A method, device, and storage medium for early warning of crop yield.

Abstract

This application discloses a method, device, and storage medium for early warning of crop yield. It constructs a pre-defined probability distribution of crop yield in a target region, quantifying the probabilistic patterns of the crop's yield in that region. When it is necessary to determine whether a yield warning is needed for a specific time period, the computing device acquires meteorological data for that period, predicts the crop yield under the influence of meteorological conditions during that period, and then, based on the pre-determined yield probability distribution, determines whether the predicted yield is abnormal (i.e., affected by meteorological conditions), thereby determining whether a yield warning is needed. Therefore, this application uses the yield probability distribution to determine whether meteorological conditions have an adverse impact on crop yield during the continuous reproductive growth period, enabling the analysis and early warning of the impact of non-extreme weather on crop yield.

Description

Technical Field

[0001] This application relates to the field of agricultural and crop yield early warning technology, and in particular to a method, device and storage medium for early warning of crop yield. Background Technology

[0002] With the continuous advancement of technology, agriculture-related industries are gradually entering a process of modernization and intelligentization.

[0003] Current research on the impact of meteorological conditions on crop yields primarily focuses on extreme weather and natural disasters. For example, it involves issuing early warnings related to crop yields in the event of potential droughts or high temperatures. However, in reality, even slight fluctuations in meteorological conditions can adversely affect crops. For instance, fluctuations in temperature, humidity, and light intensity during the crop growing season can impact yields. While such fluctuations may not cause large-scale yield reductions, they can still result in lower-than-expected yields. Therefore, current technologies neglect the impact of non-extreme weather on crop yields; that is, they ignore early warnings based on the influence of non-extreme weather on crop yields.

[0004] There is currently no effective solution to the problem that the existing technologies mentioned above ignore the impact of non-extreme weather on crop yields, i.e., they ignore the technical problem of crop yield early warning based on the impact of non-extreme weather on crop yields. Summary of the Invention

[0005] The embodiments of this disclosure provide a method, apparatus, and storage medium for early warning of crop yield, so as to at least solve the technical problem of the impact of non-extreme weather on crop yield in the prior art.

[0006] According to one aspect of the present disclosure, a method for early warning of crop yield is provided, comprising: acquiring historical yield information corresponding to a preset crop in a target area; determining a yield probability distribution corresponding to the preset crop in the target area based on the historical yield information; acquiring meteorological data within a preset time period, the preset time period including the reproductive growth period of the preset crop; predicting the yield of the preset crop under the influence of meteorological conditions within the preset time period based on the meteorological data, thereby obtaining a yield prediction value; and determining whether to issue a yield warning for the preset crop based on the yield prediction value and the yield probability distribution.

[0007] 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.

[0008] According to another aspect of the present disclosure, an early warning device for crops is also provided, comprising: a first acquisition module for acquiring historical yield information corresponding to a preset crop in a target area; a probability distribution determination module for determining a yield probability distribution corresponding to the preset crop in the target area based on the historical yield information; a second acquisition module for acquiring meteorological data within a preset time period, the preset time period including the reproductive growth period of the preset crop; a prediction model for predicting the yield of the preset crop under the influence of meteorological conditions within the preset time period based on the meteorological data, thereby obtaining a yield prediction value; and an early warning module for determining whether to issue a yield warning for the preset crop based on the yield prediction value and the yield probability distribution.

[0009] According to another aspect of the present disclosure, a crop yield early warning device 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 historical yield information corresponding to a preset crop in a target area; determining a yield probability distribution corresponding to the preset crop in the target area based on the historical yield information; acquiring meteorological data within a preset time period, the preset time period including the reproductive growth period of the preset crop; predicting the yield of the preset crop under the influence of meteorological conditions within the preset time period based on the meteorological data, obtaining a yield prediction value; and determining whether to issue a yield early warning for the preset crop based on the yield prediction value and the yield probability distribution.

[0010] In this embodiment, the method first constructs a probability distribution of the yield of a preset crop in a target area. This probability distribution quantifies the probabilistic patterns of the crop's yield in the target area. Then, when it is necessary to determine whether a yield warning for a preset crop is needed for a specific time period, the computing device can acquire meteorological data for that period and predict the crop's yield under the influence of meteorological conditions during that period. Based on the pre-determined yield probability distribution, it determines whether the predicted yield is abnormal and whether a yield warning is needed. Therefore, in this application, the yield probability distribution can be used to determine whether meteorological conditions during the crop's reproductive growth period have an adverse impact on yield, thus enabling the analysis and early warning of the impact of non-extreme weather on crop yield. Attached Figure Description

[0011] 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 according to Embodiment 1 of this disclosure; Figure 2 This is a schematic flowchart of a crop yield early warning method according to the first aspect of Embodiment 1 of this disclosure; Figure 3A This is a flowchart illustrating a process for predicting the yield of a preset crop under the influence of meteorological conditions within a preset time period, as provided in Embodiment 1 of this disclosure. Figure 3B This is a schematic diagram of the structure of a prediction model provided in Embodiment 1 of this disclosure; Figure 3C This is a schematic diagram of the structure of a feature extraction subnetwork in a prediction model provided in Embodiment 1 of this disclosure; Figure 3D This is a schematic diagram of the output prediction subnetwork in a prediction model provided in Embodiment 1 of this disclosure; Figure 4 This is a schematic diagram of a crop yield early warning device according to the first aspect of Embodiment 2 of this disclosure; and Figure 5 This is a schematic diagram of a crop yield early warning device according to the first aspect of Embodiment 3 of this disclosure. Detailed Implementation

[0012] 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.

[0013] 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.

[0014] Example 1 According to this embodiment, a method embodiment for early warning of crop yield 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.

[0015] 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 early warning of crop yields 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.

[0016] 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).

[0017] 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 method for early warning of crop yield 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 method for early warning of crop yield in the application. 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 such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0018] 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.

[0019] 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.

[0020] 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 1 This 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.

