Meteorological disaster loss assessment method and device for crops and storage medium

Abstract

The invention discloses a meteorological disaster loss assessment method and device for crops and a storage medium. Comprising the following steps: acquiring first weather conditions corresponding to a plurality of sample periods and yield change samples in each sample period; obtaining a second weather condition of each time period in the target period, and determining a first association relationship based on each first weather condition and the second weather condition; based on the first association relationship, the yield change sample and the virtual yield change condition, respectively constructing a graph structure corresponding to each growth period; using a plurality of graph neural network models to iteratively update each graph structure, using a plurality of yield prediction models to predict each iteratively updated graph structure, and determining a real yield change condition corresponding to each growth period in the target period; and fusing the real yield change conditions corresponding to the growth periods, and evaluating disaster loss in the target period according to a fusion result.

Description

Technical Field

[0001] This application relates to the field of crop growth technology, and in particular to methods, devices and storage media for assessing meteorological disaster losses to crops. Background Technology

[0002] For crops with long growing seasons and almost complete exposure to the natural environment, predictive disaster prevention based on meteorological data is the most crucial means to achieve stable, high, and high yields. Crops with long growing seasons include winter wheat and rice.

[0003] Taking winter wheat as an example. Firstly, the key growth periods for winter wheat include the sowing and emergence period, the overwintering period, the greening-jointing period, the heading-flowering period, and the grain-filling and ripening period. The heading-flowering period can be further divided into the booting stage, the heading-flowering stage, and the pre-grain-filling stage. Among these, the heading-flowering stage has the greatest impact on yield; damage during this period can lead to a yield reduction of 30% to 50%, which is difficult to compensate for through later management. The pre-grain-filling stage has the least impact; improper management during this period can reduce dry grain weight by 10% to 20%, but this can be partially mitigated through foliar fertilization and other measures. The booting stage has the least impact; problems during this period may limit the upper limit of yield, but this can be partially compensated for later if the environment is suitable by increasing the grain filling rate.

[0004] Furthermore, different weather conditions have varying impacts on winter wheat at different stages (including the heading-flowering stage and the pre-grain-filling stage), leading to different yield variations. For example, continuous rain or heavy rainfall during the heading-flowering stage may wash away pollen, preventing pollination, or the high humidity may induce Fusarium head blight. Similarly, drought during the heading-flowering stage may result in poor pollen development, hindering pollen attachment and germination. Finally, low temperatures and frost during the heading-flowering stage may cause frost damage and necrosis of winter wheat tissues.

[0005] For example, if hot, dry winds occur during the early stages of grouting, problems such as premature ripening or forced planting during the grouting process may occur. Furthermore, strong winds or heavy rains during the early stages of grouting may cause plants to lodging and stems to break.

[0006] Therefore, it is crucial to predict the yield loss of winter wheat based on the impact of weather conditions on different growth stages, and to assess and take targeted precautions against meteorological disasters based on the yield loss.

[0007] There is currently no effective solution to the technical problems existing in the above-mentioned prior art regarding how to predict crop yield loss based on weather conditions and assess meteorological disaster losses based on yield loss in order to prevent meteorological disasters affecting crops. Summary of the Invention

[0008] The embodiments of this disclosure provide a method, apparatus, and storage medium for assessing meteorological disaster losses to crops, thereby at least solving the technical problems existing in the prior art of how to predict crop yield losses based on weather conditions and assess meteorological disaster losses based on yield losses, so as to facilitate the prevention of meteorological disasters to crops.

[0009] According to one aspect of the present disclosure, a method for assessing meteorological disaster losses to crops is provided, comprising: acquiring first weather conditions corresponding to each time period within multiple sample periods, and yield change samples corresponding to disaster conditions at each growth stage within each sample period; acquiring second weather conditions for each time period within a target period, and determining a first correlation based on each first weather condition and the second weather condition, wherein the first correlation includes the correlation between the second weather condition and each first weather condition, and the correlation between any two first weather conditions; constructing graph structures corresponding to each growth stage based on the first correlation, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period; iteratively updating each graph structure using multiple pre-trained graph neural network models, and predicting each iteratively updated graph structure using multiple pre-trained yield prediction models to determine the actual yield change corresponding to each growth stage within the target period; and fusing the actual yield change corresponding to each growth stage, and assessing the disaster loss within the target period based on the fusion result.

[0010] According to another aspect of the present disclosure, a storage medium is also provided, the storage medium including a stored program, wherein, when the program is executed, a processor performs any of the methods described above.

[0011] According to another aspect of the present disclosure, a device for assessing meteorological disaster losses of crops is also provided, comprising: a first data acquisition module, configured to acquire first weather conditions corresponding to each time period within multiple sample periods, and yield change samples corresponding to disaster conditions at each growth stage within each sample period; a second data acquisition module, configured to acquire second weather conditions for each time period within a target period, and determine a first correlation based on each first weather condition and the second weather condition, wherein the first correlation includes the correlation between the second weather condition and each first weather condition, and the correlation between any two first weather conditions; a graph structure construction module, configured to construct graph structures corresponding to each growth stage based on the first correlation, yield change samples corresponding to each sample period, and virtual yield change corresponding to the target period; a prediction module, configured to iteratively update each graph structure using multiple pre-trained graph neural network models, and predict each iteratively updated graph structure using multiple pre-trained yield prediction models, to determine the actual yield change corresponding to each growth stage within the target period; and an evaluation module, configured to fuse the actual yield change corresponding to each growth stage, and evaluate the disaster loss within the target period based on the fusion result.

[0012] According to another aspect of the present disclosure, a device for assessing meteorological disaster losses of crops is also provided, comprising: a processor; and a memory connected to the processor, configured to provide the processor with instructions for processing the following steps: acquiring first weather conditions corresponding to each time period within multiple sample periods, and yield change samples corresponding to disaster conditions at each growth stage within each sample period; acquiring second weather conditions for each time period within a target period, and determining a first correlation based on each first weather condition and the second weather condition, wherein the first correlation includes the correlation between the second weather condition and each first weather condition, and the correlation between any two first weather conditions; constructing graph structures corresponding to each growth stage based on the first correlation, yield change samples corresponding to each sample period, and virtual yield change corresponding to the target period; iteratively updating each graph structure using multiple pre-trained graph neural network models, and predicting each iteratively updated graph structure using multiple pre-trained yield prediction models to determine the actual yield change corresponding to each growth stage within the target period; and fusing the actual yield change corresponding to each growth stage, and assessing the disaster loss within the target period based on the fusion result.

