the single-ingredient drugs according to the adaptability of the single-ingredient drugs to the application scenario information; Collecting plant information, including target plant information and non-target plant information, and performing second classification management on the sensitivity of the single-component agent according to the plant information; Establishing a first classification management table and a second classification management table respectively according to the first classification management and the second classification management, and setting a medicine label corresponding to the single-component medicine; Establishing indexes on the drug label fields of the first classification management table and the second classification management table, implementing cross-search through joint query, and establishing stored procedures and related functions to simplify the query process; Construct a program evaluation model, including: Collecting historical herbicidal composition preparation data information, and extracting application scenario information and target plant information and non-target plant information of the corresponding historical herbicidal composition preparation scheme according to application analysis, wherein the historical herbicidal composition preparation data information includes the performance of diuron and foramsulfuron on target plants, non-target plants and application scenarios, respectively, and the performance of the diuron and foramsulfuron ratio on target plants, non-target plants and application scenarios; Cleaning the historical herbicidal composition preparation data information, and integrating the cleaned agent information, the application scenario information, the target plant information, and the non-target plant information of the historical herbicidal composition preparation data information into a historical sample set according to the implementation project; Selecting a deep learning method to learn the historical sample set to obtain the correlation between elements in the historical sample set; An input port and an output port are established according to the learning result, wherein the numerous input ports input the application scenario information, the target plant information and the non-target plant information, and the output port outputs the herbicidal composition preparation scheme; The scheme evaluation model is trained and validated, including: constructing a training set and a validation set according to the historical herbicidal composition preparation data information; The scheme evaluation model is trained by the training set, and the training result is verified by the verification set; The scheme evaluation model is optimized according to the training and verification results.

2. The method for preparing a herbicidal composition based on application analysis according to claim 1, characterized in that: Based on the drug effect database, the environment-drug-plant interaction network is constructed, including: Retrieving the single-ingredient medicine related to the application scenario information from the medicine effect database according to the application analysis; Establishing a plant-drug interaction relationship, taking the plant information, diuron and foramsulfuron as network nodes, and establishing strong and weak network connections according to the sensitivity and toxicity of the single-ingredient pesticide to the target plant information and non-target plant information in the plant information; Setting edge weights of the plant-drug network, wherein the edge weights represent the absorption and metabolism capabilities of the plant to the single-ingredient drug; The application scenario information is added to the plant-drug network, and the attributes of the network nodes are added according to the application scenario information, and the attributes of the edge weights are added according to the impact of the application scenario information on the plant-drug network.

3. The method for preparing a herbicidal composition based on application analysis according to claim 1 or 2, characterized in that: The application of the herbicide composition is analyzed through the environment-agent-plant interaction network, and a preparation scheme of the herbicide composition is obtained according to the analysis results, including: Using complex network theory to identify the structural characteristics of the environment-pharmaceutical-plant interaction network, including the degree distribution of nodes, clustering coefficient, and small-world properties of the network; According to the structural characteristics, the centrality of the environment-agent-plant interaction network is analyzed, the node sensitivity is set, and the ratio and dosage of diuron and mesotrione are selected according to the node sensitivity to form a herbicide composition preparation plan.

4. The method for preparing a herbicidal composition based on application analysis according to claim 3, characterized in that: The centrality of the environment-pharmaceutical-plant interaction network was analyzed, including: Calculate the degree of each node, and determine the degree centrality of the node according to the degree calculation result, wherein the degree of the node is the number of edges connected to the node; Calculate the degree of each node as an intermediary, and determine the betweenness of the node according to the degree of intermediary, wherein the degree of intermediary is the frequency of passing through the node on the path connecting other nodes; Calculate the closeness centrality of the nodes, which is the distance between each node and other nodes, indicating the rapid influence of the node on the network; The eigenvector centrality is calculated, which is the weight of a node connecting to other nodes and the centrality of the other nodes.

5. The method for preparing a herbicidal composition based on application analysis according to claim 1, characterized in that: The historical herbicidal composition preparation data information in the scheme evaluation model is regularly updated, and the updated historical herbicidal composition preparation data information is relearned by the deep learning method.

6. A system for preparing a herbicidal composition based on application analysis, characterized in that: The method for preparing a herbicidal composition based on application analysis as claimed in claim 1, wherein the system comprises: An information acquisition module, which acquires plant information and application scenario information, wherein the plant information includes target plant information and non-target plant information; An interaction network generation module is used to construct an environment-drug-plant interaction network based on a drug effect database, wherein the environment-drug-plant interaction network retrieves single-component drug information related to the plant information and application scenario information from the drug effect database, and constructs a network of interactions among plant characteristics, environmental characteristics, and drug characteristics, wherein the single-component drug information is diuron and mesotrione; A preparation scheme generating module, which analyzes the ratio of diuron and foramsulfuron through the environment-agent-plant interaction network, and obtains a preparation scheme of the herbicidal composition according to the analysis result; A preparation scheme evaluation module is used to construct a scheme evaluation model. The scheme evaluation model evaluates the preparation scheme of the herbicidal composition based on historical experimental data and optimizes the preparation scheme of the herbicide according to the evaluation result.

7. The herbicidal composition preparation system based on application analysis according to claim 6, characterized in that: The interaction network generation module comprises: An information retrieving unit, for retrieving the single-ingredient medicine related to the application scenario information from the medicine effect database according to the application analysis; An interaction relationship establishment unit is used to establish a plant-drug interaction relationship, taking the plant information, diuron and foramsulfuron as network nodes, and establishing strong and weak network connections according to the sensitivity and toxicity of the single-ingredient pesticide to the target plant information and non-target plant information in the plant information; An edge weight setting unit, which sets the edge weight of the plant-drug network, wherein the edge weight represents the absorption and metabolism ability of the plant to the single-ingredient drug; A scenario information adding unit adds the application scenario information to the plant-pharmaceutical network, adds attributes of network nodes according to the application scenario information, and adds attributes of edge weights according to the impact of the application scenario information on the plant-pharmaceutical network.

Images