[0021] Under the aforementioned operating environment, according to the first aspect of this embodiment, a method for early warning of crop yield is provided. This method can be... Figure 1 The computing device implementation is shown. Figure 2 A flowchart illustrating the method is shown below. (Refer to...) Figure 2 As shown, the method includes: S202: Obtain historical yield information corresponding to preset crops in the target area; S204: Based on the historical yield information, determine the yield probability distribution corresponding to the preset crops in the target area; S206: Obtain meteorological data for a preset time period; S208: Based on the meteorological data, predict the yield of a preset crop under the influence of meteorological conditions within a preset time period, and obtain the predicted yield value; and S210: Based on the predicted output value and the output probability distribution, determine whether to issue an output warning for the preset time period.

[0022] Specifically, the computing device can acquire historical yield information corresponding to preset crops in the target area (S202), and determine the yield probability distribution corresponding to preset crops in the target area based on the historical yield information (S204).

[0023] The target region can refer to an area requiring meteorological condition warnings for crops, such as a specific city, district, or town. The selection of the target region can be manually set. Since different crops have different growth periods and suitable growing conditions, this method provides yield warnings based on meteorological conditions for a pre-defined crop within the target region. The specific pre-defined crop can be manually selected. In other words, this method can provide meteorological condition warnings for a specific crop in a specific region.

[0024] The historical yield information mentioned above can be obtained by statistically analyzing the yields of the target crop in the target area over many years. Of course, to supplement the sample size (i.e., the amount of historical yield data used to construct the yield probability distribution), historical yield information from other regions adjacent to the target area can be obtained to supplement the sample size, thereby fully constructing the yield probability distribution corresponding to the target crop in the target area.

[0025] How to construct the yield probability distribution corresponding to the preset crops in the target area will be introduced later.

[0026] The computing device can acquire meteorological data within a preset time period (S206), and then predict the yield of a preset crop under the influence of the meteorological conditions corresponding to the meteorological data within the preset time period, thus obtaining a yield prediction value (S208). The preset time period includes the reproductive growth period of the preset crop.

[0027] It should be noted that the preset time period can be the reproductive growth period of the preset crop, and thus the preset time period includes each preset growth cycle in the reproductive growth period of the preset crop. Different preset growth cycles have different effects on the preset crop.

[0028] This embodiment uses winter wheat as the preset crop for illustration. The reproductive growth period of winter wheat is from April to May each year (this period is a sensitive period for winter wheat growth), including: the booting stage (early to mid-April), the heading-flowering stage (late April to early May), and the grain-filling stage (mid to late May). The environmental requirements differ for each of these different periods. Therefore, for winter wheat, the different preset growth cycles corresponding to this winter wheat can be set as the booting stage, the heading-flowering stage, and the grain-filling stage, respectively.

[0029] Specifically, for winter wheat, when the preset crop is winter wheat, the preset time period can be the reproductive growth period of winter wheat, i.e., April to May. For example, in this embodiment, actual meteorological data from April to May can be obtained after May ends, serving as the meteorological data for the preset time period. Yield prediction can then be made using this meteorological data to forecast the yield of winter wheat that has undergone this reproductive growth period. Alternatively, meteorological data for the completed reproductive growth period (i.e., April 1 to May 10) can be obtained on May 10th, along with meteorological data for May 11 to May 30 (i.e., the remaining reproductive growth period) obtained through weather forecasts. These combined data form the meteorological data for the preset time period, and yield prediction can then be made using this meteorological data to forecast the yield of winter wheat that has undergone this reproductive growth period.

[0030] After determining the yield under the influence of meteorological conditions during the preset time period, the computing device can determine whether to issue a yield warning for the preset crop based on the corresponding yield forecast and yield probability distribution (S210).

[0031] In other words, the purpose of step S210 is to predict the yield affected by the weather conditions during the preset period, and to determine whether the final yield may be abnormal due to the weather conditions during the preset period (i.e., whether it will lead to a reduction in the yield of the preset crop) by using the predetermined yield probability distribution and the predicted yield, thereby determining whether to issue a yield warning for the preset crop.

[0032] As described in the background section, current research on the impact of meteorological conditions on crop yields primarily focuses on extreme weather and natural disasters. For example, early warnings related to crop yields are issued when extreme weather events such as drought or high temperatures are likely. However, in reality, even slight fluctuations in meteorological conditions can adversely affect crops. For instance, fluctuations in temperature, humidity, and light intensity during the crop growing season can impact yields. While such weather fluctuations may not cause large-scale yield reductions, they can still result in lower-than-expected yields. Therefore, current technologies neglect the impact of non-extreme weather on crop yields; that is, they neglect early warnings regarding the impact of non-extreme weather on crop yields.

[0033] In view of this, this application first constructs a probability distribution of the yield of a preset crop in a target area. This probability distribution quantifies the probabilistic patterns of the crop's yield in the target area. Then, when it is necessary to determine whether a yield warning is needed for a certain period, the computing device can acquire meteorological data for that period and predict the yield of the preset crop under the influence of the meteorological conditions corresponding to that period. Based on the pre-determined yield probability distribution, it is determined whether the predicted yield is abnormal, and thus whether a yield warning is needed for the preset crop. Therefore, in this application, the yield probability distribution can be used to determine whether meteorological conditions during the continuous reproductive growth period of a crop have an adverse impact on yield, thereby enabling the analysis and early warning of the impact of non-extreme weather on crop yield.

[0034] Optionally, the operation of determining the probability distribution of yield corresponding to a preset crop in the target area based on historical yield information includes: determining the logarithm of the historical yield information; determining the variance and mean corresponding to the logarithm of the historical yield information based on the logarithm of the historical yield information; and determining the log-normal distribution corresponding to the yield of the preset crop in the target area based on the determined variance and mean, as the probability distribution of yield corresponding to the preset crop in the target area. Specifically, determining whether to issue a yield warning for the preset crop based on the predicted yield value and the yield probability distribution includes: determining whether to issue a yield warning for the preset crop based on the log-normal distribution and the logarithm of the predicted yield value.

[0035] Since crop yields are data greater than 0, when determining the yield probability distribution, we can determine the log-normal distribution corresponding to the yield of the preset crop in the target area, and use it as the yield probability distribution corresponding to the preset crop in the target area.