[0013] This application provides a method for assessing meteorological disaster losses to crops. First, the processor acquires the first weather conditions corresponding to each time period within multiple sample periods, and yield change samples corresponding to disaster conditions at each growth stage within each sample period. Then, the processor acquires the second weather conditions for each time period within a target period, and determines a first correlation based on each first and second weather condition. This first correlation includes the correlation between the second weather condition and each first weather condition, as well as the correlation between any two first weather conditions. Next, based on the first correlation, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period, the processor constructs graph structures corresponding to each growth stage. Further, the processor iteratively updates each graph structure using multiple pre-trained graph neural network models, and predicts the updated graph structures using multiple pre-trained yield prediction models to determine the actual yield changes corresponding to each growth stage within the target period. Finally, the processor fuses the actual yield changes corresponding to each growth stage and assesses the disaster losses within the target period based on the fusion result.

[0014] As described above, this application obtains the first weather conditions corresponding to each time period within multiple sample periods and the second weather conditions corresponding to each time period within the target period, and determines the first weather conditions of multiple sample periods and the second weather conditions of the target period, as well as the first correlation between the first weather conditions of any two sample periods. Therefore, when the first weather conditions of each time period within each sample period affect different growth stages of crops, thereby damaging crop yields within the sample period, multiple graph structures corresponding to different growth stages can be constructed based on the first correlation. These graph structures are used to represent the correlation between the virtual yield changes in the target period and the yield change samples of each sample period.

[0015] Then, multiple pre-trained graph neural network models are used to iteratively update each graph structure corresponding to different periods. Furthermore, multiple pre-trained yield prediction models are used to predict each iteratively updated graph structure, thereby predicting the actual yield changes corresponding to each growth stage within the target period.

[0016] In other words, operators can determine the potential damage caused by weather conditions affecting different growth stages of crops based on the predicted changes in actual yields corresponding to each growth stage, and assess the damage.

[0017] This solves the technical problem in existing technologies of how to predict crop yield loss based on weather conditions and assess meteorological disaster losses based on yield loss, so as to facilitate the prevention of meteorological disasters affecting crops. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this disclosure and form part of this application, illustrate exemplary embodiments of this disclosure and are used to explain this disclosure, but do not constitute an undue limitation of this disclosure. In the drawings: Figure 1 This is a hardware structure block diagram of a computing device for implementing the method described in Embodiment 1 of this disclosure; Figure 2 This is a schematic diagram of a meteorological disaster loss assessment system for crops according to Embodiment 1 of this disclosure; Figure 3 This is a modular schematic diagram of a meteorological disaster loss assessment system for crops according to Embodiment 1 of this disclosure; Figure 4 This is a flowchart illustrating the meteorological disaster loss assessment method for crops according to Embodiment 1 of this disclosure; Figure 5A This is a schematic diagram of the structure of the first figure according to Embodiment 1 of this disclosure; Figure 5B This is a schematic diagram of the structure of the second figure according to Embodiment 1 of this disclosure; Figure 6 This is a schematic diagram of the structure of the first neural network model according to Embodiment 1 of this disclosure; Figure 7 This is a schematic diagram of the structure of the third MLP model and the first production prediction model according to Embodiment 1 of this disclosure; Figure 8 This is a schematic diagram of the structure of the second neural network model according to Embodiment 1 of this disclosure; Figure 9 This is a schematic diagram of the structure of the fourth MLP model and the second production prediction model according to Embodiment 1 of this disclosure; Figure 10 This is a schematic diagram of a meteorological disaster loss assessment device for crops according to Embodiment 2 of this disclosure; Figure 11 This is a schematic diagram of a meteorological disaster loss assessment device for crops according to Embodiment 3 of this disclosure. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this disclosure.

[0020] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] Example 1 According to this embodiment, a method embodiment for assessing meteorological disaster losses to crops is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0022] The method embodiments provided in this example can be executed on mobile terminals, computer terminals, servers, or similar computing devices. Figure 1 A hardware block diagram of a computing device for implementing a method for assessing meteorological disaster losses to crops is shown. Figure 1 As shown, a computing device may include one or more processors (processors may include, but are not limited to, microprocessors such as MCUs or programmable logic devices such as FPGAs), a memory for storing data, a transmission device for communication functions, and an input / output interface. The memory, transmission device, and input / output interface are connected to the processor via a bus. In addition, it may also include a display, keyboard, and cursor control device connected to the input / output interface. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, a computing device may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0023] It should be noted that the aforementioned one or more processors and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element in a computing device. As involved in the embodiments of this disclosure, the data processing circuits serve as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0024] The memory can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the meteorological disaster loss assessment method for crops in this embodiment of the present disclosure. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the above-mentioned application method for meteorological disaster loss assessment of crops. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the computing device via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0025] The transmission device is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the computing device's communication provider. In one example, the transmission device includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0026] The display can be, for example, a touchscreen liquid crystal display (LCD), which allows users to interact with the user interface of the computing device.

[0027] It should be noted here that, in some optional embodiments, the above... Figure 1 The computing device shown may include hardware elements (including circuitry), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware and software elements. It should be noted that... Figure 1This is only one instance of a specific particular instance, and is intended to illustrate the types of components that may exist in the aforementioned computing devices.

[0028] Figure 2 This is a schematic diagram of the meteorological disaster loss assessment system for crops described in this embodiment. (Refer to...) Figure 2 As shown, the system includes: a terminal device 100, a processor 200 communicatively connected to the terminal device 100, and a memory 300 communicatively connected to the processor 200. Operators can send meteorological disaster loss assessment requests for crops to the processor 200 via the terminal device 100. In response to the received assessment request, the processor 200 retrieves from the memory 300 the first weather conditions corresponding to each time period within multiple sample periods, yield change samples corresponding to disaster conditions at each growth stage within each sample period, and the second weather conditions for each time period within a target period, and predicts the actual yield changes corresponding to each growth stage within the target period.

[0029] In addition, the processor 200 is also used to fuse the actual yield changes corresponding to each growth stage, assess the disaster loss based on the fusion results, and return the assessment results to the terminal device 100.

[0030] It should be noted that the terminal device 100, processor 200, and memory 300 in the system can all be adapted to the hardware structure described above.

[0031] Figure 3 This is a modular schematic diagram of the meteorological disaster loss assessment system for crops described in this embodiment. (Reference) Figure 3 As shown, the system includes: a data acquisition module, a correlation calculation module, a graph structure construction module, an iterative update module, a yield prediction module, a fusion and evaluation module, and a storage module.

[0032] The data acquisition module is connected to the storage module and is used to obtain from the storage module the first weather conditions corresponding to each time period within multiple sample periods, the yield change samples corresponding to the disaster conditions of each growth stage within each sample period, and the second weather conditions for each time period within the target period.

