Recommendation method and system for cultivation improvement means to enhance heat tolerance of rice

Abstract

The present invention discloses a method and a system for recommending cultivation improvement means for improving the heat tolerance of rice, including: obtaining multi-source environmental parameters of a target rice planting area for heat damage early warning of the target rice planting area; constructing a mapping relationship between high-temperature stress and morphological traits corresponding to different rice growth stages, and when a heat damage early warning is generated, reading the morphological trait indexes of the current rice growth stage according to the mapping relationship; obtaining the current high-temperature characteristics of the target rice planting area according to the read morphological trait indexes, and using a knowledge graph related to high-temperature stress defense measures to obtain a recommended path including cultivation improvement measures; and determining an optimal combination of cultivation improvement measures according to soil environmental characteristics based on the cultivation improvement measures for recommended output. The present invention formulates scientific and reasonable cultivation improvement means for high-temperature heat damage of rice, alleviates the problem of rice yield reduction caused by high-temperature weather, and ensures high and stable yields of rice.

Description

Technical Field

[0001] The present invention relates to the technical field of rice cultivation, and more specifically, to a method and system for recommending cultivation improvement means for improving the heat tolerance of rice. Background Art

[0002] With the large-scale emission of greenhouse gases such as carbon dioxide, methane and nitrogen oxides, the trend of global warming continues to intensify, and food production is greatly affected. As one of the important food crops, high temperature directly affects the stable and high yield of rice. The suitable temperature for rice seedling stage is 25.0 - 30.0 °C. When rice is in the flowering stage, when the environmental temperature rises to 37.0 - 42.0 °C, the seed setting rate of rice will drop significantly. When rice tillering stage suffers from high temperature stress, the tiller number and effective panicle number both decrease. The heading and flowering stage is most susceptible to high temperature. Under high temperature stress, it will cause reduced spikelet fertility, anther dehiscence, abnormal pollen tube growth, etc., thus reducing the seed setting rate, 1000-grain weight and yield of rice.

[0003] The current countermeasures to deal with the impact of high temperature weather on rice yield mainly include the following aspects: 1) Selection of heat-tolerant rice varieties; 2) Application of plant growth regulators; 3) Sowing in stages. At present, in the production of field rice, single measures are mainly adopted to deal with high temperature heat damage. For example, the method of delaying sowing is used to avoid the high temperature during the flowering stage of rice, but this will increase the probability of encountering high temperature during the panicle differentiation stage of rice. In addition, although the application of heat-tolerant varieties and plant growth regulators can reduce high temperature heat damage to a certain extent, it is still difficult to meet the needs of production. Because some varieties are more tolerant to high temperature during the flowering stage, while some varieties are less sensitive to high temperature during the filling stage, it is difficult to select varieties. Plant growth regulators are not applicable to all high temperature stress situations. Therefore, how to scientifically and reasonably select the cultivation improvement countermeasures for rice under high temperature stress to relieve high temperature and improve spikelet fertility is an urgent problem to be solved. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a method and system for recommending cultivation improvement means for improving the heat tolerance of rice, aiming to formulate scientific and reasonable cultivation improvement means for rice high temperature heat damage and regulate the heat tolerance of rice plants.

[0005] The first aspect of the present invention provides a method for recommending cultivation improvement means for improving the heat tolerance of rice, including:

[0006] Obtaining multi-source environmental parameters of the target rice planting area, obtaining heat damage monitoring parameters of different rice growth stages according to the preprocessed multi-source environmental parameters, and performing heat damage early warning on the target rice planting area through the heat damage monitoring parameters;

[0007] Screen morphological traits related to high - temperature stress at different rice growth stages, and construct corresponding mapping relationships. When a heat damage warning is generated, read the morphological trait indicators of the current rice growth stage according to the mapping relationship;

[0008] Obtain the current high - temperature characteristics of the target rice planting area according to the read morphological trait indicators, connect the high - temperature characteristics to the knowledge graph related to high - temperature stress defense measures, and obtain a recommended path including cultivation improvement measures according to the knowledge graph;

[0009] Extract the included cultivation improvement measures based on the recommended path, and determine a combination of cultivation improvement measures according to the soil environmental characteristics of the target rice planting area as the recommended result of the cultivation improvement measures.