[0036] Continuing with the example of winter wheat, before determining the probability distribution of yield, historical yield information was identified. Let the sample of historical yield per mu (a Chinese unit of area, approximately 0.067 hectares) of winter wheat in the target area be X = {x1, x2, ..., x...} M}

[0037] Specifically, it is possible to determine each sample x. m Logarithm of (m=1~M) Then, calculate x for each sample. m logarithm y m Given the mean μ and variance σ, the log-normal distribution corresponding to the sample X (i.e., the log-normal distribution corresponding to the winter wheat yield in the target region) can be determined based on the mean μ and variance σ: Among them, in the formula Indicates output, Let be the probability density function.

[0038] Optionally, meteorological data may include: daily average temperature, daily maximum temperature, daily minimum temperature, sunshine duration, and relative soil moisture content, etc. Daily meteorological data for a preset time period can be obtained. However, the input parameters (meteorological data) of the prediction model are not limited here, and other input parameters can be added. For example, meteorological data may also include daily precipitation.

[0039] The operation of predicting the yield of a preset crop under the influence of meteorological conditions within a preset time period based on meteorological data includes: inputting meteorological data into a pre-trained prediction model to obtain output results, the output results including a first probability value under each preset yield interval; and determining the yield prediction value based on the first probability value under each preset yield interval and the median yield of each preset yield interval.

[0040] Figure 3A This is a flowchart illustrating a process for predicting the yield of a preset crop under the influence of meteorological conditions corresponding to meteorological data within a preset time period, as provided in Embodiment 1 of this disclosure.

[0041] like Figure 3A As shown, after meteorological data is input into the prediction model, the prediction model can output multiple first probability values. Each first probability value corresponds to a preset yield range. That is, the number of first probability values ​​will be output as many as there are preset yield ranges. Figure 3A In the diagram, I represents the input climate parameters (meteorological data), and Q = [q1, q2, q3, ..., q]. J The values ​​represent the probability values ​​for J preset yield intervals. Each preset yield interval corresponds to a range of yields per acre, and the J preset yield intervals are consecutive. For example, the first preset yield interval is [0~10 kg], the second preset yield interval is [10 kg~20 kg], the third preset yield interval is [20 kg~30 kg], and so on. These examples are merely illustrations of the preset yield intervals; in practice, each preset yield interval can be determined based on the historical actual yields of the target crop in the target region.

[0042] To improve forecast accuracy, the preset yield ranges can be set with finer granularity. For example, each preset yield range can be defined in 10 kg increments, with the median values ​​for each range being y1, y2, y3, ..., y J .

[0043] Therefore, the production assessment module can assess the production based on Q output by the prediction model to obtain the predicted production value R', calculated using the following formula: Specifically, the unit of the predicted yield value can be yield per acre. That is, the median value and the corresponding probability value corresponding to each preset yield interval can be multiplied to obtain the multiplied result, and then the multiplied results of each preset yield interval can be summed to obtain the predicted yield value.

[0044] Optionally, before inputting meteorological data into a pre-trained prediction model to obtain the output result, the method further includes: acquiring sample meteorological data and sample yield values ​​corresponding to the sample meteorological data; inputting the sample meteorological data into the prediction model to obtain the prediction result output by the prediction model; the prediction result includes a second probability value under each preset yield interval; determining the yield prediction value corresponding to the sample meteorological data based on the second probability value under each preset yield interval and the median yield of each preset yield interval; and training the prediction model with the training objective of minimizing the difference between the sample yield value and the yield prediction value corresponding to the sample meteorological data.

[0045] Specifically, a sample set can be constructed first, which corresponds to the meteorological data of the target area and the preset crop yield. The sample set can be constructed only for the reproductive growth period (a sensitive period for crop growth), and then the prediction model can be trained.

[0046] Specifically, historical meteorological data (sample meteorological data) for each reproductive growth stage of the target region (and adjacent regions) and the actual yield of the corresponding preset crop (sample yield value) can be obtained. A reproductive growth stage of the preset crop can be divided into multiple preset growth cycles, and the sample meteorological data includes meteorological data within each preset growth cycle. Since different preset growth cycles within the reproductive growth stage have different impacts on the preset crop, during the training phase, the prediction model can extract features from data within different reproductive growth stages of a single sample (sample meteorological data), then fuse the data from different reproductive growth stages (weighted fusion) to obtain the output of the prediction model. The prediction model is trained with the objective of minimizing the difference between the predicted yield value determined by this output and the actual sample yield value.

[0047] Therefore, the prediction model can be trained using the constructed sample set. During training, the computing device inputs the meteorological data corresponding to a specific sample into the prediction model. The prediction model outputs a second probability value for each preset yield interval corresponding to that sample. Then, through the yield assessment module, based on the formula used in the aforementioned yield assessment module, and according to the second probability value output by the prediction model for each preset yield interval and the median value corresponding to each yield interval, the predicted yield value corresponding to that sample is calculated. The prediction model can be trained with the goal of minimizing the difference between the sample yield value and the predicted yield value corresponding to the same sample (i.e., supervised training of the prediction model).

[0048] The loss function used to train the prediction model can be specifically shown below: Where N is the number of samples, Let n be the predicted output value corresponding to the nth sample. This is the sample output value corresponding to the nth sample.

[0049] Continuing with the example of winter wheat, a reproductive growth period for winter wheat includes three pre-defined growth cycles: the booting stage (approximately April 1st to April 20th), the heading-flowering stage (approximately April 21st to May 10th), and the grain-filling stage (approximately May 11th to May 30th). For winter wheat, a single sample of meteorological data can include meteorological data D for the booting stage. a Meteorological data during the heading-flowering period (D) b And meteorological data during the grouting period D c One sample of meteorological data may include meteorological data D for the heading stage within one year. a Meteorological data during the heading-flowering period (D) b And meteorological data during the grouting period D c .

[0050] Specifically, for the meteorological data within each preset growth cycle, the covariance matrix corresponding to that meteorological data can be determined. Then, eigenvalue decomposition is performed on the covariance matrix to obtain the eigenvectors [λ1, λ2, ..., λ] corresponding to the covariance matrix. n [and the transformation matrix associated with this eigenvector. Where n is the number of types of meteorological data.]