[0033] The correlation calculation module is connected to the data acquisition module to calculate the first correlation between the second weather condition and each first weather condition, as well as the first correlation between any two first weather conditions.

[0034] The graph structure construction module is communicatively connected to the correlation calculation module, and the graph structure construction module includes a first graph structure construction unit and a second graph structure construction unit. The first graph structure construction unit is used to construct a first graph structure corresponding to the first growth stage. The second graph structure construction unit is used to construct a second graph structure corresponding to the second growth stage.

[0035] The iterative update module includes a first iterative update unit communicatively connected to the first graph structure building unit, and a second iterative update unit communicatively connected to the second graph structure building unit. The first iterative update unit is used to iteratively update the first graph structure using a first graph neural network model. The second iterative update unit is used to iteratively update the second graph structure using a second graph neural network model.

[0036] The production forecasting module includes a first production forecasting unit communicatively connected to a first iterative update unit and a second production forecasting unit communicatively connected to a second iterative update unit. The first production forecasting unit uses a first production forecasting model and, based on the iteratively updated first graph structure, to predict the first actual production change corresponding to a first growth period. The second production forecasting unit uses a second production forecasting model and, based on the iteratively updated second graph structure, to predict the second actual production change corresponding to a second growth period.

[0037] The fusion and evaluation module communicates with the yield prediction module to fuse the first actual yield change corresponding to the first growth period and the second actual yield change corresponding to the second growth period, and to evaluate the disaster loss of the target period based on the fusion results.

[0038] Under the aforementioned operating environment, according to the first aspect of this embodiment, a method for assessing meteorological disaster losses to crops is provided. This method comprises... Figure 3 The system implementation shown. Figure 4 A flowchart illustrating the method is shown below. (Refer to...) Figure 4 As shown, the method includes: S402: Obtain the first weather conditions corresponding to each time period within multiple sample periods, and the yield change samples corresponding to the disaster conditions at each growth stage within each sample period; S404: Obtain the second weather conditions for each time period within the target period, and determine the first correlation based on each first weather condition and the second weather conditions, wherein the first correlation includes the correlation between the second weather condition and each first weather condition, as well as the correlation between any two first weather conditions; S406: Based on each first correlation, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period, construct graph structures corresponding to each growth stage respectively. S408: Multiple pre-trained graph neural network models are used to iteratively update each graph structure, and multiple pre-trained yield prediction models are used to predict the actual yield changes corresponding to each growth stage within the target period; and S410: Integrate the actual yield changes corresponding to each growth stage, and assess the disaster losses within the target period based on the integration results.

[0039] Specifically, firstly, the processor 200 responds to the meteorological disaster loss assessment request for crops sent by the operator through the terminal device 100, and uses the data acquisition module to obtain from the storage module the first weather conditions corresponding to each time period within multiple sample periods, and the yield change samples corresponding to the disaster conditions at each growth stage within each sample period (S402). In this embodiment, winter wheat is used as an example. Since the heading-flowering period and the early grain-filling stage have a significant impact on winter wheat yield, while the booting stage has a smaller impact, the multiple growth stages in this embodiment include a first growth stage corresponding to the heading-flowering period of winter wheat and a second growth stage corresponding to the early grain-filling stage of winter wheat.

[0040] Furthermore, the sample period is, for example, in years. Each period within the sample period could be, for example, the weather conditions for 90 days from January to March of a year. For example, compared to the first sample period... The corresponding first weather condition .in, This represents the weather conditions on day i of the first sample period, where i = 1 to 90. And within this, . This represents the average temperature corresponding to day i in the first sample period. This represents the maximum temperature corresponding to day i in the first sample period. This represents the rainfall amount corresponding to day i in the first sample period. This represents the maximum wind force corresponding to day i in the first sample period. This represents the solar luminosity corresponding to day i in the first sample period. This represents the illumination duration corresponding to the i-th day of the first sample period. That is, j = 1~6.

[0041] With the second sample period The corresponding first weather condition .in, This represents the weather conditions on day i in the second sample period, where i = 1 to 90. And within this, . This represents the average temperature corresponding to day i in the second sample period. This represents the maximum temperature corresponding to day i in the second sample period. This represents the rainfall amount corresponding to day i in the second sample period. This represents the maximum wind force corresponding to day i in the second sample period. This represents the daylight intensity corresponding to day i in the second sample period. This represents the illumination duration corresponding to the i-th day of the second sample period. That is, j = 1~6.

[0042] And so on.

[0043] With the nth sample period The corresponding first weather condition .in, This represents the weather conditions on day i in the nth sample period, where i = 1 to 90. And within this, . This represents the average temperature corresponding to day i in the nth sample period. This represents the maximum temperature corresponding to day i in the nth sample period. This represents the rainfall amount corresponding to the i-th day of the n-th sample period. This represents the maximum wind force corresponding to day i in the nth sample period. This represents the solar luminosity corresponding to the i-th day of the n-th sample period. This represents the illumination duration corresponding to the i-th day of the n-th sample period. That is, j = 1~6.

[0044] The data acquisition module then collects data from the storage module for each sample period. ~ Within this sample, yield changes correspond to disaster conditions at various growth stages. For example, across multiple sample periods. ~ Within, sample period ~ The initial weather conditions affect the first growth stage of crops, thus leading to a reduction in crop yield. The data acquisition module, for example, can determine the relationship with the sample period. Corresponding production change sample , with the sample period Corresponding production change sample ... and the sample period Corresponding production change sample .

[0045] Similarly, for example, in multiple sample periods ~ Within, sample period ~ The first weather condition affects the second growth stage of crops, thus leading to a reduction in crop yield. The data acquisition module, for example, can determine the relationship with the sample period. Corresponding production change sample , with the sample period Corresponding production change sample ... and the sample period Corresponding production change sample .

[0046] Based on the above method, the data acquisition module can obtain from the storage module the first weather conditions corresponding to each time period within multiple sample periods, as well as the yield change samples corresponding to the disaster conditions at each growth stage within each sample period.

[0047] Then, the data acquisition module retrieves the second weather conditions for each time period within the target period from the storage module. The target period... The various time periods within this period could be, for example, the weather conditions during the 90 days from January to March of the year. For example, compared to the target period... The corresponding weather conditions on the second day .in, This represents the weather conditions on day i of the target period, where i = 1 to 90. And within this, . This represents the average temperature on day i of the target period. This represents the maximum temperature corresponding to day i of the target period. This represents the rainfall amount corresponding to day i of the target period. This represents the maximum wind force corresponding to day i of the target period. This represents the daylight intensity corresponding to day i of the target period. This represents the duration of illumination corresponding to day i of the target period. That is, j = 1~6.