[0010] In this solution, heat damage monitoring parameters at different rice growth stages are obtained according to pre - processed multi - source environmental parameters, and heat damage warning for the target rice planting area is carried out through the heat damage monitoring parameters. Specifically:

[0011] Divide the target rice planting area into grids, match the pre - processed multi - source environmental parameters with the grid blocks, calculate the similarity between the data blocks and the neighborhood data blocks after calculating the number of matches. If the similarity is less than the preset similarity threshold, it is regarded as the same category area;

[0012] After all grid blocks are compared, obtain the area division of the target rice planting area, aggregate all multi - source environmental parameters within the area block, and retrieve historical heat damage data according to the rice variety information of the target rice planting area;

[0013] Cluster the retrieved historical heat damage data based on the rice growth stage, obtain the occurrence frequency of heat damage data at different rice growth stages, and normalize the occurrence frequency of the heat damage data to read the monitoring coefficients at different rice growth stages;

[0014] Compare the aggregated multi - source environmental parameters with the multi - source environmental parameters under the suitable conditions at the corresponding rice growth stage to calculate the environmental parameter deviation, identify abnormal environmental parameters through the environmental parameter deviation, and perform weighted summation of the environmental parameter deviation of the abnormal environmental parameters using the monitoring coefficient to generate a heat damage score;

[0015] If there are no abnormal environmental parameters, then analyze the change law according to the aggregated multi - source environmental parameters for prediction, identify abnormal environmental parameters using the multi - source environmental parameters after a preset time, calculate the heat damage score, and carry out heat damage warning for the area block through the heat damage score.

[0016] In this solution, screen morphological traits related to high - temperature stress at different rice growth stages, and construct corresponding mapping relationships. Specifically:

[0017] Using a big data search engine to obtain the impact instance of high temperature stress on rice plants in a preset search space, reading the rice plant organs involved from the impact instance, and soft clustering the rice plant organs according to different rice growth stages;

[0018] Obtaining a subset of rice plant organs related to high temperature stress at different rice growth stages, performing principal component analysis on the subset of rice plant organs, and obtaining a variance contribution rate corresponding to the principal component to represent the importance, and screening the characteristic rice plant organs at different rice growth stages when subjected to high temperature stress based on the importance;

[0019] According to the impact example, the morphological trait data sequences of the iconic rice plant organs at different rice growth stages before and after being subjected to high temperature stress are read, the low-dimensional data distribution of the two morphological trait data sequences is obtained, and the characteristic scatter plots corresponding to the two low-dimensional data distributions are compared and analyzed;

[0020] The data feature points in the feature scatter plot whose position deviation is greater than a preset distance threshold are read, and a mapping relationship is established according to the organ structure characteristics and high temperature stress corresponding to the data feature points.

[0021] In this scheme, the current high temperature characteristics of the target rice planting area are obtained based on the read morphological trait indicators, specifically:

[0022] Obtaining the rice growth stage of the rice plants in the regional block with heat damage warning in the target rice planting area, reading the corresponding morphological trait index based on the mapping relationship, and reading the index parameters of the rice plants in the regional block based on the morphological trait index;

[0023] Constructing a parameter matrix according to the index parameters, calculating the residual vector between the parameter matrix and the benchmark parameter matrix corresponding to the morphological trait index under normal conditions, and establishing a corresponding relationship between the degree of high temperature stress and the change of morphological traits based on impact example training;

[0024] The residual vector is further represented by the corresponding relationship to obtain a high temperature feature representing the severity of heat damage to the regional block.

