Method, device and equipment for processing the impact of meteorological events on crop yield

Abstract

The present invention relates to the field of agricultural information technology, and provides a method, device and medium for processing the impact of meteorological events on crop yields. The method includes: determining a climate characteristic vector within the range of a target site according to the meteorological data, and determining the biomass of crops planted in the field according to variety parameters, soil data and field management data; biomass represents an index of crop productivity; inputting the climate characteristic vector and biomass into a crop yield prediction model, and the crop yield prediction model outputs a predicted yield; based on the predicted yield and the corresponding actual yield, determining a prediction error, and determining a risk coefficient for each climate feature based on the prediction error. The present invention can predict the impact of extreme meteorological events on crop yields, provide an important reference for the long-term planning of agricultural production, and achieve the dual goals of risk warning and response.

Description

Technical Field

[0001] The present invention relates to the field of agricultural information technology, and in particular to a method, device and equipment for processing the impact of meteorological events on crop yield. Background Art

[0002] As the impact of global climate change becomes increasingly significant, extreme weather events such as drought, floods, and frost pose a serious threat to crop production. These extreme events not only affect the normal growth of crops, but may also lead to reduced or even total crop failure, which has a huge impact on agricultural production safety and social and economic development. Therefore, accurately diagnosing and assessing the risk of extreme weather events to crop yields is of great significance for guiding agricultural production and formulating disaster prevention and mitigation measures.

[0003] There are currently three methods for studying the impact of extreme climate events on crops, namely field experiments, process-based crop models, and statistical models or machine learning methods.

[0004] First of all, the field experiment method refers to artificially controlling extreme climate variables, forcing crop growth, and analyzing the impact of extreme climate variables on crops. This method can intuitively demonstrate the impact of extreme climate on crops, but may be limited by experimental conditions and cannot fully simulate real extreme climate conditions.

[0005] The process-based crop model method fully considers factors such as environmental conditions and cultivation management measures during crop growth, and simulates crop development, organ formation, and yield formation to predict crop growth and yield under specific environments. However, it usually involves a large number of parameters and variables, and the model establishment and calibration process is relatively complicated. In addition, since the crop growth and development process is affected by many factors, the model's prediction results may have certain limitations. The large amount of climate, soil, and crop growth data required to run the crop model may also be difficult to obtain and process. In addition, when there are extreme climate events, the crop model cannot capture them well, resulting in low accuracy when exploring the impact of extreme climate events on crops.

[0006] The statistical model method collects and analyzes historical climate data and crop yield data, establishes a regression relationship between climate factors and crop yields, and then realizes prediction. The structure of the model and the mathematical formulas used are intuitive and easy to understand. The disadvantage is that it is highly dependent on the quantity and quality of data, and cannot separate the growth stage. In addition, the climate and crop growth system are complex and dynamic, and may be affected by many factors. If the environment changes significantly (such as climate change, improved planting technology, etc.), the predictive ability of historical data analysis and statistical models may be affected. The widely used linear regression model can neither capture the nonlinear relationship between climate factors and crop yields, nor the nonlinear relationship between climate factors. Summary of the invention

[0007] In order to solve the above problems, the present invention provides a method, device and equipment for processing the impact of meteorological events on crop yield.

[0008] The present invention provides a method for processing the impact of meteorological events on crop yield, comprising:

[0009] Acquire meteorological data of each station within the target station range in the first time period, as well as variety parameters, soil data and field management data of crops planted in the fields within the target station range; the target station range is divided based on preset climate conditions;

[0010] Determine a climate characteristic vector within the range of the target site according to the meteorological data, and randomly permute each climate characteristic in the climate characteristic vector a preset number of times to determine a plurality of climate characteristic vectors;

[0011] Determining the biomass of crops planted in the field according to the variety parameters, soil data and field management data; the biomass represents an indicator of crop productivity;

[0012] Inputting a plurality of climate characteristic vectors and the biomass into a crop yield prediction model respectively, and the crop yield prediction model outputs a predicted yield;

[0013] Based on the predicted unit yield and the corresponding actual unit yield, determining a prediction error, and determining a risk factor for each climate feature based on the prediction error;

[0014] The crop yield prediction model is a model for predicting crop yields, which is obtained through machine learning training by taking the climate characteristic vector and biomass determined according to the sample data, and the actual yield corresponding to the sample data as input.