[0051] Take meteorological data D from the heading stage of a sample meteorological data set. a Let's take an example (assuming the heading stage in this sample of meteorological data is 20 days). The meteorological data includes five types: daily average temperature, daily maximum temperature, daily minimum temperature, sunshine duration, and relative soil moisture content.a The matrix form is shown below: in, It represents the average temperature of each day over 20 days (i.e., the daily average temperature). This represents the maximum daily temperature over 20 days (i.e., the maximum daily temperature). This represents the minimum daily temperature over 20 days. This indicates the duration of sunshine on each day within a 20-day period. This indicates the relative soil moisture content for each day over 20 days.

[0052] D can be determined a The corresponding covariance matrix P a ∈R 5×5 This covariance matrix can represent the correlation between meteorological data in various dimensions (daily average temperature, daily maximum temperature, daily minimum temperature, sunshine duration, and relative soil moisture content) during the heading stage. Then, this covariance matrix P... a Perform eigenvalue decomposition, that is: P a =T a ∧T a T (1) Among them, T a That is, the transformation matrix T mentioned above that is related to the eigenvectors. a Let be an orthogonal matrix, and let ∧ be a diagonal matrix. The eigenvector β can be obtained from matrix ∧. a =[λ1, λ2, λ3, λ4, λ5], that is, the elements on the diagonal of matrix ∧ form the eigenvector β. a .

[0053] The following formula (2) can be used to determine the relationship between meteorological data D and the data. a The corresponding covariance matrix: (2) Here, tr represents the trace of the matrix.

[0054] And the covariance matrix P is calculated using the following formula (3). a Feature extraction: (3) Here, diag() means to determine the elements on the diagonal of the matrix within the parentheses and combine them into a vector.

[0055] Similarly, for meteorological data D during the heading-flowering period... b And meteorological data during the grouting period Dc In all cases, the corresponding covariance matrix can be determined in the same way, and then the corresponding transformation matrix and eigenvector can be obtained. That is, the covariance matrix and eigenvector related to D can be determined. b The corresponding covariance matrix P b , with D c The corresponding covariance matrix P c And further determine the relationship with the covariance matrix P b The corresponding eigenvector β b and the corresponding transformation matrix T b , and covariance matrix P c The corresponding eigenvector β c and the corresponding transformation matrix T c .

[0056] refer to Figure 3B As shown, the prediction model can include a feature extraction sub-network, a feature fusion module, and a yield prediction sub-network. For each preset growth cycle, the meteorological data, transformation matrix, and feature vector within that preset growth cycle can be input into the feature extraction sub-network to obtain the meteorological features corresponding to that preset growth cycle. Figure 3B The example uses three preset growth cycles of winter wheat (booting stage, heading-flowering stage, and grain-filling stage) to illustrate how the feature extraction subnetwork can obtain the meteorological characteristics corresponding to each preset growth cycle. Figure 3B The meteorological characteristics a during the booting stage, b during the heading-flowering stage, and c during the grain-filling stage were obtained.

[0057] The feature extraction subnetwork may include a feature extraction subnetwork corresponding to each preset growth cycle, to extract features from the meteorological data of the corresponding preset growth cycle. Alternatively, the same feature extraction subnetwork may be used to extract features from the meteorological data of each preset growth cycle (i.e., the feature extraction subnetworks for meteorological data of different preset growth cycles share weights).

[0058] Then, the meteorological features of different preset growth cycles can be fused using the feature fusion module to obtain fused features. These fused features are then input into the yield prediction sub-network to obtain the final output (i.e., the probability values ​​corresponding to each preset yield interval). The operation of fusing meteorological features of different preset growth cycles through the feature fusion module can include operations such as splicing and summing the meteorological features. Furthermore, since the impact of weather on the final yield of the preset crop may vary depending on the preset growth cycle, a weight can be determined for each preset growth cycle. According to the weights corresponding to the preset growth cycles, the feature fusion module can perform weighted fusion of meteorological features from different preset growth cycles (e.g., weighted summation or weighted splicing) to obtain fused features. The weight of each preset growth cycle is related to its degree of influence on the yield of the preset crop. The weights of each preset growth cycle can be pre-set. For example, for winter wheat, the heading-flowering stage has the greatest impact on yield, the grain-filling stage has a slightly smaller impact, and the booting stage has the least impact. Therefore, the weight w for the booting stage is... a Heading-flowering period w b and the early stage of grouting w c The weights can be set according to the following relationship: w b >w a >w c That is, the weight corresponding to a preset growth cycle is positively correlated with the degree of influence of meteorological conditions on the preset crop yield within that preset growth cycle.

[0059] For details, please refer to Figure 3C A feature extraction subnetwork may include convolutional and fully connected layers corresponding to meteorological data D, a feedforward neural network corresponding to the transformation matrix T, and a multilayer perceptron for generating meteorological features. This explanation focuses on the heading stage; the feature extraction process is the same for other preset growth cycles. Specifically, the meteorological data D from the heading stage can be... a The data is input into a convolutional layer, and the convolutional layer processes the meteorological data D. a After performing convolution to obtain the convolution result, the convolution result is input into the fully connected layer, and the output of the fully connected layer is compared with the meteorological data D. a The corresponding feature vectors. The convolutional kernel of the convolutional layer can be, for example, a 5×3 kernel (the number of rows in the kernel is the same as the dimension of the meteorological data). Simultaneously, the transformation matrix T can be... a The vectors are expanded and concatenated and input into the feedforward neural network. The output of the feedforward neural network is the transformation matrix T. a The corresponding feature vector. Then, it will be compared with the meteorological data D. a The corresponding eigenvectors and transformation matrix T aThe corresponding eigenvector, and the eigenvector β a The inputs are fed into a multilayer perceptron (MLP) to obtain meteorological feature a.