[0025] In this solution, the high temperature characteristics are connected to the knowledge graph related to high temperature stress defense measures, and the recommended path containing cultivation improvement measures is obtained according to the knowledge graph, specifically:

[0026] According to the example of the impact of high temperature stress on rice plants, several knowledge graphs related to high temperature stress defense measures are obtained, and a preset number of relevant knowledge graphs are selected based on the richness of entities and relationships in the relevant knowledge graphs;

[0027] Construct an interaction matrix of heat damage situations and high-temperature stress defense measures in each relevant knowledge graph, and use a graph convolutional network to perform representation learning on the interaction matrix to obtain the embedding vectors of entities and relationships in the relevant knowledge graph;

[0028] Initialize a random walk model, perform entity localization according to the high-temperature characteristics, use the located entity as the starting point, use cultivation improvement measures as entity category constraints, combine length constraints to generate random walk rules, and sample paths containing cultivation improvement measures based on the random walk rules;

[0029] Introduce a federated learning framework, use the random walk model in each relevant knowledge graph as a local model, generate a confidence score for the local model according to the average improvement effect of the obtained cultivation improvement measures, and upload and aggregate the local models with the confidence score greater than the preset threshold to construct a global model;

[0030] Use the global model of ensemble learning to obtain recommended paths containing cultivation improvement measures in the relevant knowledge graph.

[0031] In this solution, the cultivation improvement measures included in the recommended path are extracted as follows:

[0032] Construct corresponding embedding vector sequences according to the cultivation improvement measures of different recommended paths, perform path encoding on the embedding vector sequences based on Bi-LSTM, and generate an embedding matrix of cultivation improvement measures;

[0033] Use the attention mechanism to generate a weight vector for the embedding matrix, and perform weighting using the weight vector to obtain the encoded representation of the cultivation improvement measures.

[0034] In this solution, the cultivation improvement measure combination is determined according to the soil environmental characteristics of the target rice planting area as the recommended result of the cultivation improvement measures, specifically:

[0035] Based on multi-source environmental parameters, analyze the soil environmental characteristics of the heat damage warning area blocks in the target rice planting area, and use a multi-layer perceptron to learn the encoded representation of the cultivation improvement measures and the soil environmental characteristics;

[0036] Combine the activation function operation to obtain the interaction prediction probability of the heat damage warning area block for the cultivation improvement measures, sort the cultivation improvement measures according to the interaction prediction probability, and sequentially select the cultivation improvement measures according to the preset quantity to determine the cultivation improvement measure combination.

[0037] The second aspect of the present invention provides a cultivation improvement measure recommendation system for improving the heat tolerance of rice, including: a multi-source environmental data collection module, a heat damage warning module, a rice morphological analysis module, a knowledge graph module, a cultivation improvement measure evaluation module, and a recommendation output module;

[0038] The multi-source environmental data acquisition module is responsible for collecting multi-source environmental parameters of the target rice planting area and generating heat damage monitoring parameters for different rice growth stages;

[0039] The heat damage early warning module is responsible for carrying out heat damage early warning for the target rice planting area by analyzing the obtained heat damage monitoring parameters;

[0040] The rice morphological trait analysis module is responsible for constructing the mapping relationship between high-temperature stress and rice morphological traits. When a heat damage early warning is generated, it reads the morphological trait indicators of the current rice growth stage according to the mapping relationship and evaluates the current high-temperature characteristics of the target rice planting area;

[0041] The knowledge graph module is responsible for connecting the analyzed high-temperature characteristics to the knowledge graph related to high-temperature stress defense measures to obtain a recommended path including cultivation improvement measures;

[0042] The cultivation improvement measure evaluation module is responsible for evaluating the cultivation improvement measures included in the extracted recommended path and determining the combination of cultivation improvement measures according to the soil environmental characteristics of the target rice planting area;

[0043] The recommended output module is responsible for taking the combination of cultivation improvement measures as the recommended result and visually displaying the recommended result through a preset method.

[0044] Compared with the prior art, the beneficial effects of the present disclosure are:

[0045] The present invention analyzes the damage mechanism of high-temperature stress on rice floret fertility, formulates scientific and reasonable cultivation improvement measures for rice high-temperature heat damage, alleviates high temperature, improves floret fertility, regulates the heat tolerance of rice plants, and through the recommended combination of cultivation improvement measures, alleviates the problem of rice yield reduction caused by high-temperature weather and ensures high and stable rice yields. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments or exemplary of the present invention, the following will briefly introduce the drawings required for use in the embodiments or exemplary descriptions. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the drawings shown.