[0015] According to an embodiment of the present invention, the method further includes: a step of obtaining the crop yield prediction model, comprising:

[0016] Obtain a sample data set, and divide the sample data set into a training set and a test set;

[0017] Configure the parameter grid for the initial random forest model. The parameters to be tuned include the number of trees, the maximum depth of the tree, the minimum number of samples required to split an internal node, the minimum number of samples on a leaf node, and the maximum number of features used for splitting.

[0018] Perform parameter grid search based on sample data of the training set, determine the optimal parameter combination through cross-validation and specified scoring indicators, and obtain the optimal random forest model;

[0019] The optimal random forest model is verified based on the sample data of the validation set, and the performance value of the model on the validation set is evaluated by calculating the determination coefficient and the root mean square error. When it is determined that the performance value meets the preset conditions, the optimal random forest model is used as the crop yield prediction model.

[0020] According to an embodiment of the present invention, the step of obtaining a sample data set includes:

[0021] Collecting meteorological data, soil data, field management data of each site within the target site range during the second time period and variety parameters of crops planted in the field within the target site range;

[0022] Simulate crop growth based on meteorological data, soil data, field management data and variety parameters to obtain information on biomass, potential yield per unit area and growth stage division;

[0023] Based on the division information of growth stages and the meteorological data of each station within the target station, the climate characteristics within the target station are obtained.

[0024] According to one embodiment of the present invention, the method further includes a step of acquiring variety parameters of crops, including:

[0025] Collecting meteorological data of each site within the target site range in a third time period, and soil data and field management data of the fields within the target site range;

[0026] Set different variety parameters, input the meteorological data, soil data, and field management data of each site into the crop growth simulation model, and the crop growth simulation model outputs the yield simulation value;

[0027] Based on the simulated yield value and the measured yield value, a simulation evaluation index is determined, and based on the simulation evaluation index, variety parameters of the crop variety are determined; the variety parameters are growth parameters of the planted crop variety.

[0028] According to one embodiment of the present invention, the method further includes:

[0029] Identify climate features to be analyzed that have a risk factor higher than a preset factor;

[0030] Obtaining meteorological information of each station within the target station range in the fourth time period and the actual yield of crops planted in the field within the target station range;

[0031] Collect sliding windows with different preset window sizes and step lengths to analyze the positive and negative correlations between different climate characteristics to be analyzed and actual yields in each growth stage;

[0032] Based on the positive and negative correlations between different climate characteristics to be analyzed and the actual unit yield, the impact of the meteorological events corresponding to the climate characteristics to be analyzed on the actual unit yield of crops is determined.

[0033] According to one embodiment of the present invention, the method further includes:

[0034] The actual unit yield of the crops planted in the field within the scope of the first station is obtained, the predicted unit yield is compared with the actual unit yield, and the comparison result is determined and displayed.

[0035] The present invention also provides a device for processing the impact of meteorological events on crop yield, comprising:

[0036] An acquisition module, used to acquire meteorological data of each site within a target site range in a first time period, as well as variety parameters, soil data and field management data of crops planted in a field within the target site range; the target site range is divided based on preset climate conditions;

[0037] A first determination module is used to determine a climate characteristic vector within the range of the target site according to the meteorological data, and to perform a preset number of random permutations on each climate characteristic in the climate characteristic vector to determine a plurality of climate characteristic vectors;

[0038] A second determination module is used to determine the biomass of crops planted in the field according to the variety parameters, soil data and field management data; the biomass represents an indicator of crop productivity;

[0039] A prediction module, used for inputting a plurality of climate characteristic vectors and the biomass into a crop yield prediction model respectively, and the crop yield prediction model outputs a predicted yield;

[0040] A processing module, configured to determine a prediction error based on the predicted unit yield and the corresponding actual unit yield, and determine a risk coefficient of each climate feature based on the prediction error;

[0041] The crop yield prediction model is a model for predicting crop yields, which is obtained through machine learning training by taking the climate characteristic vector and biomass determined according to the sample data, and the actual yield corresponding to the sample data as input.