[0048] Subsequently, the data acquisition module sends the acquired data to the correlation calculation module, which then determines a first correlation based on each of the first and second weather conditions (S404). This first correlation includes the correlation between the second weather condition and each of the first weather conditions, as well as the correlation between any two first weather conditions. Furthermore, the correlation calculation module can, for example, determine each sample period based on Euclidean distance. ~ First weather conditions ~ With target cycle Second weather conditions The first relationship between them ~ Similarly, the period of each sample can be calculated. ~ First weather conditions ~ With target cycle Second weather conditions The first relationship between them ~ .

[0049] Similarly, it is also possible to calculate any two sample periods. ~ First weather conditions ~ The primary correlation between them. For example, the sample period. First weather conditions With sample period First weather conditions The first relationship between them Sample period First weather conditions With sample period First weather conditions The first relationship between them ..., sample period First weather conditions With sample period First weather conditions The first relationship between them .

[0050] For example, the sample period First weather conditions With sample period First weather conditions The first relationship between them Sample period First weather conditions With sample period First weather conditions The first relationship between them ..., sample period First weather conditions With sample period First weather conditions The first relationship between them .etc.

[0051] Furthermore, the correlation calculation module can also be based on each primary correlation relationship. ~ Multiple second associations are determined by comparing them with a pre-set association threshold. ~ Similarly, the correlation calculation module can be based on each primary correlation. ~ Multiple second associations are determined by comparing them with a pre-set association threshold. ~ Similarly, the correlation calculation module can determine multiple second correlations based on the first correlation between any two first weather conditions and a pre-set correlation threshold, which will not be elaborated here.

[0052] Specifically, if the first association is greater than the association threshold, the first association is determined as the second association; if the first association is less than or equal to the association threshold, the first association is removed. The above will be described in detail later, and therefore will not be repeated here.

[0053] Then, based on the first correlation, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period, the graph structure construction module constructs the graph structure corresponding to each growth period (S406). Specifically, when the correlation calculation module determines multiple second correlations, the first graph structure construction unit in the graph structure construction module uses each second correlation as the first edge and the first virtual yield change corresponding to the first growth period and the yield change samples corresponding to the first growth period as the first nodes to construct the first graph structure corresponding to the first growth period.

[0054] In this application, the first growth period can be, for example, the heading-flowering stage of winter wheat. Therefore, the yield change sample corresponding to the first growth period could be, for example, the influence of the first weather condition on the heading-flowering stage of winter wheat, resulting in a change in the yield of winter wheat.

[0055] Similarly, the second graph structure building unit in the graph structure building module uses each second association as the second edge, and the second virtual yield change corresponding to the second growth period and the yield change sample corresponding to the second growth period as the second node to build the second graph structure corresponding to the second growth period.

[0056] In this application, the second growth stage can be, for example, the early grain-filling stage of winter wheat. Therefore, the yield change sample corresponding to the second growth stage could be, for example, the influence of the first weather condition on the early grain-filling stage of winter wheat, leading to a change in winter wheat yield.

[0057] The above content will be described in detail later, so it will not be repeated here.

[0058] Furthermore, the iterative update module utilizes multiple pre-trained graph neural network models to iteratively update each graph structure. Specifically, when the first graph structure construction unit in the graph structure construction module constructs the first graph structure, the first iterative update unit in the iterative update module iteratively updates the first graph structure. The first iterative update unit includes a first graph neural network model, which is used to iteratively update the first graph structure.

[0059] Similarly, when the second graph structure building unit in the graph structure building module builds the second graph structure, the second iterative update unit in the iterative update module iteratively updates the second graph structure. The second iterative update unit includes a second graph neural network model, which is used to iteratively update the second graph structure.

[0060] The yield prediction module then uses multiple pre-trained yield prediction models to predict the iteratively updated graph structures, determining the actual yield changes corresponding to each growth stage within the target period (S408). Specifically, when the first iterative update unit in the iterative update module iteratively updates the first graph structure, the first yield prediction unit in the yield prediction module uses the pre-trained first yield prediction model to predict the iteratively updated first graph structure. This predicts the first actual yield change corresponding to the first growth stage within the target period.

[0061] Similarly, when the second iterative update unit in the iterative update module iteratively updates the second graph structure, the second production prediction unit in the production prediction module uses a pre-trained second production prediction model to predict the iteratively updated second graph structure. This allows for the prediction of the second actual production change corresponding to the second growth period within the target cycle.

[0062] The fusion and evaluation module then merges the first true yield change corresponding to the first growth stage and the second true yield change corresponding to the second growth stage. This determines the yield change of crops (e.g., winter wheat) caused by the impact of the second weather condition on the first and second growth stages within the target period (i.e., the fusion result).

[0063] Finally, the fusion and assessment module assesses the disaster loss based on the fusion results.

[0064] As described in the background section, different weather conditions have varying impacts on winter wheat at different stages (including the heading-flowering stage and the pre-grain-filling stage), leading to different yield variations. For example, continuous rain or heavy rainfall during the heading-flowering stage may wash away pollen, preventing pollination, or the high humidity may induce Fusarium head blight. Similarly, drought during the heading-flowering stage may result in poor pollen development, hindering pollen attachment and germination. Furthermore, low temperatures and frost during the heading-flowering stage may cause frost damage and necrosis of winter wheat tissues.

[0065] For example, if hot, dry winds occur during the early stages of grouting, problems such as premature ripening or forced planting during the grouting process may occur. Furthermore, strong winds or heavy rains during the early stages of grouting may cause plants to lodging and stems to break.

[0066] Therefore, it is crucial to predict the yield loss of winter wheat based on the impact of weather conditions on different growth stages, and to assess and take targeted precautions against meteorological disasters based on the yield loss.

[0067] In view of this, this application provides a method for assessing meteorological disaster losses to crops. Referring to the above description, this application obtains the first weather conditions corresponding to each time period within multiple sample periods and the second weather conditions corresponding to each time period within a target period, and determines the first weather conditions of multiple sample periods and the second weather conditions of the target period, as well as the first weather conditions of any two sample periods. Therefore, when the first weather conditions of each time period within each sample period affect different growth stages of crops, thereby damaging crop yields within the sample periods, multiple graph structures corresponding to different growth stages can be constructed based on the first correlation. These graph structures are used to represent the correlation between the virtual yield changes in the target period and the yield change samples of each sample period.