[0047] Figure 1 Shows the flowchart of the cultivation improvement measure recommendation method for improving the heat tolerance of rice in the present invention;

[0048] Figure 2 Shows the flowchart of screening morphological trait indicators related to high-temperature stress in the embodiments of the present invention;

[0049] Figure 3The flowchart showing the recommended path for obtaining cultivation improvement measures according to an embodiment of the present invention is shown;

[0050] Figure 4 The block diagram showing the recommended system for cultivation improvement means to improve the heat tolerance of rice according to the present invention is shown. Detailed implementation manners

[0051] In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.

[0052] Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention may be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0053] Figure 1 The flowchart showing the recommended method for cultivation improvement means to improve the heat tolerance of rice according to the present invention is shown.

[0054] As Figure 1 shown, in the first embodiment of the present invention, a recommended method for cultivation improvement means to improve the heat tolerance of rice is provided, including:

[0055] S102, obtaining multi-source environmental parameters of the target rice planting area, obtaining heat damage monitoring parameters at different rice growth stages according to the preprocessed multi-source environmental parameters, and performing heat damage early warning on the target rice planting area through the heat damage monitoring parameters;

[0056] S104, screening morphological traits related to high temperature stress in different rice growth stages and constructing corresponding mapping relationships. When a heat damage early warning is generated, the morphological trait indexes of the current rice growth stage are read according to the mapping relationships;

[0057] S106, obtaining the current high temperature characteristics of the target rice planting area according to the read morphological trait indexes, connecting the high temperature characteristics to the knowledge graph related to high temperature stress defense measures, and obtaining a recommended path including cultivation improvement measures according to the knowledge graph;

[0058] S108, extracting the included cultivation improvement measures based on the recommended path, and determining a combination of cultivation improvement measures according to the soil environmental characteristics of the target rice planting area as the recommended result of the cultivation improvement measures.

[0059] It should be noted that the climate data, temperature and humidity data, soil physical and chemical properties and other multi-source environmental parameters of the target rice-growing area are obtained, the target rice-growing area is divided into grids, the pre-processed multi-source environmental parameters are matched with the grid blocks, and the similarity between the matching data blocks and the neighborhood data blocks is calculated. If the similarity is less than the preset similarity threshold, they are regarded as the same category area; when all grid fast comparisons are completed, the regional division of the target rice-growing area is obtained, all multi-source environmental parameters in the regional block are aggregated, and the historical heat damage data is retrieved according to the rice variety information of the target rice-growing area; the retrieved historical heat damage data are clustered based on the rice growth stage to obtain different rice growth stages. The frequency of heat damage data in the corresponding rice growth stage is normalized to read the monitoring coefficient of different rice growth stages; the environmental parameter deviation is calculated by comparing the aggregated multi-source environmental parameters with the multi-source environmental parameters under the suitable growth conditions of the corresponding rice growth stage, the abnormal environmental parameters are identified by the environmental parameter deviation, the environmental parameter deviation of the abnormal environmental parameters is weightedly summed by using the monitoring coefficient to generate a heat damage score; if there is no abnormal environmental parameter, a prediction is made based on the change law of the aggregated multi-source environmental parameters, the abnormal environmental parameters are identified by using the multi-source environmental parameters after a preset time, the heat damage score is calculated, and a heat damage warning is issued for the regional block according to the heat damage score.

[0060] Figure 2 A flow chart of screening morphological trait indicators related to high temperature stress according to an embodiment of the present invention is shown.

[0061] According to an embodiment of the present invention, morphological traits related to high temperature stress are screened in different rice growth stages, and corresponding mapping relationships are constructed, specifically:

[0062] S202, using a big data search engine to obtain an impact instance of high temperature stress on rice plants in a preset search space, reading the rice plant organs involved from the impact instance, and soft clustering the rice plant organs according to different rice growth stages;

[0063] S204, obtaining a subset of rice plant organs related to high temperature stress at different rice growth stages, performing principal component analysis on the subset of rice plant organs, and obtaining a variance contribution rate corresponding to the principal component to represent an importance degree, and screening representative rice plant organs at different rice growth stages when subjected to high temperature stress based on the importance degree;

[0064] S206, reading the morphological trait data sequences of the iconic rice plant organs at different rice growth stages before and after being subjected to high temperature stress according to the impact examples, obtaining low-dimensional data distributions of the two morphological trait data sequences, and performing comparative analysis on the characteristic scatter plots corresponding to the two low-dimensional data distributions;

[0065] S208, Read data feature points in the feature scatter plot with a position deviation greater than a preset distance threshold, and establish a mapping relationship based on the organ structure features and high-temperature stress corresponding to the data feature points.