[0042] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for processing the impact of any of the above-mentioned meteorological events on crop yield is implemented.

[0043] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, a method for processing the impact of any of the above-mentioned meteorological events on crop yield is implemented.

[0044] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements a method for processing the impact of any of the above-mentioned meteorological events on crop yield.

[0045] The present invention provides a method, device and equipment for processing the impact of meteorological events on crop yield. The method determines the climate characteristic vector according to meteorological data, determines the biomass of crops planted in the field according to variety parameters, soil data and field management data, and collects and establishes a model of the relationship between extreme meteorological events and crop productivity to analyze the climate characteristic vector and biomass, outputs a predicted yield, and then determines the prediction error based on the predicted yield and the corresponding actual yield, and then determines the risk coefficient of each climate characteristic based on the prediction error, so as to predict the impact of the occurrence of extreme meteorological events on crop yield, provide an important reference for the long-term planning of agricultural production, and achieve the dual goals of risk warning and response. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 It is a flow chart of a method for processing the impact of meteorological events on crop yield provided by the present invention.

[0048] Figure 2 It is a schematic diagram of the structure of a device for processing the impact of meteorological events on crop yield provided by the present invention.

[0049] Figure 3It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Combine the following Figure 1-Figure 3 The invention describes the method, device, equipment and medium for processing the impact of meteorological events on crop yield.

[0052] Figure 1 A flow chart showing the method for processing the impact of meteorological events on crop yield provided by the present invention is shown in FIG. Figure 1 , the method comprises the following steps, as follows:

[0053] Step 11, obtaining meteorological data of each site within the target site range in the first time period, as well as variety parameters, soil data and field management data of crops planted in the fields within the target site range; the target site range is divided based on preset climate conditions.

[0054] Step 12: determine the climate characteristic vector within the target site range according to the meteorological data, and perform a preset number of random permutations on each climate characteristic in the climate characteristic vector to determine a plurality of climate characteristic vectors.

[0055] Step 13: Determine the biomass of crops planted in the field based on variety parameters, soil data, and field management data; biomass is an indicator of crop productivity.

[0056] Step 14: Input multiple climate characteristic vectors and biomass into the crop yield prediction model respectively, and the crop yield prediction model outputs the predicted yield per unit area; wherein the crop yield prediction model is a model for predicting crop yield per unit area, which is obtained through machine learning training by using the climate characteristic vector and biomass determined according to the sample data, and the actual yield per unit area corresponding to the sample data as input.

[0057] Step 15: Determine the prediction error based on the predicted yield per unit area and the corresponding actual yield per unit area, and determine the risk coefficient of each climate feature based on the prediction error.

[0058] With respect to the above-mentioned steps 11 to 15, it should be noted that, in the present invention, the site is a meteorological site for collecting meteorological data. A large number of meteorological sites will be arranged nationwide. The site range is an area, which includes multiple meteorological sites. The site range is obtained based on the division of preset climatic conditions. The climatic conditions here can be mainly hydrothermal conditions. Due to the vast area and complex terrain of the North China Plain, the distribution of hydrothermal conditions at the site shows significant unevenness. Therefore, the K-means clustering algorithm can be used to classify the sites, and the sites with similar hydrothermal conditions can be classified into one category.

[0059] In the present invention, the climate is changeable, so it is necessary to obtain meteorological data within a period of time, and use the meteorological data as a benchmark to determine the climate characteristics within the above-mentioned target site. Climate characteristics are associated with meteorological events, and each meteorological event corresponds to one or more climate characteristics. The climate characteristics include typical climate characteristics and extreme climate characteristics, but the present invention is more about studying the impact of extreme meteorological events on the yield of crops, so the climate characteristics mainly analyzed by the present invention can specifically refer to extreme climate characteristics.