[0060] refer to Figure 3D As shown, the output prediction subnetwork can consist of a neural network (such as a multilayer perceptron) and multiple (J) softmax layers. One softmax layer outputs a probability value for a preset output range. After the feature fusion model completes the feature fusion operation and obtains the fused features, these features can be input into the output prediction subnetwork. Through multiple softmax layers, the probability values ​​corresponding to each preset output range can be obtained.

[0061] It should be noted that the computing device can also input meteorological data, the corresponding covariance matrix, eigenvectors, and transformation matrix corresponding to a preset growth cycle into the feature extraction sub-network (or input the meteorological data, covariance matrix, and eigenvectors corresponding to the preset growth cycle into the feature extraction sub-network) to obtain the meteorological features corresponding to that preset growth cycle. This allows the feature extraction sub-network to fully utilize the correlation of the meteorological data within a preset growth cycle, the correlation between different dimensions, and the importance of each dimension to determine the meteorological features. Consequently, subsequent yield predictions based on the meteorological features of each preset growth cycle can be more accurate.

[0062] Furthermore, when predicting the yield of a preset crop using a trained prediction model, it is also necessary to extract features from meteorological data according to different preset growth cycles to obtain the prediction results.

[0063] Optionally, the prediction model includes a feature extraction subnetwork and a yield prediction subnetwork. Meteorological data is input into the pre-trained prediction model to obtain output results. Specifically, this includes: determining meteorological data for each preset growth cycle within a preset time period; for each preset growth cycle, determining the covariance matrix corresponding to the meteorological data within that preset growth cycle, and performing eigenvalue decomposition on the covariance matrix to obtain eigenvectors corresponding to the covariance matrix and transformation matrices related to the eigenvectors; for each preset growth cycle, generating meteorological features corresponding to that preset growth cycle based on the feature extraction subnetwork, according to the meteorological data, eigenvectors, and transformation matrices corresponding to that preset growth cycle; and based on the preset time period... The meteorological characteristics corresponding to each preset growth cycle within a preset time period are used to obtain the output result through the yield prediction sub-network. The prediction model also includes a feature fusion module. Specifically, this involves determining the weight corresponding to each preset growth cycle, and then using the feature fusion module to weight and fuse the meteorological characteristics corresponding to each preset growth cycle to obtain fused features. The weight corresponding to each preset growth cycle is positively correlated with the degree of influence of meteorological conditions within that cycle on the preset crop yield. Finally, the fused features are input into the yield prediction sub-network to obtain the output result.

[0064] In other words, when using a prediction model to predict actual yield, meteorological data within a preset time period can be divided into meteorological data for each preset growth cycle. Then, through a feature extraction sub-network, the meteorological characteristics of the meteorological data for each preset growth cycle can be determined. Based on the meteorological characteristics of the meteorological data for each preset growth cycle, the yield prediction sub-network can be used to predict the yield under the influence of meteorological conditions within the preset time period.

[0065] Consistent with the training phase, for a preset growth cycle, the meteorological data D corresponding to that preset growth cycle can be determined. i The corresponding covariance matrix P i Then, for this covariance matrix P i Perform eigenvalue decomposition to obtain P i =T i ∧T i T Thus, the transformation matrix T is obtained. i and eigenvector β i (Obtained through a diagonal matrix ∧, where the elements on the diagonal of the diagonal matrix constitute β) i =[λ1, λ2, λ3, λ4, λ5]). Then, the meteorological data D for the preset growth cycle is... i Transformation matrix Ti and eigenvector β i All data are input into the feature extraction subnetwork to obtain meteorological features corresponding to the preset growth cycle. Thus, the feature extraction subnetwork can be used to obtain meteorological features corresponding to different preset growth cycles.

[0066] Then, according to the weights corresponding to each preset growth cycle, the meteorological features corresponding to each preset growth cycle can be weighted and fused to obtain the fused features. The fused features are then input into the yield prediction sub-network to obtain the output result (i.e., the first probability value under each preset yield interval).

[0067] Based on this, this embodiment can extract features from meteorological data within different preset growth cycles of a preset crop to obtain meteorological features corresponding to each preset growth cycle. Then, it can perform weighted fusion of the meteorological features of each preset growth cycle according to the impact of different preset growth cycles on the yield of the preset crop. Finally, it can predict the yield based on the fused features, which more meticulously depicts the characteristics of meteorological data for each preset growth cycle and allows the prediction model to fully consider the impact of different preset growth cycles on the yield of the crop, thereby improving the accuracy of the crop yield prediction to a certain extent.

[0068] Optionally, based on the yield forecast and the yield probability distribution, determine whether to issue a yield warning for a preset crop, including: issuing a yield warning for the preset crop if the yield forecast does not fall within the confidence interval corresponding to the preset confidence level.

[0069] Since the determined yield probability distribution is a log-normal distribution, the logarithm Y' of the predicted yield can be determined, and then it can be determined whether this logarithm falls within the confidence interval corresponding to a preset confidence level. The preset confidence level mentioned above can be set manually. For example, the preset confidence level can be set to 95%. If the logarithm does not fall within the 95% confidence interval of the log-normal distribution, an early warning can be issued. Alternatively, other methods can be used to determine whether to issue a yield warning. For example, the probability value corresponding to the predicted yield in the determined yield probability distribution can be determined. If this probability value is lower than a preset probability value, a yield warning can be issued for a specific crop.

[0070] It should be noted that when the computing device issues a yield warning for the preset crop, it can generate corresponding warning information. This warning information can then be displayed to technical personnel involved in the planting of the preset crop (e.g., on a website accessible to them), or sent via SMS or email to their mobile devices (e.g., smartphones). Upon receiving the warning information, the relevant technical personnel can determine whether to take appropriate measures (such as protective or remedial measures) for the preset crop.

[0071] 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.

[0072] Therefore, according to this embodiment, this method can predict the yield of a preset crop after it has been affected by meteorological conditions within a preset time period, and determine whether the meteorological conditions within the corresponding preset time period have a certain adverse effect on the crop, thereby determining whether to issue a yield warning for the preset crop.

[0073] 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.

[0074] 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.