[0066] It should be noted that when rice is under high-temperature stress during the tillering stage, both the tiller number and the effective panicle number decrease. The heading and flowering stage is most susceptible to high temperature. Under high-temperature stress, it will cause a decrease in spikelet fertility, anther dehiscence, abnormal pollen tube growth, etc. According to the heat-sensitive periods such as the tillering stage and the heading and flowering stage, obtain examples of the impact of high-temperature stress on rice plants. Obtain subsets of rice plant organs related to high-temperature stress in different rice growth stages. For example, high temperature during the heading and flowering stage of rice will affect the formation of spikelets, pollen development, the structure of stamens and pistils, flowering and fertilization, and spikelet physiological metabolism. High temperature will increase the viscosity of pollen, resulting in difficulty in pollen dispersal in the anther, and the filaments of spikelets will lose water and wilt, which will also cause hindrance to spikelet flowering. Observing the shape and state of anthers and spikelets can reflect the heat damage of rice to a certain extent.

[0067] It should be noted that obtain the rice growth stage of rice plants in the area blocks with heat damage warnings in the target rice planting area, read the corresponding morphological trait indicators based on the mapping relationship, and read the index parameters of the rice plants in the area blocks based on the morphological trait indicators; construct a parameter matrix according to the index parameters, calculate the residual vector between the parameter matrix and the benchmark parameter matrix corresponding to the morphological trait indicators in the normal state, and establish a corresponding relationship between the high-temperature stress degree and the morphological trait changes based on the training of impact examples. For example, when the temperature is higher, the anther dehiscence of rice is smaller, that is, the degree of anther dehiscence obstruction is greater, which proves that the high-temperature stress degree is higher; further represent the residual vector through the corresponding relationship to obtain the high-temperature feature characterizing the severity of heat damage in the area block.

[0068] Figure 3 The flowchart showing the recommended path for obtaining cultivation improvement measures according to the embodiments of the present invention is shown.

[0069] According to the embodiments of the present invention, connect the high-temperature feature to the knowledge graph related to high-temperature stress defense measures, and obtain the recommended path including cultivation improvement measures according to the knowledge graph. Specifically:

[0070] S302, obtain a number of knowledge graphs related to high-temperature stress defense measures according to the impact examples of high-temperature stress on rice plants, and select a preset number of related knowledge graphs based on the richness of entities and relationships in the related knowledge graphs;

[0071] S304, construct an interaction matrix between the heat damage situation and high-temperature stress defense measures in each related knowledge graph, and use a graph convolutional network to perform representation learning on the interaction matrix to obtain the embedding vectors of entities and relationships in the related knowledge graphs;

[0072] S306. Initialize the random walk model, locate entities based on the high-temperature characteristics, use the located entities as the starting points, take the cultivation improvement measures as the entity category constraints, generate random walk rules in combination with length constraints, and sample paths containing cultivation improvement measures based on the random walk rules;

[0073] S308. Introduce the federated learning framework, use the random walk models in each relevant knowledge graph as local models, generate confidence scores for the local models according to the average improvement effects of the obtained cultivation improvement measures, and upload and aggregate the local models with confidence scores greater than the preset threshold to construct a global model;

[0074] S310. Use the global model of ensemble learning to obtain recommended paths containing cultivation improvement measures in the relevant knowledge graph.

[0075] It should be noted that the random walk cultivation improvement measure recommendation model is used as the global model. Preferably, the random walk can be a meta-path model, an RWNN model, etc. The training parameters of each received local model will be aggregated into new parameter information and passed into the global model as its initial parameters. The validation set is trained and verified. The global model parameters in each round of training are obtained by taking the difference between the received local model parameters and the previous round of global model parameters in turn and then averaging while retaining part of the previous round of parameter information. After training, the global model is further optimized for parameters, and the optimized model parameter information is then passed to each local model, and then the updated local model parameters are passed back to the global model of the terminal server.