[0060] For the acquisition of meteorological data, the daily meteorological observation data of meteorological stations can be collected based on the daily value data set of ground climate data, including daily maximum temperature, daily minimum temperature, daily average temperature, precipitation, sunshine hours and relative humidity. When some data need to be converted into data with other meanings for subsequent use, they can be converted in an appropriate way, such as converting sunshine hours into solar radiation data.

[0061] In the present invention, different crops are divided into different varieties. For example, corn can be divided into corn No. 1, corn No. 2, etc. during the cultivation process. Different varieties of crops have different variety parameters. Variety parameters characterize the growth status of crops. For example, variety name, vernalization index, photoperiod index, maximum grain weight, potential filling rate, flowering time, flowering period, etc., are not listed here one by one.

[0062] In the present invention, soil data requires the percentage of each particle size at each soil layer depth (%), organic matter content (%), soil bulk density (BD), field water holding capacity (DUL), wilting coefficient (LL15), total nitrogen content (%), pH value and soil organic carbon (SOC), etc.

[0063] In the present invention, the field management data of the site is obtained from the observation data of the meteorological station, including sowing date, planting density, sowing depth, row spacing, irrigation amount and irrigation date, fertilization amount and fertilization date, growth period (seedling stage, flowering stage, maturity stage), yield, etc.

[0064] In the present invention, a crop model based on simulating the crop growth process is first configured. The process-based crop model method fully considers factors such as environmental conditions and cultivation management measures during crop growth, and simulates and studies processes such as crop development, organ formation, and yield formation to predict the growth status and yield of crops under specific environments. The crop model is based on crop physiological ecology principles and mathematical methods, and can accurately describe the crop growth and development process, and can comprehensively consider various factors in the crop growth process, such as soil, crop physiological and ecological characteristics, so as to more comprehensively analyze the impact of climate on crops. Therefore, the climate feature vector within the target site is determined according to meteorological data, and the climate feature vector includes at least one climate feature. The present invention aims to determine the impact of climate features on the yield risk of crops. To this end, in the present invention, RF importance analysis is an important application of the random forest model, which evaluates the contribution of each feature to the model prediction performance by permuting the features of the model and calculating the prediction error after permutation. This operation not only helps to understand the importance of each feature in the model, but also provides valuable reference for subsequent feature selection and model optimization. Therefore, each climate feature in the climate feature vector is randomly permuted a preset number of times to determine a plurality of climate feature vectors.

[0065] In the present invention, the biomass of crops planted in the field can be determined using a configured crop model based on variety parameters, soil data and field management data, and the biomass represents an indicator of crop productivity.

[0066] In the present invention, a crop yield prediction model that can predict the yield can be pre-trained and configured. At this time, the obtained multiple climate characteristic vectors and biomass are input into the crop yield prediction model, and the crop yield prediction model outputs the predicted yield. Among them, the crop yield prediction model is a model for predicting the crop yield obtained by machine learning training using the climate characteristic vector and biomass determined according to the sample data and the actual yield corresponding to the sample data as input.

[0067] The present invention uses permutation importance to rank the importance of climate features. The permutation importance method randomly permutes the value of each feature and then monitors the degree of degradation of model performance. In this process, the number of permutations of each feature and the resulting change in model prediction error (RMSE) are recorded. By comparing these changes, the degree of impact of each feature on model performance can be evaluated, thereby determining their importance ranking.

