Cultivation improvement means recommendation method and recommendation system for improving heat resistance of rice

Abstract

The invention discloses a cultivation improvement means recommendation method and recommendation system for improving heat resistance of rice. The method comprises the following steps: acquiring multi-source environmental parameters of a target rice planting area, and carrying out heat damage early warning on the target rice planting area; constructing a mapping relation corresponding to high temperature stress and morphological characters in different rice growth stages, and when a heat damage early warning is generated, reading morphological character indexes of the current rice growth stage according to the mapping relation; obtaining current high-temperature characteristics of the target rice planting area according to the read morphological character indexes, and obtaining a recommended path containing cultivation improvement measures by using a high-temperature stress defense measure related knowledge graph; and determining an optimal cultivation improvement measure combination according to the soil environment characteristics based on the cultivation improvement measures, and performing recommendation output. Scientific and reasonable cultivation improvement means are made for rice high-temperature heat damage, the problem of rice yield reduction caused by high-temperature weather is relieved, and high and stable yield of rice is ensured.

Description

Technical Field

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

[0002] With the massive 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. 25.0-30.0℃ is the suitable temperature for rice seedlings. When rice is in the flowering stage, the ambient temperature rises to 37.0-42.0℃, and the rice fruit setting rate will drop significantly. Rice suffers from high temperature stress during the tillering stage, and its tiller number and effective panicle number are reduced. The heading and flowering stage is most susceptible to high temperature. High temperature stress will cause the fertility of spikelets to decrease, anthers to crack, and pollen tube growth to be abnormal, thereby reducing the rice fruit setting rate, thousand-grain weight and yield.

[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-resistant rice varieties; 2) Application of plant growth regulators; 3) Staged sowing. At present, in field rice production, single measures are mainly used to deal with high temperature heat damage, such as delaying sowing to avoid high temperatures during the flowering period of rice, but this will increase the probability of encountering high temperatures during the rice panicle differentiation period. In addition, although the application of heat-resistant varieties and plant growth regulators can alleviate high temperature heat damage to a certain extent, it is still difficult to meet the needs of production. Because some varieties are more resistant to high temperatures during the flowering period, but some varieties are less sensitive to high temperatures during the filling period, there is a certain difficulty for the varieties. Plant growth regulators are not suitable for all high temperature stress situations. Therefore, how to scientifically and rationally select cultivation and improvement measures for rice under high temperature stress, so as to alleviate high temperatures and improve the fertility of spikelets, 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 measures for improving the heat resistance of rice, with the aim of formulating scientific and reasonable cultivation improvement measures for high temperature heat damage of rice and regulating the heat resistance of rice plants.

[0005] The first aspect of the present invention provides a recommended method for improving the heat tolerance of rice through cultivation, comprising:

[0006] 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;

[0007] 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;

[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] The cultivation improvement measures included in the recommended path are extracted, and a combination of cultivation improvement measures is determined according to the soil environmental characteristics of the target rice-growing area as a recommendation result of the cultivation improvement measures.

[0010] In this scheme, heat damage monitoring parameters at different rice growth stages are obtained based on the pre-processed multi-source environmental parameters, and heat damage early warning is performed in the target rice planting area through the heat damage monitoring parameters, specifically:

[0011] 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;

[0012] When all grid fast comparisons are completed, 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;

[0013] 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;

[0014] 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;

[0015] 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.

[0016] In this scheme, morphological traits related to high temperature stress were screened at different rice growth stages, and corresponding mapping relationships were constructed, 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 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;

[0028] 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;

[0029] 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.

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

[0031] In this scheme, the cultivation improvement measures included in the extraction based on the recommended path are specifically:

[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 the cultivation improvement measures;

[0033] 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.

[0034] In this plan, a combination of cultivation improvement measures is determined based on the soil environmental characteristics of the target rice planting area as the recommended results of cultivation improvement measures, specifically:

[0035] 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.

[0036] 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.

[0037] The second aspect of the present invention provides a cultivation improvement method recommendation system for improving the heat resistance of rice, 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;

[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 at different rice growth stages;

[0039] 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;

[0040] 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.

[0041] 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;

[0042] 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;

[0043] 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.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention analyzes the damage mechanism of high temperature stress on the fertility of rice spikelets, formulates scientific and reasonable cultivation and improvement measures for high temperature heat damage of rice, alleviates high temperature, improves the fertility of spikelets, regulates the heat resistance of rice plants, and alleviates the problem of rice production reduction caused by high temperature weather by recommending a combination of cultivation and improvement measures, thereby ensuring 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 embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0047] Figure 1 A flow chart showing a recommended method for improving the cultivation of rice to improve its heat tolerance according to the present invention;

[0048] Figure 2 A flow chart showing the method of screening morphological trait indicators related to high temperature stress according to an embodiment of the present invention is shown;

[0049] Figure 3A flow chart of obtaining a recommended path including cultivation improvement measures according to an embodiment of the present invention is shown;

[0050] Figure 4 A block diagram of the cultivation improvement method recommendation system for improving the heat resistance of rice according to the present invention is shown. DETAILED DESCRIPTION

[0051] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0053] Figure 1 A flow chart showing a recommended method for the cultivation improvement means for improving the heat resistance of rice according to the present invention.

[0054] like Figure 1 As shown, in the first embodiment of the present invention, a recommended method for improving the heat tolerance of rice is provided, comprising:

[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 pre-processed multi-source environmental parameters, and performing heat damage early warning for the target rice planting area according to 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, and when a heat damage warning is generated, reading the morphological trait indicators of the current rice growth stage according to the mapping relationship;

[0057] S106, obtaining the current high temperature characteristics of the target rice planting area according to the read morphological trait indicators, connecting the high temperature characteristics to a 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 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.