[0068] Then, multiple pre-trained graph neural network models are used to iteratively update each graph structure corresponding to different periods. Furthermore, multiple pre-trained yield prediction models are used to predict each iteratively updated graph structure, thereby predicting the actual yield changes corresponding to each growth stage within the target period.

[0069] In other words, operators can determine the potential damage caused by weather conditions affecting different growth stages of crops based on the predicted changes in actual yields corresponding to each growth stage, and assess the damage.

[0070] This solves the technical problem in existing technologies of how to predict crop yield loss based on weather conditions and assess meteorological disaster losses based on yield loss, so as to facilitate the prevention of meteorological disasters affecting crops.

[0071] Optionally, the operation of constructing graph structures corresponding to each growth stage based on each first association, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period includes: determining multiple second associations based on each first association and a pre-set association threshold; and constructing graph structures corresponding to each growth stage using each second association as edges and the yield change samples corresponding to each sample period and the virtual yield change corresponding to the target period as nodes. Further optionally, the operation of determining multiple second associations based on each first association and a pre-set association threshold includes: determining the relationship between each first association and the association threshold; determining a first association as a second association if the first association is greater than the association threshold; and removing a first association if the first association is less than or equal to the association threshold. Further optionally, the multiple growth periods include a first growth period and a second growth period, and the operation of constructing a graph structure corresponding to each growth period with each second association as an edge and the output change sample corresponding to each sample period and the virtual output change corresponding to the target period as nodes includes: constructing a first graph structure corresponding to the first growth period with each second association as a first edge and the first virtual output change corresponding to the first growth period and the output change sample corresponding to the first growth period as first nodes; and constructing a second graph structure corresponding to the second growth period with each second association as a second edge and the second virtual output change corresponding to the second growth period and the output change sample corresponding to the second growth period as second nodes.

[0072] Specifically, as can be seen from the above description, the correlation calculation module can, for example, use the Euclidean distance calculation method to calculate the correlation of each sample period. ~ First weather conditions ~ With target cycle Second weather conditions The first relationship between them ~ .

[0073] For example, the first association Indicates the weather conditions on the second day With sample period First weather conditions The relationships between them, where k = 1 to n. The specific calculation formula is as follows: in, Indicates the weather conditions on the second day With each first weather condition The first relationship between them This represents the second weather condition on day i within the target period. This represents the first weather condition on day i within the k-th sample period.

[0074] Similarly, the correlation calculation module can calculate the first correlation between the first weather conditions of any two sample periods in the same way as described above, which will not be elaborated here.

[0075] The correlation calculation module then maps the determined first correlations to the range of 0 to 1.

[0076] Furthermore, the correlation calculation module determines the relationship between each first correlation and a pre-set correlation threshold. For example, if a first correlation is greater than the correlation threshold, it is identified as a second correlation. And if a first correlation is less than or equal to the correlation threshold, it is discarded.

[0077] Then, the correlation calculation module sends the determined second associations to the graph structure construction module.

[0078] Furthermore, the first graph structure construction unit in the graph structure construction module uses multiple second relationships as first edges and the first virtual yield change corresponding to the first growth period and the yield change sample corresponding to the first growth period as first nodes to construct a first graph structure corresponding to the first growth period. The first graph structure is used to represent the relationship between the first virtual yield change in the target period and the yield change samples in each sample period.

[0079] Figure 5A This is a schematic diagram of the structure described in the first figure according to an embodiment of this application. (Reference) Figure 5A As shown, the first node With the first node The first side corresponds to First node With the first node The first side corresponds to First node With the first node The first side corresponds to First node With the first node The first side corresponds to ...etc. In the first diagram structure, each first node represents the impact of the first weather condition on the first growth period of crops, resulting in a loss of crop yield in the corresponding sample period.

[0080] Meanwhile, the second graph structure construction unit in the graph structure construction module uses multiple second relationships as second edges and the second virtual yield change corresponding to the second growth period and the yield change samples corresponding to the second growth period as second nodes to construct a second graph structure corresponding to the second growth period. The second graph structure is used to represent the relationship between the second virtual yield change in the target period and the yield change samples in each sample period.

[0081] Figure 5B This is a schematic diagram of the second structure according to an embodiment of this application. (Reference) Figure 5B As shown, the second node With the second node There is a second side. Second node With the second node There is a second side. Second node With the second node There is a second side. Second node With the second node There is a second side. ...etc. In the second diagram structure, each second node represents the impact of the first weather condition on the second growth stage of crops, resulting in a loss of crop yield in the corresponding sample period.

[0082] Thus, by constructing the first graph structure in the above manner, the technical effect of visually representing the changes in the first virtual yield corresponding to the first growth period within the target period and the correlation between each yield change sample is achieved, based on the determined second correlation. Similarly, by constructing the second graph structure in the above manner, the technical effect of visually representing the changes in the second virtual yield corresponding to the second growth period within the target period and the correlation between each yield change sample is achieved, based on the determined second correlation.

[0083] Optionally, the plurality of graph neural network models include a first graph neural network model corresponding to the first growth stage, and the operation of iteratively updating each graph structure using the pre-trained plurality of graph neural network models includes: inputting the first graph structure into the first graph neural network model, and constructing a first node attribute vector using the calculation rules corresponding to the first node, wherein the first node attribute vector includes the node attributes of the first node, and the product of the edge attributes of two adjacent first edges and the node attributes of the corresponding first node; and iteratively updating the first node attribute vector using the plurality of first MLP models in the first graph neural network model to generate node features corresponding to the first graph structure. Optionally, the yield prediction model includes a first yield prediction model corresponding to the first growth period, and the operation of using multiple pre-trained yield prediction models to predict the iteratively updated graph structures to determine the actual yield changes corresponding to each growth period within the target period includes: determining multiple first target feature vectors, wherein the first target feature vectors include the node features of the first node corresponding to the first virtual yield change after iterative updates, and the product of the edge features of two adjacent first edges and the node features of the corresponding first node; and inputting the multiple first target feature vectors into the first yield prediction model to output the first actual yield change corresponding to the first growth period within the target period.

[0084] Specifically, Figure 6 This is a schematic diagram of the structure of the first neural network model according to an embodiment of this application. (Reference) Figure 3 and Figure 6 As shown, the first graph structure construction unit of the graph structure construction module, when constructing the first graph structure, sends the first graph structure to the first iterative update unit in the iterative update module. The first iterative update unit is equipped with a pre-trained first graph neural network model, which includes multiple first MLP models for iteratively updating the node attributes of the first node of the first graph structure.

[0085] Therefore, the first iterative update unit first constructs the first node attribute vector using the calculation rules corresponding to the nodes. For example, the first iterative update unit constructs the first node attribute vector corresponding to the first node in the first graph structure according to the calculation rules. .in, Represents the first node Node attributes and the first edge The product of edge attributes, Indicates the first node The node attributes and the first edge The product of the edge properties.