[0068] The specific method is: using the method of permutation features, randomly permuting each climate feature a preset number of times (such as 500 times), and calculating the model prediction error (RMSE) after permutation. By comparing the error changes before and after permutation, the importance score of each climate feature is obtained. In order to more intuitively show the importance of climate features, the importance scores of the top climate features are normalized to the maximum and minimum, and converted into a value between 0 and 1 as the risk coefficient of extreme meteorological events. In this way, it is easier to compare the relative importance of different climate features on crop yields. Finally, the feature name and the normalized importance score are organized into a DataFrame and output as an Excel file. Through RF importance analysis, the key features that have a greater impact on the model prediction performance in each region, that is, the extreme meteorological disaster indicators that have a greater impact on crop yields, can be quickly identified, and their risk coefficients can be quantified.

[0069] The method for processing the impact of meteorological events on crop yield provided by the present invention determines the climate characteristic vector according to meteorological data, determines the biomass of crops planted in the field according to variety parameters, soil data and field management data, collects and establishes a model of the relationship between extreme meteorological events and crop productivity to analyze the climate characteristic vector and biomass, outputs a predicted yield, and then determines the prediction error based on the predicted yield and the corresponding actual yield, and then determines the risk coefficient of each climate characteristic based on the prediction error, so as to predict the impact of the occurrence of extreme meteorological events on crop yield, provide an important reference for the long-term planning of agricultural production, and achieve the dual goals of risk warning and response.

[0070] In a further method of the above method, it should be noted that the steps of training and obtaining the crop yield prediction model are as follows:

[0071] Obtain a sample data set, and divide the sample data set into a training set and a test set;

[0072] Configure the parameter grid for the initial random forest model. The parameters to be tuned include the number of trees, the maximum depth of the tree, the minimum number of samples required to split an internal node, the minimum number of samples on a leaf node, and the maximum number of features used for splitting.

[0073] Perform parameter grid search based on sample data of the training set, determine the optimal parameter combination through cross-validation and specified scoring indicators, and obtain the optimal random forest model;

[0074] The optimal random forest model is verified based on the sample data of the validation set, and the performance value of the model on the validation set is evaluated by calculating the determination coefficient and the root mean square error. When it is determined that the performance value meets the preset conditions, the optimal random forest model is used as the crop yield prediction model.

[0075] In this regard, it should be noted that the sample data in the sample data set include climate characteristics, biomass and yield per unit area. With yield per unit area as the dependent variable and extreme climate characteristics and biomass of each meteorological station in the region facing different growth periods as independent variables, the random forest machine learning model was input for model training. All samples were randomly divided into two groups, one of which was used as a training set to train the model, and the other was used as a test set to evaluate the performance of the model, running multiple parameter combinations to find the optimal parameters of the model.

[0076] Extract the independent variable columns (climate characteristics and biomass) and the dependent variable column (yield per unit area) from the read data, assign values ​​to X and y respectively, standardize the data, and use StandardScaler to eliminate the dimensional differences between different features to improve the training effect of the model.

[0077] Use the train_test_split function to split the sample data set into a training set and a test set, with the test set accounting for 20% and the training set accounting for 80%. This split method helps evaluate the performance of the model on unseen data.

[0078] A grid of parameters is defined for the Random Forest model, which will be tuned in a subsequent grid search. For Random Forest, the parameters tuned include the number of trees (n_estimators), the maximum depth of the tree (max_depth), the minimum number of samples required to split an internal node (min_samples_split), the minimum number of samples on a leaf node (min_samples_leaf), and the maximum number of features considered for splitting (max_features).

[0079] Use GridSearchCV to perform a grid search to find the optimal parameter combination through cross validation (cv=5) and the specified scoring metric ('r2'). Perform a grid search on the random forest model to train the model and find the optimal parameters.

[0080] Extract the best model estimator (best_estimator_), the best parameter combination (best_params_), and the best score (best_score_) from the GridSearchCV object.

[0081] The validation set data is used to predict the optimal random forest model. The performance of the model on the validation set is evaluated by calculating the coefficient of determination (R²) and the root mean square error (RMSE). These indicators can reflect the accuracy and error size of the model of the impact of extreme meteorological disasters on winter wheat. When the performance value meets the preset conditions, the optimal random forest model is used as the crop yield prediction model.