[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, 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.

[0066] It should be noted that rice suffers from high temperature stress during the tillering stage, and its tiller number and effective panicle number are reduced. The heading and flowering stage is most susceptible to high temperature. High temperature stress will cause reduced fertility of spikelets, cracking of anthers, abnormal growth of pollen tubes, etc. According to heat-sensitive periods such as the tillering stage and the heading and flowering stage, examples of the impact of high temperature stress on rice plants are obtained. A subset of rice plant organs related to high temperature stress in different rice growth stages is obtained. For example, high temperature during the heading and flowering stage of rice will affect the formation of rice spikelets, pollen development, stamen structure, flowering fertilization and physiological metabolism of spikelets. High temperature will increase the viscosity of pollen, making it difficult for pollen to be dispersed in the anther, and the spikelet filaments will lose water and wilt, which will also cause the flowering of spikelets to be blocked. The shape and properties of anthers and spikelets can reflect the heat damage of rice to a certain extent.

[0067] It should be noted that the rice growth stage of the rice plants in the regional block where there is a heat damage warning in the target rice-growing area is obtained, the corresponding morphological trait indicators are read based on the mapping relationship, and the indicator parameters of the rice plants in the regional block are read based on the morphological trait indicators; a parameter matrix is ​​constructed according to the indicator parameters, and the residual vector of the parameter matrix and the benchmark parameter matrix corresponding to the morphological trait indicators under the normal state is calculated; based on the influencing example training, a corresponding relationship between the degree of high temperature stress and the change in morphological traits is established; for example, when the temperature is higher, the smaller the cracking of the rice anther, that is, the greater the degree of obstruction of the anther cracking, the higher the degree of high temperature stress; the residual vector is further represented by the corresponding relationship, and the high temperature characteristics that characterize the severity of heat damage in the regional block are obtained.

[0068] Figure 3 A flow chart of obtaining a recommended path including cultivation improvement measures according to an embodiment of the present invention is shown.

[0069] According to an embodiment of the present invention, the high temperature feature is connected to a knowledge graph related to high temperature stress defense measures, and a recommended path including cultivation improvement measures is obtained according to the knowledge graph, specifically:

[0070] S302, obtaining a number of knowledge graphs related to high temperature stress defense measures according to the example of the impact of high temperature stress on rice plants, and selecting a preset number of relevant knowledge graphs based on the richness of entities and relationships in the relevant knowledge graphs;

[0071] S304, constructing an interaction matrix of heat damage conditions and high temperature stress defense measures in each relevant knowledge graph, performing representation learning on the interaction matrix using a graph convolutional network, and obtaining embedding vectors of entities and relationships in the relevant knowledge graph;

[0072] S306, initializing a random walk model, locating entities according to the high temperature characteristics, taking the located entities as the starting point, taking the cultivation improvement measures as entity category constraints, combining the length constraints to generate random walk rules, and sampling paths containing the cultivation improvement measures based on the random walk rules;

[0073] S308, introducing a federated learning framework, using the random walk model in each relevant knowledge graph as a local model, and generating a confidence score for the local model based on the average improvement effect of the acquired cultivation improvement measures, uploading and aggregating the local models whose confidence score is greater than a preset threshold, and constructing a global model;

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

[0075] It should be noted that the random walk-based cultivation improvement measure recommendation model is used as a global model. Preferably, the random walk can be a meta-path model, an RWNN model, etc. The received training parameters of each 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 retaining a part of the parameter information of the previous round, and then calculating the difference and averaging the received local model parameters and the global model parameters of the previous round in turn. After training, the global model is further optimized, and the optimized model parameter information is then passed to each local model, and the updated local model parameters are passed back to the global model of the terminal server.

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

[0077] Based on multi-source environmental parameters, the soil environmental characteristics of the heat damage warning area block in the target rice planting area are analyzed, and the encoding representation L of the cultivation improvement measures and the soil environmental characteristics T are learned using a multi-layer perceptron; combined with the activation function operation, the interactive prediction probability y of the heat damage warning area block for the cultivation improvement measures is obtained, which is expressed as: y = σ (MLP (LT)), σ represents the activation function. The cultivation improvement measures are sorted according to the interactive prediction probability, and the cultivation improvement measures are selected in turn according to the preset number to determine the combination of cultivation improvement measures to improve the heat resistance of rice plants.

[0078] Figure 4 A block diagram of the cultivation improvement method 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 method 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 at different rice growth stages;

[0081] 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;

[0082] 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.

[0083] 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;

[0084] 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;

[0085] 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.

[0086] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: 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 can be through some interfaces, and the indirect coupling or communication connection of the 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, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

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

[0089] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0090] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for 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 each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0091] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

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; The cultivation improvement measures included in the recommended path are extracted, and a combination of cultivation improvement measures is determined according to the soil environmental characteristics of the target rice-growing area as a recommendation result of the cultivation improvement measures.

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 fast comparisons are completed, 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 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; 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.

4. The method for improving the heat tolerance 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.

5. The method for recommending cultivation improvement means for improving the heat tolerance of rice according to claim 1, characterized in that: 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.

6. The method for recommending cultivation improvement means 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.

7. 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.

8. A cultivation improvement method recommendation system for improving the heat tolerance of rice, characterized in that: It includes multi-source environmental data collection module, heat damage warning module, rice shape and state analysis module, knowledge graph module, cultivation improvement measure evaluation module and 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