[0086] For example, the first iteration update unit constructs the first node attribute vector corresponding to the first node in the first graph structure according to the calculation rules. .in, Indicates the first node The node attributes and the first edge The product of edge attributes, Indicates the first node The node attributes and the first edge The product of the edge properties.

[0087] The construction method of the first node attribute vector corresponding to other first nodes is the same as described above, and will not be repeated here.

[0088] Then, multiple first MLP models iteratively update the first node attribute vectors corresponding to each first node, thereby outputting the node features of each first node after iterative updates.

[0089] Furthermore, the first iterative update unit sends the iteratively updated first graph structure to the production prediction module, so that the production prediction module can determine the first target feature vector corresponding to the iteratively updated first graph structure. The first target feature vector includes the node features of the first node corresponding to the first virtual production change after the iterative update, and the product of the edge features of two adjacent first edges and the node features of the corresponding first node. For example, the first target feature vector... .in, This represents the node features of the first node after iterative updates. Edge features of the first edge The product of This represents the node features of the first node after iterative updates. Edge features of the first edge The product of.

[0090] For example, the first target feature vector .in, This represents the node features of the first node after iterative updates. Edge features of the first edge The product of This represents the node features of the first node after iterative updates. Edge features of the first edge The product of.

[0091] Figure 7 This is a schematic diagram illustrating the structure of the third MLP model and the first production prediction model according to embodiments of this application. (Reference) Figure 7As shown, when the yield prediction module determines multiple first target feature vectors, these first target feature vectors are input into the first yield prediction unit. The first yield prediction unit then uses a third MLP model to fuse the multiple first target feature vectors and generate a fused first target feature vector. Further, the third MLP model inputs the fused target feature vector into a pre-trained first yield prediction model, which then makes predictions based on the fused target feature vector, thereby outputting the first true yield change within the target period corresponding to the first growth stage. The first yield prediction model is a neural network model, and it includes an input layer, a fully connected layer, an activation function layer, and an output layer.

[0092] Furthermore, in this embodiment, the first actual yield change can be, for example, the predicted yield reduction of winter wheat within the target period when the weather conditions of the second day affect the heading-flowering period of winter wheat.

[0093] Optionally, the multiple graph neural network models include a second graph neural network model corresponding to the second growth stage, and the multiple pre-trained graph neural network models are used to iteratively update each graph structure, including: inputting the second graph structure into the second graph neural network model, and constructing a second node attribute vector using the calculation rules corresponding to the second node, wherein the second node attribute vector includes the node attributes of the second node, and the product of the edge attributes of two adjacent second edges and the node attributes of the corresponding second node; and iteratively updating the second node attribute vector using multiple second MLP models in the second graph neural network model to generate node features corresponding to the second graph structure. Optionally, the yield prediction model includes a second yield prediction model corresponding to the second growth period, and the operation of using multiple pre-trained yield prediction models to predict the iteratively updated graph structures to determine the actual yield changes corresponding to each growth period within the target period includes: determining multiple second target feature vectors, wherein the second target feature vectors include the node features of the second node corresponding to the second virtual yield change after iterative updates, and the product of the edge features of two adjacent second edges and the node features of the corresponding second node; and inputting the second target feature vectors into the second yield prediction model to output the second actual yield change corresponding to the second growth period within the target period.

[0094] Specifically, Figure 8 This is a schematic diagram of the second neural network model according to an embodiment of this application. (Reference) Figure 8As shown, the second graph structure construction unit of the graph structure construction module, when constructing the second graph structure, sends the second graph structure to the second iterative update unit in the iterative update module. The second iterative update unit is equipped with a pre-trained second graph neural network model, which includes multiple second MLP models for iteratively updating the node attributes of the second nodes of the second graph structure.

[0095] Therefore, firstly, the second iterative update unit constructs the second node attribute vector using the computation rules corresponding to the second node. For example, the second iterative update unit constructs the second node attribute vector corresponding to the second node in the second graph structure according to the computation rules. .in, Indicates the second node The node attributes and the second side The product of edge attributes, Indicates the second node The node attributes and the second side The product of the edge properties.

[0096] For example, the second iteration update unit constructs the second node attribute vector corresponding to the second node in the second graph structure according to the calculation rules. .in, Indicates the second node The node attributes and the second side The product of edge attributes, Indicates the second node The node attributes and the second side The product of the edge properties.

[0097] The construction method of the attribute vector of the second node corresponding to other second nodes is the same as that described above, and will not be repeated here.

[0098] Then, multiple second MLP models iteratively update the second node attribute vectors corresponding to each constructed second node, thereby outputting the node features of each second node after iterative updates.

[0099] Furthermore, the second iterative update unit sends the iteratively updated second graph structure to the production prediction module, so that the production prediction module can determine the second target feature vector corresponding to the iteratively updated second graph structure. The second target feature vector includes the node features of the second node corresponding to the second virtual production change after the iterative update, and the product of the edge features of two adjacent second edges and the node features of the corresponding second node. For example, the second target feature vector... .in, This represents the node features of the second node after iterative updates. Edge features of the second side The product of This represents the node features of the second node after iterative updates. Edge features of the second side The product of.

[0100] For example, the second target feature vector .in, This represents the node features of the second node after iterative updates. Edge features of the second side The product of This represents the node features of the second node after iterative updates. Edge features of the second side The product of.

[0101] Figure 9 This is a schematic diagram illustrating the structure of the fourth MLP model and the second yield prediction model according to embodiments of this application. (Reference) Figure 9 As shown, when the yield prediction module determines multiple second target feature vectors, these vectors are input into the second yield prediction unit. The second yield prediction unit then uses a fourth MLP model to fuse the multiple second target feature vectors and generate a fused second target feature vector. Further, the fourth MLP model inputs the fused target feature vector into a pre-trained second yield prediction model, which then predicts the second actual yield change within the target period corresponding to the second growth stage. The second yield prediction model is a neural network model, comprising an input layer, a fully connected layer, an activation function layer, and an output layer.

[0102] Furthermore, in this embodiment, the second actual yield change can be, for example, the predicted yield reduction of winter wheat within the target period when the weather conditions of the second day affect the early grain-filling stage of winter wheat.