[0082] In the present invention, further, obtaining a sample data set includes:

[0083] Collecting meteorological data, soil data, field management data of each site within the target site range during the second time period and variety parameters of crops planted in the field within the target site range;

[0084] Simulate crop growth based on meteorological data, soil data, field management data and variety parameters to obtain information on biomass, potential yield per unit area and growth stage division;

[0085] Based on the division information of growth stages and the meteorological data of each station within the target station, the climate characteristics within the target station are obtained.

[0086] In this regard, it should be noted that in the present invention, in order to distinguish the meteorological data used in the prediction, the meteorological data, soil data, field management data of each station within the target station range in the second time period and the variety parameters of the crops planted in the field within the target station range can be determined from the sample data. Then, based on the meteorological data, soil data, field management data and variety parameters, the growth of crops is simulated to obtain the biomass, potential yield and growth stage division information. Based on the growth stage division information and the meteorological data of each station within the target station range, the climate characteristics within the target station range are obtained.

[0087] It should be noted that the meteorological data of the site will be divided into multiple indicator data, which are regarded as climate indicators. Different indicators and indicator combinations can characterize meteorological events. Therefore, for extreme meteorological events, it is necessary to construct climate characteristics from climate indicators.

[0088] The mean value and outlier range of climate indicators of each type of station after partitioning by the K-means clustering algorithm were calculated to obtain the typical climate characteristics and extreme climate characteristics of stations of different categories. The extreme climate characteristics of each station year by year, which characterize different growth stages and include different types of agricultural extreme meteorological disaster evaluation indicators such as cold, heat, drought and waterlogging, were statistically calculated, such as the maximum precipitation in 5 days, the maximum temperature in 5 days, the ratio of heavy rain days, the ratio of high temperature anomaly days, the ratio of low temperature anomaly days, the maximum temperature drop, negative accumulated temperature, SPEI, high temperature and low humidity dry hot wind, the ratio of high temperature days in each growth stage, and the ratio of low temperature days in each growth stage.

[0089] Table 1 shows the characteristics of extreme climate

[0090]

[0091] In a further method of the above method, the steps of obtaining the variety parameters of crops are mainly described as follows:

[0092] Collecting meteorological data of each station within the target station range in a third time period, and soil data and field management data of the fields within the target station range;

[0093] Set different variety parameters, input the meteorological data, soil data, and field management data of each site into the crop growth simulation model, and the crop growth simulation model outputs the yield simulation value;

[0094] Based on the simulated yield value and the measured yield value, the simulation evaluation index is determined, and the variety parameters of the crop varieties are determined based on the required simulation evaluation index; the variety parameters are the growth parameters of the planted crop varieties.

[0095] It should be noted that in the present invention, in order to distinguish the meteorological data from that used in prediction, the meteorological data, soil data, and field management data of each site within the target site range in the third time period can be collected when determining the data used for quality parameters.

[0096] After inputting meteorological data, soil data, and field management data into the APSIM-Wheat model, the “trial and error method” was used to debug the variety parameters of the APSIM-Wheat crop model, that is, different variety parameters and their combinations were set, and the four model simulation accuracy evaluation indicators, namely the linear regression determination coefficient (R^2) between the measured values ​​and the simulated values, the root mean square error (RMSE), the normalized root mean square error (NRMSE), and the graphical distribution 1:1 line, were used to obtain variety parameters that meet the actual situation. For example, the number of days from sowing to emergence, the number of days from sowing to flowering, the number of days from sowing to maturity, and the yield per unit area were used as X and Y variables, respectively, to make a scatter plot of the site, and the values ​​of each accuracy evaluation index at each site were counted to achieve an accurate simulation of the crop model.