[0076] It should be noted that corresponding embedding vector sequences are constructed for the cultivation improvement measures of different recommended paths. Based on Bi-LSTM, the embedding vector sequences are encoded for paths respectively from the forward and backward directions, and an embedding matrix of the cultivation improvement measures is generated according to the hidden layer features; the attention mechanism is used to generate a weight vector a for the embedding matrix. w1 and w2 are trainable matrices. is the transpose of the embedding matrix L q ; weighted by the weight vector to obtain the encoded representation of the cultivation improvement measures. Preferably, corresponding embedding vector sequences are constructed for the cultivation improvement measures of different recommended paths, the cultivation improvement measures used in the target planting area in history are obtained, the embedding vector sequences are spliced, and the spliced vector sequences are used for path encoding to improve the accuracy of recommendation.

[0077] Based on the analysis of the soil environmental characteristics of the heat damage warning area in the target rice planting area using multi-source environmental parameters, a multi-layer perceptron is used to learn the encoded representation L of the cultivation improvement measures and the soil environmental characteristics T; the interaction prediction probability y of the heat damage warning area for the cultivation improvement measures is obtained through the operation of the activation function, expressed as: y = σ(MLP(L T)), where σ represents the activation function. The cultivation improvement measures are sorted according to the interaction prediction probability, and the cultivation improvement measures are sequentially selected according to the preset quantity to determine the cultivation improvement measure combination to improve the heat resistance of rice plants.

[0078] Figure 4 The block diagram of the cultivation improvement means recommendation system for improving the heat resistance of rice according to the present invention is shown.

[0079] The second embodiment of the present invention provides a cultivation improvement means recommendation system 4 for improving the heat resistance of rice, including a multi-source environmental data acquisition module 41, a heat damage warning module 42, a rice shape and state analysis module 43, a knowledge graph module 44, a cultivation improvement measure evaluation module 45, and a recommendation output module 46;

[0080] The multi-source environmental data acquisition module is responsible for collecting multi-source environmental parameters of the target rice planting area and generating heat damage monitoring parameters for different rice growth stages;

[0081] The heat damage warning module is responsible for performing heat damage warning on the target rice planting area by analyzing the obtained heat damage monitoring parameters;

[0082] The rice shape and state analysis module is responsible for constructing the mapping relationship between high-temperature stress and rice morphological traits. When a heat damage warning is generated, the morphological trait indicators of the current rice growth stage are read according to the mapping relationship to evaluate the current high-temperature characteristics of the target rice planting area;

[0083] The knowledge graph module is responsible for connecting the analyzed high-temperature characteristics to the knowledge graph related to high-temperature stress defense measures to obtain a recommended path including cultivation improvement measures;

[0084] The cultivation improvement measure evaluation module is responsible for evaluating the cultivation improvement measures included in the recommended path extraction and determining the cultivation improvement measure combination according to the soil environmental characteristics of the target rice planting area;

[0085] The recommendation output module is responsible for taking the cultivation improvement measure combination as the recommendation result and visually displaying the recommendation result in a preset manner.

[0086] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed with each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be electrical, mechanical, or other forms.

[0087] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0088] In addition, each functional unit in the embodiments of the present invention can be all integrated in a processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware, or in the form of hardware plus software functional units.

[0089] Those of ordinary skill in the art can understand that all or part of the steps to implement the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments. The foregoing storage media include: removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks and other various media that can store program codes.

[0090] Alternatively, if the above-mentioned integrated units of the present invention are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art 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 enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The foregoing storage media include: removable storage devices, ROM, RAM, magnetic disks, or optical disks and other various media that can store program codes.

[0091] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claimed rights.