[0103] Therefore, finally, after the fusion and evaluation module receives the first actual yield change corresponding to the first growth period and the second actual yield change corresponding to the second growth period from the first yield prediction unit in the yield prediction module, the first actual yield change is evaluated. Changes in second actual output Perform a weighted summation and determine the fusion result. The specific calculation formula is as follows: in, This represents the actual output change during the target period (i.e., the fusion result), and α represents the fusion coefficient corresponding to the first actual output change. This represents the reduction in crop yield (i.e., the first true yield change) within the target period when the second weather condition affects the first growth stage of the crop. β represents the fusion coefficient corresponding to the second true yield change. This indicates the reduction in crop yield during the target period (i.e., the change in second actual yield) when the second weather condition affects the second growth stage of crops.

[0104] It is worth noting that the fusion coefficient α corresponding to the first actual output change and the fusion coefficient β corresponding to the second actual output change can be determined by the operator based on experience, and will not be elaborated here.

[0105] Therefore, once the fusion and assessment module determines the fusion results, disaster assessment can be further conducted based on the fusion results, and targeted prevention measures can be taken based on the disaster assessment results.

[0106] Thus, according to the first aspect of this embodiment, it is possible to achieve the technical effect of predicting changes in crop yield based on weather conditions and taking targeted preventative measures based on disaster assessment results.

[0107] In addition, refer to Figure 1 As shown, according to a second aspect of this embodiment, a storage medium is provided. The storage medium includes a stored program, wherein, when the program is executed, a processor performs any of the methods described above.

[0108] Therefore, according to this embodiment, it is possible to achieve the technical effect of predicting changes in crop yield based on weather conditions and taking targeted preventive measures based on disaster assessment results.

[0109] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0110] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution 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 (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0111] Example 2 Figure 10 A meteorological disaster loss assessment device 1000 for crops according to this embodiment is shown, which corresponds to the method described in Embodiment 1. Reference Figure 10 As shown, the device 1000 includes: a first data acquisition module 1010, used to acquire first weather conditions corresponding to each time period within multiple sample periods, and yield change samples corresponding to disaster conditions at each growth stage within each sample period; a second data acquisition module 1020, used to acquire second weather conditions for each time period within a target period, and determine a first correlation based on each first weather condition and the second weather condition, wherein the first correlation includes the correlation between the second weather condition and each first weather condition, and the correlation between any two first weather conditions; and a graph structure construction module 1030, used to construct a graph based on the first correlation. The system constructs graph structures corresponding to each growth stage, including the relationships, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period. A prediction module 1040 iteratively updates each graph structure using multiple pre-trained graph neural network models and predicts each updated graph structure using multiple pre-trained yield prediction models to determine the actual yield change corresponding to each growth stage within the target period. An evaluation module 950 fuses the actual yield changes corresponding to each growth stage and evaluates the disaster loss within the target period based on the fusion result.

[0112] Optionally, the graph structure construction module 1030 includes: a first association determination module, used to determine multiple second associations based on each first association and a pre-set association threshold; and a graph structure construction submodule, used to construct graph structures corresponding to each growth period, with each second association as an edge and the yield change samples corresponding to each sample period and the virtual yield change situation corresponding to the target period as nodes.

[0113] Optionally, the first association determination module includes: a relationship judgment module, used to judge the relationship between each first association and the association threshold; a second association determination module, used to determine the first association as the second association when the first association is greater than the association threshold; and an association elimination module, used to eliminate the first association when the first association is less than or equal to the association threshold.

[0114] Optionally, the multiple growth periods include a first growth period and a second growth period, and the graph structure construction submodule includes: a first graph structure construction unit, used to construct a first graph structure corresponding to the first growth period using each of the second associations as first edges and the first virtual yield change situation corresponding to the first growth period and the yield change sample corresponding to the first growth period as first nodes; and a second graph structure construction unit, used to construct a second graph structure corresponding to the second growth period using each of the second associations as second edges and the second virtual yield change situation corresponding to the second growth period and the yield change sample corresponding to the second growth period as second nodes.

[0115] Optionally, the plurality of graph neural network models include a first graph neural network model corresponding to the first growth stage, and the prediction module 1040 includes: a first graph structure input module, used to input the first graph structure into the first graph neural network model, and construct a first node attribute vector using the calculation rules corresponding to the first node, wherein the first node attribute vector includes the node attributes of the first node, and the product of the edge attributes of two adjacent first edges and the node attributes of the corresponding first node; and a first iterative update module, used to iteratively update the first node attribute vector using a plurality of first MLP models in the first graph neural network model to generate node features corresponding to the first graph structure.

[0116] Optionally, the yield prediction model includes a first yield prediction model corresponding to the first growth period, and the prediction module 1040 includes: a first target feature vector determination module, used to determine a plurality of first target feature vectors, wherein the first target feature vectors include the node features of the first node corresponding to the first virtual yield change after iterative updates, and the product of the edge features of two adjacent first edges and the node features of the corresponding first node; and a first yield output module, used to input the plurality of first target feature vectors into the first yield prediction model, and output the first real yield change corresponding to the first growth period within the target period.

[0117] Optionally, the plurality of graph neural network models include a second graph neural network model corresponding to the second growth stage, and the prediction module 1040 includes: a second graph structure input module, used to input the second graph structure into the second graph neural network model, and construct a second node attribute vector using the calculation rules corresponding to the second node, wherein the second node attribute vector includes the node attributes of the second node, and the product of the edge attributes of two adjacent second edges and the node attributes of the corresponding second node; and a second iterative update module, used to iteratively update the second node attribute vector using a plurality of second MLP models in the second graph neural network model to generate node features corresponding to the second graph structure.

[0118] Optionally, the yield prediction model includes a second yield prediction model corresponding to the second growth period, and the prediction module 1040 includes: a second target feature vector determination module, used to determine a plurality of second target feature vectors, wherein the second target feature vectors include the node features of the second node corresponding to the second virtual yield change after iterative updates, and the product of the edge features of two adjacent second edges and the node features of the corresponding second node; and a second yield output module, used to input the second target feature vectors into the second yield prediction model and output the second real yield change corresponding to the second growth period within the target period.

[0119] Therefore, according to this embodiment, it is possible to achieve the technical effect of predicting changes in crop yield based on weather conditions and taking targeted preventive measures based on disaster assessment results.