[0097] In a further method of the above method, the process of determining the influence of the meteorological event corresponding to the climate feature to be analyzed on the actual unit yield of the crop is mainly explained as follows:

[0098] Identify climate features to be analyzed that have a risk factor higher than a preset factor;

[0099] Obtaining meteorological information of each station within the target station range in the fourth time period and the actual yield of crops planted in the field within the target station range;

[0100] Collect sliding windows with different preset window sizes and step lengths to analyze the positive and negative correlations between different climate characteristics to be analyzed and actual yields in each growth stage;

[0101] Based on the positive and negative correlations between different climate characteristics to be analyzed and the actual unit yield, the impact of the meteorological events corresponding to the climate characteristics to be analyzed on the actual unit yield of crops is determined.

[0102] In this regard, it should be noted that in the present invention, in order to distinguish the meteorological data from that used in prediction, the data used to determine the impact of meteorological events on the actual yield of crops can be collected from the meteorological data, soil data, and field management data of each site within the target site range within the fourth time period.

[0103] Sliding windows with preset different window sizes and step lengths are collected to analyze the positive and negative correlations between different climate characteristics to be analyzed and the actual yield per unit area in each growth stage. Based on the positive and negative correlations between different climate characteristics to be analyzed and the actual yield per unit area, the impact of meteorological events corresponding to the climate characteristics to be analyzed on the actual yield per unit area of ​​crops is determined.

[0104] In order to more comprehensively understand the dynamic relationship between extreme climate in different regions and winter wheat yield, three sliding windows with different window sizes and step lengths were specially designed to carefully analyze the positive and negative correlations between different indicators and yield within a specific range in each growth stage. This step not only helps to reveal the potential impact mechanism of extreme climate events on crop yields, but also provides a scientific basis for subsequent response strategies.

[0105] Furthermore, the actual yield per unit area of ​​the crops planted in the fields within the target site is obtained, the predicted yield per unit area is compared with the actual yield per unit area, and the comparison result is determined and displayed.

[0106] The following is a description of the processing device for the impact of meteorological events on crop yields provided by the present invention. The processing device for the impact of meteorological events on crop yields described below and the processing method for the impact of meteorological events on crop yields described above can be referenced to each other.

[0107] Figure 2 A schematic diagram of a process flow diagram of a device for processing the impact of meteorological events on crop yield provided by the present invention is shown in FIG. Figure 2 The device comprises an acquisition module 21, a first determination module 22, a second determination module 23, a prediction module 24 and a processing module 25, wherein:

[0108] The acquisition module 21 is used to acquire the meteorological data of each station within the target station range in the first time period, as well as the variety parameters, soil data and field management data of the crops planted in the fields within the target station range; the target station range is obtained based on the preset climate conditions;

[0109] A first determination module 22 is used to determine a climate characteristic vector within the range of a target site according to meteorological data, and to perform a preset number of random permutations on each climate characteristic in the climate characteristic vector to determine a plurality of climate characteristic vectors;

[0110] The second determination module 23 is used to determine the biomass of crops planted in the field according to variety parameters, soil data and field management data; biomass is an indicator of crop productivity;

[0111] A prediction module 24, used for inputting the plurality of climate characteristic vectors and the biomass into a crop yield prediction model, and the crop yield prediction model outputs a predicted yield;

[0112] The processing module 25 is used to determine the prediction error based on the predicted unit yield and the corresponding actual unit yield, and determine the risk coefficient of each climate feature based on the prediction error.

[0113] Among them, the crop yield prediction model is a model for predicting crop yields, which is obtained through machine learning training by taking the climate characteristic vector and biomass determined according to the sample data, as well as the actual yield corresponding to the sample data as input.

[0114] Since the principle of the device of the embodiment of the present invention is the same as that of the method of the above embodiment, a more detailed explanation is omitted here.

[0115] It should be noted that, in the embodiments of the present invention, relevant functional modules may be implemented by a hardware processor.