Claims

1. A recommended method for improving the heat tolerance of rice, characterized in that: The following steps are involved: Acquire multi-source environmental parameters of the target rice planting area, acquire heat damage monitoring parameters at different rice growth stages according to the pre-processed multi-source environmental parameters, and perform heat damage early warning for the target rice planting area through the heat damage monitoring parameters; Screening morphological traits related to high temperature stress in different rice growth stages and constructing corresponding mapping relationships, and when a heat damage warning is generated, reading the morphological trait indicators of the current rice growth stage according to the mapping relationship; Obtain the current high temperature characteristics of the target rice planting area according to the read morphological trait indicators, connect the high temperature characteristics to the knowledge graph related to high temperature stress defense measures, and obtain a recommended path including cultivation improvement measures according to the knowledge graph; Extracting the cultivation improvement measures included in the recommended path, and determining a combination of cultivation improvement measures according to soil environmental characteristics of the target rice planting area as a recommendation result of the cultivation improvement measures; The morphological traits related to high temperature stress were screened at different rice growth stages, and the corresponding mapping relationships were constructed, specifically: Using a big data search engine to obtain the impact instance of high temperature stress on rice plants in a preset search space, reading the rice plant organs involved from the impact instance, and soft clustering the rice plant organs according to different rice growth stages; Obtaining a subset of rice plant organs related to high temperature stress at different rice growth stages, performing principal component analysis on the subset of rice plant organs, and obtaining a variance contribution rate corresponding to the principal component to represent the importance, and screening the characteristic rice plant organs at different rice growth stages when subjected to high temperature stress based on the importance; According to the impact example, the morphological trait data sequences of the iconic rice plant organs at different rice growth stages before and after being subjected to high temperature stress are read, the low-dimensional data distribution of the two morphological trait data sequences is obtained, and the characteristic scatter plots corresponding to the two low-dimensional data distributions are compared and analyzed; Reading data feature points in the feature scatter plot whose position deviation is greater than a preset distance threshold, and establishing a mapping relationship according to the organ structure characteristics and high temperature stress corresponding to the data feature points; The high temperature characteristics are connected to the knowledge graph related to high temperature stress defense measures, and the recommended path including cultivation improvement measures is obtained according to the knowledge graph, specifically: According to the example of the impact of high temperature stress on rice plants, several knowledge graphs related to high temperature stress defense measures are obtained, and a preset number of relevant knowledge graphs are selected based on the richness of entities and relationships in the relevant knowledge graphs; Construct an interaction matrix of heat damage and high temperature stress defense measures in each relevant knowledge graph, use a graph convolutional network to perform representation learning on the interaction matrix, and obtain embedding vectors of entities and relationships in the relevant knowledge graph; Initialize the random walk model, locate the entity according to the high temperature feature, use the located entity as the starting point, use the cultivation improvement measures as entity category constraints, combine the length constraints to generate random walk rules, and sample the path containing the cultivation improvement measures based on the random walk rules; Introducing a federated learning framework, the random walk model in each relevant knowledge graph is used as a local model, and the confidence score of the local model is generated according to the average improvement effect of the acquired cultivation improvement measures. The local models with confidence scores greater than a preset threshold are uploaded and aggregated to construct a global model. The global model of ensemble learning is used to obtain the recommended paths containing cultivation improvement measures in the relevant knowledge graph.

2. The method for improving the heat resistance of rice according to claim 1, characterized in that: The heat damage monitoring parameters at different rice growth stages are obtained based on the pre-processed multi-source environmental parameters, and the heat damage early warning of the target rice planting area is carried out based on the heat damage monitoring parameters, specifically: The target rice planting area is divided into grids, and the pre-processed multi-source environmental parameters are used to match the grid blocks. The similarity between the matched data blocks and the neighboring data blocks is calculated. If the similarity is less than the preset similarity threshold, they are regarded as the same category areas; When all grid blocks are compared, the regional division of the target rice planting area is obtained, all multi-source environmental parameters in the regional block are aggregated, and historical heat damage data are retrieved according to the rice variety information of the target rice planting area; Clustering the retrieved historical heat damage data based on the rice growth stage, obtaining the occurrence frequency of heat damage data at different rice growth stages, and normalizing the occurrence frequency of heat damage data to read the monitoring coefficients at different rice growth stages; Comparing the aggregated multi-source environmental parameters with the multi-source environmental parameters under the suitable growth conditions of the corresponding rice growth stage to calculate the environmental parameter deviation, identifying the abnormal environmental parameters through the environmental parameter deviation, and performing weighted summation of the environmental parameter deviations of the abnormal environmental parameters using the monitoring coefficient to generate a heat damage score; If there are no abnormal environmental parameters, a prediction is made based on the change rules of the aggregated multi-source environmental parameters, and the multi-source environmental parameters after a preset time are used to identify abnormal environmental parameters, and a heat damage score is calculated. The heat damage score is used to issue a heat damage warning for the regional block.