[0075] Example 2 Figure 4 A crop yield warning device 400 according to the first aspect of this embodiment is shown, the yield warning device 400 corresponding to the method described according to the first aspect of Embodiment 1. Reference Figure 4As shown, the yield early warning device 400 includes: a first acquisition module 410, used to acquire historical yield information corresponding to a preset crop in a target area; a probability distribution determination module 420, used to determine the yield probability distribution corresponding to the preset crop in the target area based on the historical yield information; a second acquisition module 430, used to acquire meteorological data within a preset time period, which includes the reproductive growth period of the preset crop; a prediction module 440, used to predict the yield of the preset crop under the influence of meteorological conditions within the preset time period based on the meteorological data, and obtain a yield prediction value; and an early warning module 450, used to determine whether to issue a yield early warning for the preset crop based on the yield prediction value and the yield probability distribution.

[0076] Optionally, meteorological data may include at least one of the following: daily average temperature, daily maximum temperature, daily minimum temperature, sunshine duration, and relative soil moisture content.

[0077] Optionally, the prediction module 440 is specifically used to input meteorological data into a pre-trained prediction model to obtain output results, the output results including a first probability value under each preset yield interval; and to determine the yield prediction value based on the first probability value under each preset yield interval and the median yield of each preset yield interval.

[0078] Optionally, the prediction model includes: a feature extraction subnetwork and a yield prediction subnetwork; the prediction module 440 is specifically used to: determine the meteorological data for each preset growth cycle within a preset time period; for the meteorological data within each preset growth cycle, determine the covariance matrix corresponding to the meteorological data within that preset growth cycle, and perform eigenvalue decomposition on the covariance matrix to obtain the eigenvectors corresponding to the covariance matrix and the transformation matrix related to the eigenvectors; for each preset growth cycle, based on the meteorological data, eigenvectors, and transformation matrix corresponding to that preset growth cycle, generate meteorological features corresponding to that preset growth cycle based on the feature extraction subnetwork; and based on the meteorological data, eigenvectors, and transformation matrix corresponding to each preset growth cycle within a preset time period... The meteorological characteristics are used to obtain the output result through the yield prediction sub-network. The prediction model also includes a feature fusion module. Specifically, this involves determining the weight corresponding to each preset growth cycle within a preset time period, and then using the feature fusion module to weight and fuse the meteorological characteristics corresponding to each preset growth cycle to obtain fused features. The weight corresponding to each preset growth cycle is positively correlated with the degree of influence of meteorological conditions within that cycle on the preset crop yield. Finally, the fused features are input into the yield prediction sub-network to obtain the output result.

[0079] Optionally, the probability distribution determination module 420 is specifically used to: determine the logarithm of historical yield information; determine the variance and mean corresponding to the logarithm of historical yield information based on the logarithm of historical yield information; and determine the log-normal distribution corresponding to the yield of a preset crop in the target area based on the determined variance and mean, as the yield probability distribution corresponding to the preset crop in the target area. Specifically, determining whether to issue a yield warning for the preset crop based on the yield prediction value and the yield probability distribution includes: determining whether to issue a yield warning for the preset crop based on the log-normal distribution and the logarithm of the yield prediction value.

[0080] Optionally, before inputting meteorological data into a pre-trained prediction model to obtain the output result, the device 400 further includes: a training module 460, used to acquire sample meteorological data and sample yield values ​​corresponding to the sample meteorological data; input the sample meteorological data into the prediction model to obtain the prediction result output by the prediction model; the prediction result includes a second probability value under each preset yield interval; determine the yield prediction value corresponding to the sample meteorological data based on the second probability value under each preset yield interval and the median yield of each preset yield interval; and train the prediction model with the training objective of minimizing the difference between the sample yield value and the yield prediction value corresponding to the sample meteorological data.

[0081] Optionally, the early warning module 450 is specifically used to issue an early warning for a preset crop when the predicted yield value does not fall within the confidence interval corresponding to the preset confidence level.

[0082] Therefore, according to this embodiment, this method can predict the yield of a preset crop after it has been affected by meteorological conditions within a preset time period, and determine whether the meteorological conditions within the corresponding preset time period have a certain adverse effect on the crop, thereby determining whether to issue a yield warning for the preset crop.

[0083] Example 3 Figure 5 A crop yield early warning device according to the first aspect of this embodiment is shown, which corresponds to the method described according to the first aspect of Embodiment 1. (Reference) Figure 5As shown, the yield early warning device includes: a processor 510; and a memory 520 connected to the processor 510, used to provide the processor 510 with instructions to process the following steps: acquiring historical yield information corresponding to a preset crop in a target area; determining the yield probability distribution corresponding to the preset crop in the target area based on the historical yield information; acquiring meteorological data within a preset time period, which includes the reproductive growth period of the preset crop; predicting the yield of the preset crop under the influence of meteorological conditions within the preset time period based on the meteorological data, and obtaining a yield prediction value; and determining whether to issue a yield early warning for the preset crop based on the yield prediction value and the yield probability distribution.

[0084] Optionally, meteorological data may include at least one of the following: daily average temperature, daily maximum temperature, daily minimum temperature, sunshine duration, and relative soil moisture content.

[0085] Optionally, the operation of predicting the yield of a preset crop under the influence of meteorological conditions corresponding to meteorological data within a preset time period, based on meteorological data, includes: inputting meteorological data into a pre-trained prediction model to obtain an output result, the output result containing a first probability value under each preset yield interval; and determining the yield prediction value based on the first probability value under each preset yield interval and the median yield of each preset yield interval.