[0120] Example 3 Figure 11 A meteorological disaster loss assessment device 1100 for crops according to this embodiment is shown, which corresponds to the method described in Embodiment 1. Reference Figure 11As shown, the device 1100 includes: a processor 1110; and a memory 1120 connected to the processor 1110, used to provide the processor 1110 with instructions to process the following steps: acquiring first weather conditions corresponding to each time period within multiple sample periods, and yield change samples corresponding to disaster conditions at each growth stage within each sample period; acquiring second weather conditions for each time period within a target period, and determining a first correlation based on each first weather condition and the second weather condition, wherein the first correlation includes the correlation between the second weather condition and each first weather condition, and the correlation between any two first weather conditions; constructing graph structures corresponding to each growth stage based on the first correlation, yield change samples corresponding to each sample period, and virtual yield change corresponding to the target period; iteratively updating each graph structure using multiple pre-trained graph neural network models, and predicting each iteratively updated graph structure using multiple pre-trained yield prediction models to determine the actual yield change corresponding to each growth stage within the target period; and fusing the actual yield change corresponding to each growth stage, and assessing the disaster loss within the target period based on the fusion result.

[0121] Therefore, according to this embodiment, it is possible to achieve the technical effect of predicting changes in crop yield based on weather conditions and taking targeted preventive measures based on disaster assessment results.

[0122] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0123] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0124] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0125] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0126] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0127] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0128] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for assessing meteorological disaster losses to crops, characterized in that, include: Obtain the first weather conditions corresponding to each time period within multiple sample periods, and the yield change samples corresponding to the disaster conditions at each growth stage within each sample period; Obtain the second weather conditions for each time period within the target period, and determine the first correlation based on each first weather condition and the second weather conditions, wherein the first correlation includes the correlation between the second weather conditions and each first weather condition, as well as the correlation between any two first weather conditions; Based on the first correlation, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period, a graph structure corresponding to each growth stage is constructed respectively. Multiple pre-trained graph neural network models are used to iteratively update each graph structure, and multiple pre-trained yield prediction models are used to predict each iteratively updated graph structure to determine the actual yield changes corresponding to each growth stage within the target period. as well as The actual yield changes corresponding to each growth stage are fused together, and the disaster losses within the target period are assessed based on the fusion results.

2. The method according to claim 1, characterized in that, Based on each first correlation, the operation of constructing a graph structure corresponding to each growth stage, including the yield change samples corresponding to each sample period and the virtual yield change corresponding to the target period, includes: Based on each first association relationship and a pre-set association relationship threshold, multiple second association relationships are determined; Using each second association as an edge, and with the yield change samples corresponding to each sample period and the virtual yield change corresponding to the target period as nodes, a graph structure corresponding to each growth period is constructed respectively.

3. The method according to claim 2, characterized in that, The operation of determining multiple second relationships based on each first relationship and a pre-set relationship threshold includes: Determine the relationship between each of the first association relationships and the association relationship threshold; If the first association relationship is greater than the association relationship threshold, the first association relationship is determined as the second association relationship; and If the first association is less than or equal to the association threshold, the first association is removed.

4. The method according to claim 2, characterized in that, The multiple growth periods include a first growth period and a second growth period. Using each second association as an edge and the output change samples corresponding to each sample period and the virtual output change corresponding to the target period as nodes, the operation of constructing a graph structure corresponding to each growth period includes: Using each of the second relationships as the first edge, and the first virtual yield change corresponding to the first growth period and the yield change sample corresponding to the first growth period as the first node, a first graph structure corresponding to the first growth period is constructed; and Using each of the second relationships as the second edge, and the second virtual yield change situation corresponding to the second growth period and the yield change sample corresponding to the second growth period as the second node, a second graph structure corresponding to the second growth period is constructed.

5. The method according to claim 4, characterized in that, The plurality of graph neural network models includes a first graph neural network model corresponding to the first growth stage, and the operation of iteratively updating each graph structure using the pre-trained plurality of graph neural network models includes: The first graph structure is input into the first graph neural network model, and a first node attribute vector is constructed using the calculation rules corresponding to the first node. The first node attribute vector includes the node attributes of the first node, and the product of the edge attributes of two adjacent first edges and the corresponding node attributes of the first node. The first node attribute vector is iteratively updated using multiple first MLP models in the first graph neural network model to generate node features corresponding to the first graph structure.

6. The method according to claim 5, characterized in that, The yield prediction model includes a first yield prediction model corresponding to the first growth stage, and utilizes multiple pre-trained yield prediction models to predict the actual yield changes corresponding to each growth stage within the target period. The operation includes: Multiple first target feature vectors are determined, wherein the first target feature vectors include the node features of the first node corresponding to the first virtual output change after iterative updates, and the product of the edge features of two adjacent first edges and the node features of the corresponding first node; and The plurality of first target feature vectors are input into the first yield prediction model, and the first actual yield change corresponding to the first growth period within the target period is output.

7. The method according to claim 4, characterized in that, The plurality of graph neural network models includes a second graph neural network model corresponding to the second growth stage, and iterative updates are performed on each graph structure using the pre-trained plurality of graph neural network models, including: The second graph structure is input into the second graph neural network model, and a second node attribute vector is constructed using the calculation rules corresponding to the second node. This second node attribute vector includes the node attributes of the second node, and the product of the edge attributes of two adjacent second edges and the corresponding node attributes of the second node. The second node attribute vector is iteratively updated using multiple second MLP models in the second graph neural network model to generate node features corresponding to the second graph structure.

8. The method according to claim 7, characterized in that, The yield prediction model includes a second yield prediction model corresponding to the second growth stage, and utilizes multiple pre-trained yield prediction models to predict the actual yield changes corresponding to each growth stage within the target period. The operation includes: Multiple second target feature vectors are determined, wherein the second target feature vectors include the node features of the second node corresponding to the second virtual output change after iterative updates, and the product of the edge features of two adjacent second edges and the node features of the corresponding second node; and The second target feature vector is input into the second yield prediction model, and the second actual yield change corresponding to the second growth period within the target period is output.

9. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, the method described in any one of claims 1 to 8 is performed by a processor.

10. A device for assessing meteorological disaster losses to crops, characterized in that, include: The first data acquisition module is used to acquire the first weather conditions corresponding to each time period within multiple sample periods, and the yield change samples corresponding to the disaster conditions of each growth stage within each sample period. The second data acquisition module is used to acquire the second weather conditions for each time period within the target period, and determine the first correlation based on each first weather condition and the second weather condition, wherein the first correlation includes the correlation between the second weather condition and each first weather condition, as well as the correlation between any two first weather conditions; The graph structure construction module is used to construct graph structures corresponding to each growth stage based on the first association relationship, the yield change samples corresponding to each sample period, and the virtual yield change corresponding to the target period. The prediction module is used to iteratively update each graph structure using multiple pre-trained graph neural network models, and to predict each iteratively updated graph structure using multiple pre-trained yield prediction models, so as to determine the actual yield changes corresponding to each growth stage within the target period. as well as The evaluation module is used to integrate the actual yield changes corresponding to each growth stage and to evaluate the disaster losses within the target period based on the integration results.

Images