3. The method for improving the heat resistance of rice according to claim 1, characterized in that: The current high temperature characteristics of the target rice planting area are obtained based on the read morphological trait indicators, specifically: Obtaining the rice growth stage of the rice plants in the regional block with heat damage warning in the target rice planting area, reading the corresponding morphological trait index based on the mapping relationship, and reading the index parameters of the rice plants in the regional block based on the morphological trait index; Constructing a parameter matrix according to the index parameters, calculating the residual vector between the parameter matrix and the benchmark parameter matrix corresponding to the morphological trait index under normal conditions, and establishing a corresponding relationship between the degree of high temperature stress and the change of morphological traits based on impact example training; The residual vector is further represented by the corresponding relationship to obtain a high temperature feature representing the severity of heat damage to the regional block.

4. The method for improving the heat tolerance of rice according to claim 1, characterized in that: The cultivation improvement measures included in the extraction based on the recommended path are specifically: Construct corresponding embedding vector sequences according to the cultivation improvement measures of different recommended paths, perform path encoding on the embedding vector sequences based on Bi-LSTM, and generate an embedding matrix of the cultivation improvement measures; The attention mechanism is used to generate a weight vector for the embedding matrix, and the weight vector is used for weighting to obtain the encoding representation of the cultivation improvement measures.

5. The method for recommending cultivation improvement means for improving the heat tolerance of rice according to claim 1, characterized in that: According to the soil environment characteristics of the target rice planting area, a combination of cultivation improvement measures is determined as the recommended results of cultivation improvement measures, specifically: Based on multi-source environmental parameters, the soil environmental characteristics of the heat damage warning area in the target rice planting area are analyzed, and the encoding representation of cultivation improvement measures and soil environmental characteristics are learned using a multi-layer perceptron. The interactive prediction probability of the heat damage warning area block for the cultivation improvement measures is obtained in combination with the activation function operation, the cultivation improvement measures are sorted according to the interactive prediction probability, and the cultivation improvement measures are selected in sequence according to a preset number to determine the combination of cultivation improvement measures.

6. A cultivation improvement method recommendation system for improving the heat tolerance of rice, characterized in that: A method for recommending cultivation improvement measures for improving heat resistance of rice as claimed in any one of claims 1 to 5, comprising a multi-source environmental data acquisition module, a heat damage early warning module, a rice shape and state analysis module, a knowledge graph module, a cultivation improvement measure evaluation module and a recommendation output module; The multi-source environmental data acquisition module is responsible for collecting multi-source environmental parameters of the target rice planting area and generating heat damage monitoring parameters at different rice growth stages; The heat damage early warning module is responsible for providing a heat damage early warning for the target rice planting area by analyzing the acquired heat damage monitoring parameters; The rice shape and property analysis module is responsible for constructing a mapping relationship between high temperature stress and rice morphological traits. When a heat damage warning is generated, the morphological trait indicators of the current rice growth stage are read according to the mapping relationship to evaluate the current high temperature characteristics of the target rice planting area. The knowledge graph module is responsible for connecting the high temperature characteristics obtained through analysis to the knowledge graph related to high temperature stress defense measures to obtain the recommended path including cultivation improvement measures; The cultivation improvement measure evaluation module is responsible for evaluating the cultivation improvement measures included in the recommended path extraction and determining the combination of cultivation improvement measures according to the soil environmental characteristics of the target rice planting area; The recommendation output module is responsible for combining the cultivation improvement measures as recommendation results, and visually displaying the recommendation results in a preset manner.

Images