[0086] Optionally, the prediction model includes: a feature extraction subnetwork and a yield prediction subnetwork; meteorological data is input into the pre-trained prediction model to obtain output results, specifically including: determining the meteorological data for each preset growth cycle within a preset time period; for the meteorological data within each preset growth cycle, determining the covariance matrix corresponding to the meteorological data within that preset growth cycle, and performing eigenvalue decomposition on the covariance matrix to obtain the eigenvectors corresponding to the covariance matrix and the transformation matrix related to the eigenvectors; for each preset growth cycle, based on the meteorological data, eigenvectors, and transformation matrix corresponding to that preset growth cycle, generating meteorological features corresponding to that preset growth cycle based on the feature extraction subnetwork; and based on the meteorological data within the preset time period... The meteorological characteristics corresponding to each preset growth cycle are used to obtain the output result through the yield prediction sub-network. The prediction model also includes a feature fusion module. Specifically, this involves determining the weight corresponding to each preset growth cycle, and then using the feature fusion module to weight and fuse the meteorological characteristics corresponding to each preset growth cycle to obtain fused features. The weight corresponding to each preset growth cycle is positively correlated with the degree of influence of meteorological conditions within that cycle on the preset crop yield. Finally, the fused features are input into the yield prediction sub-network to obtain the output result.

[0087] Optionally, the operation of determining the probability distribution of yield corresponding to a preset crop in the target area based on historical yield information includes: determining the logarithm of the historical yield information; determining the variance and mean corresponding to the logarithm of the historical yield information based on the logarithm of the historical yield information; and determining the log-normal distribution corresponding to the yield of the preset crop in the target area based on the determined variance and mean, as the probability distribution of yield corresponding to the preset crop in the target area. Specifically, determining whether to issue a yield warning for the preset crop based on the predicted yield value and the probability distribution of yield includes: determining whether to issue a yield warning for the preset crop based on the log-normal distribution and the logarithm of the predicted yield value.

[0088] Optionally, before inputting meteorological data into a pre-trained prediction model to obtain the output result, the memory 520 is also used to provide the processor 510 with instructions to process the following steps: acquiring sample meteorological data and sample yield values ​​corresponding to the sample meteorological data; inputting the sample meteorological data into the prediction model to obtain the prediction result output by the prediction model; the prediction result includes a second probability value under each preset yield interval; determining the yield prediction value corresponding to the sample meteorological data based on the second probability value under each preset yield interval and the median yield of each preset yield interval; and training the prediction model with the training objective of minimizing the difference between the sample yield value and the yield prediction value corresponding to the sample meteorological data.

[0089] Optionally, based on the yield forecast and the yield probability distribution, determine whether to issue a yield warning for a preset crop, including: issuing a yield warning for the preset crop if the yield forecast does not fall within the confidence interval corresponding to the preset confidence level.

[0090] Therefore, according to this embodiment, this method can predict the yield of a preset crop after it has been affected by meteorological conditions within a preset time period, and determine whether the meteorological conditions within the corresponding preset time period have a certain adverse effect on the crop, thereby determining whether to issue a yield warning for the preset crop.

[0091] 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.

[0092] 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.

[0093] 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.

[0094] 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.

[0095] 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.

[0096] 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.

[0097] 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 early warning of crop yield, characterized in that, include: Obtain historical yield information corresponding to preset crops in the target region; Based on the historical yield information, determine the yield probability distribution corresponding to the preset crops in the target area; Acquire meteorological data within a preset time period, wherein the preset time period includes the reproductive growth period of the preset crop; Based on the meteorological data, the yield of the preset crop under the influence of meteorological conditions during the preset time period is predicted, and the yield prediction value is obtained. as well as Based on the predicted yield value and the yield probability distribution, it is determined whether to issue a yield warning for the preset crop, wherein the operation of determining the yield probability distribution corresponding to the preset crop in the target area based on the historical yield information includes: Determine the logarithm of the historical production information; Based on the logarithmic values ​​of the historical production information, determine the variance and mean corresponding to the logarithmic values ​​of the historical production information; and Based on the variance and the mean, a log-normal distribution corresponding to the yield of a preset crop in the target region is determined as the yield probability distribution corresponding to the preset crop in the target region. Specifically, determining whether to issue a yield warning for the preset crop based on the predicted yield value and the yield probability distribution includes: Based on the log-normal distribution and the logarithm of the predicted yield value, it is determined whether to issue a yield warning for the preset crop. The operation of predicting the yield of the preset crop under the influence of meteorological conditions within the preset time period, based on the meteorological data, to obtain the predicted yield value includes: The meteorological data is input into a pre-trained prediction model to obtain an output result, which includes a first probability value under each preset yield interval; and the predicted yield value is determined based on the first probability value under each preset yield interval and the median yield of each preset yield interval, wherein the prediction model includes: a feature extraction subnetwork and a yield prediction subnetwork. The operation of inputting the meteorological data into a pre-trained prediction model to obtain the output results specifically includes: Meteorological data for each preset growth cycle within the preset time period are determined; for the meteorological data within each preset growth cycle, a covariance matrix corresponding to the meteorological data within that preset growth cycle is determined, and the covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvectors corresponding to the covariance matrix and transformation matrices related to the eigenvectors; for each preset growth cycle, based on the meteorological data, eigenvectors, and transformation matrices corresponding to that preset growth cycle, meteorological features corresponding to that preset growth cycle are generated based on the feature extraction subnetwork; based on the meteorological features corresponding to each preset growth cycle within the preset time period, the output result is obtained through a yield prediction subnetwork, and wherein the prediction model further includes a feature fusion module; The operation of obtaining the output result through a yield prediction sub-network based on the meteorological characteristics corresponding to each preset growth cycle within the preset time period specifically includes: The weights corresponding to each preset growth cycle are determined, and based on the weights corresponding to each preset growth cycle, the meteorological features corresponding to each preset growth cycle are weighted and fused through the feature fusion module to obtain fused features. The weights corresponding to the preset growth cycles are positively correlated with the degree of influence of meteorological conditions within the preset growth cycles on the preset crop yield. The fused features are then input into the yield prediction sub-network to obtain the output results.

2. The method according to claim 1, characterized in that, The meteorological data includes at least one of the following: daily average temperature, daily maximum temperature, daily minimum temperature, sunshine duration, and relative soil moisture content.

3. The method according to claim 1, characterized in that, Before inputting the meteorological data into a pre-trained prediction model and obtaining the output, the method further includes: Obtain sample meteorological data and the corresponding sample yield values; The sample meteorological data is input into the prediction model to obtain the prediction results output by the prediction model; the prediction results include a second probability value under each preset yield range; Based on the second probability value under each preset production range and the median production value of each preset production range, determine the predicted production value corresponding to the sample meteorological data; and The prediction model is trained with the goal of minimizing the difference between the sample yield value and the yield prediction value corresponding to the sample meteorological data.

4. The method according to claim 1, characterized in that, Based on the predicted output value and the output probability distribution, determine whether to issue an output warning for the preset time period, including: If the predicted output value does not fall within the confidence interval corresponding to the preset confidence level, an output warning will be issued for the preset time period.

5. 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 4 is performed by a processor.

6. A yield early warning device for crops, characterized in that, include: The first acquisition module is used to acquire historical yield information corresponding to preset crops in the target area; The probability distribution determination module is used to determine the probability distribution of yield corresponding to a preset crop in the target area based on the historical yield information. The second acquisition module is used to acquire meteorological data within a preset time period, wherein the preset time period includes the reproductive growth period of the preset crop; The prediction module is used to predict the yield of the preset crop under the influence of meteorological conditions within the preset time period based on the meteorological data, and obtain the yield prediction value. as well as An early warning module is used to determine whether to issue a yield warning for a preset crop based on the predicted yield value and the yield probability distribution. The probability distribution determination module is used to determine the logarithm of the historical yield information; determine the variance and mean corresponding to the logarithm of the historical yield information based on the logarithm; and determine a log-normal distribution corresponding to the yield of the preset crop in the target region based on the variance and the mean, as the yield probability distribution corresponding to the preset crop in the target region. The early warning module is used to determine whether to issue a yield warning for the preset crop based on the log-normal distribution and the logarithm of the predicted yield value. Furthermore, the prediction module is used to input the meteorological data into a pre-trained prediction model to obtain an output result, the output result including a first probability value under each preset yield interval; and to determine the yield prediction value based on the first probability value under each preset yield interval and the median yield of each preset yield interval, wherein the prediction model includes: a feature extraction subnetwork and a yield prediction subnetwork. The prediction module is used to determine meteorological data for each preset growth cycle within the preset time period; for the meteorological data within each preset growth cycle, determine the covariance matrix corresponding to the meteorological data within that preset growth cycle, and perform eigenvalue decomposition on the covariance matrix to obtain the eigenvector corresponding to the covariance matrix and the transformation matrix related to the eigenvector; for each preset growth cycle, based on the meteorological data, eigenvector, and transformation matrix corresponding to that preset growth cycle, generate meteorological features corresponding to that preset growth cycle based on the feature extraction subnetwork; and obtain the output result through the yield prediction subnetwork based on the meteorological features corresponding to each preset growth cycle within the preset time period. The prediction model further includes a feature fusion module. The prediction module is used to determine the weight corresponding to each preset growth cycle, and according to the weight corresponding to each preset growth cycle, the feature fusion module performs weighted fusion of the meteorological features corresponding to each preset growth cycle to obtain fused features, wherein the weight corresponding to the preset growth cycle is positively correlated with the degree of influence of the meteorological conditions within the preset growth cycle on the preset crop yield; and the fused features are input into the yield prediction sub-network to obtain the output result.

7. A crop yield early warning device, characterized in that, include: processor; as well as A memory, connected to the processor, for providing the processor with instructions to perform the following processing steps: Obtain historical yield information corresponding to preset crops in the target region; Based on the historical yield information, determine the yield probability distribution corresponding to the preset crops in the target area; Acquire meteorological data within a preset time period, wherein the preset time period includes the reproductive growth period of the preset crop; Based on the meteorological data, the yield of the preset crop under the influence of meteorological conditions during the preset time period is predicted, and the yield prediction value is obtained. as well as Based on the predicted yield value and the yield probability distribution, it is determined whether to issue a yield warning for the preset crop, wherein the operation of determining the yield probability distribution corresponding to the preset crop in the target area based on the historical yield information includes: The process involves determining the logarithm of the historical yield information; determining the variance and mean corresponding to the logarithm of the historical yield information based on the logarithm; and determining a log-normal distribution corresponding to the yield of a preset crop in the target region based on the variance and the mean, as the yield probability distribution corresponding to the preset crop in the target region. Specifically, determining whether to issue a yield warning for the preset crop based on the predicted yield value and the yield probability distribution includes: Based on the log-normal distribution and the logarithm of the predicted yield, determine whether to issue a yield warning for the preset crop. Furthermore, the operation of predicting the yield of the preset crop under the influence of meteorological conditions within the preset time period based on the meteorological data, and obtaining the yield prediction value, includes: The meteorological data is input into a pre-trained prediction model to obtain an output result, which includes a first probability value under each preset yield interval; and the predicted yield value is determined based on the first probability value under each preset yield interval and the median yield of each preset yield interval, wherein the prediction model includes: a feature extraction subnetwork and a yield prediction subnetwork. The operation of inputting the meteorological data into a pre-trained prediction model to obtain the output results specifically includes: Meteorological data for each preset growth cycle within the preset time period are determined; for the meteorological data within each preset growth cycle, a covariance matrix corresponding to the meteorological data within that preset growth cycle is determined, and the covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvectors corresponding to the covariance matrix and transformation matrices related to the eigenvectors; for each preset growth cycle, based on the meteorological data, eigenvectors, and transformation matrices corresponding to that preset growth cycle, meteorological features corresponding to that preset growth cycle are generated based on the feature extraction subnetwork; based on the meteorological features corresponding to each preset growth cycle within the preset time period, the output result is obtained through a yield prediction subnetwork, and wherein the prediction model further includes a feature fusion module; The operation of obtaining the output result through the yield prediction sub-network based on the meteorological characteristics corresponding to each preset growth cycle within the preset time period specifically includes: determining the weight corresponding to each preset growth cycle, and, based on the weight corresponding to each preset growth cycle, performing weighted fusion of the meteorological characteristics corresponding to each preset growth cycle through the feature fusion module to obtain fused features, wherein the weight corresponding to the preset growth cycle is positively correlated with the degree of influence of meteorological conditions within the preset growth cycle on the preset crop yield; and inputting the fused features into the yield prediction sub-network to obtain the output result.

Images