[0116] The processing device for the influence of meteorological events on crop yield provided by the present invention determines the climate characteristic vector according to meteorological data, determines the biomass of crops planted in the field according to variety parameters, soil data and field management data, collects and establishes a model of the relationship between extreme meteorological events and crop productivity to analyze the climate characteristic vector and biomass, outputs a predicted yield, and then determines the prediction error based on the predicted yield and the corresponding actual yield, and then determines the risk coefficient of each climate characteristic based on the prediction error, so as to predict the influence of the occurrence of extreme meteorological events on crop yield, provide an important reference for the long-term planning of agricultural production, and achieve the dual goals of risk warning and response.

[0117] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3As shown, the electronic device may include: a processor 31 (processor), a communication interface 32 (Communications Interface), a memory 33 (memory) and a communication bus 34, wherein the processor 31, the communication interface 32, and the memory 33 communicate with each other through the communication bus 34. The processor 31 can call the logic instructions in the memory 33 to execute a method for processing the impact of meteorological events on crop yields, the method comprising: obtaining meteorological data of each station within the target station range within a first time period, and variety parameters, soil data and field management data of crops planted in the field within the target station range; the target station range is divided based on preset climatic conditions; determining a climate characteristic vector within the target station range according to the meteorological data, and performing a preset number of random permutations on each climate characteristic in the climate characteristic vector to determine a plurality of climate characteristic vectors; determining the biomass of crops planted in the field according to the variety parameters, soil data and field management data; biomass is an indicator of crop productivity; inputting a plurality of climate characteristic vectors and biomass into a crop yield prediction model respectively, and the crop yield prediction model outputs a predicted yield; wherein the crop yield prediction model is a model for predicting crop yields obtained by machine learning training using the climate characteristic vector and biomass determined according to sample data and the actual yield corresponding to the sample data as inputs; determining a prediction error based on the predicted yield and the corresponding actual yield, and determining a risk coefficient for each climate characteristic based on the prediction error.

[0118] In addition, the logic instructions in the above-mentioned memory 33 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0119] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the processing method for the impact of meteorological events on crop yields provided by the above methods. The method includes: obtaining meteorological data of each station within a target site range in a first time period, as well as variety parameters, soil data and field management data of crops planted in a field within the target site range; the target site range is divided based on preset climate conditions; determining a climate feature vector within the target site range based on the meteorological data, and performing a climate feature analysis on each climate feature in the climate feature vector; The method comprises the following steps: performing a preset number of random permutations to determine a plurality of climate characteristic vectors; determining the biomass of crops planted in the field according to variety parameters, soil data and field management data; the biomass is an indicator of crop productivity; inputting the plurality of climate characteristic vectors and the biomass into a crop yield prediction model respectively, and the crop yield prediction model outputs a predicted yield; wherein the crop yield prediction model is a model for predicting crop yields obtained by machine learning training using the climate characteristic vector and biomass determined according to sample data and the actual yield corresponding to the sample data as inputs; determining a prediction error based on the predicted yield and the corresponding actual yield, and determining a risk coefficient for each climate characteristic based on the prediction error.

[0120] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the processing method for the impact of meteorological events on crop yields provided by the above methods, the method comprising: obtaining meteorological data of each station within a target site range in a first time period, as well as variety parameters, soil data and field management data of crops planted in a field within the target site range; the target site range is divided based on preset climatic conditions; determining a climate feature vector within the target site range based on the meteorological data, and performing a preset number of random permutations on each climate feature in the climate feature vector to determine multiple Climate characteristic vector; determine the biomass of crops planted in the field based on variety parameters, soil data and field management data; biomass is an indicator of crop productivity; input multiple climate characteristic vectors and biomass into the crop yield prediction model respectively, and the crop yield prediction model outputs the predicted yield; wherein, the crop yield prediction model is a model for predicting crop yields, which is obtained through machine learning training by taking the climate characteristic vector and biomass determined according to sample data and the actual yield corresponding to the sample data as input; determine the prediction error based on the predicted yield and the corresponding actual yield, and determine the risk coefficient of each climate characteristic based on the prediction error.

[0121] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0122] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing the impact of meteorological events on crop yield, characterized in that: include: Acquire meteorological data of each station within the target station range in the first time period, as well as variety parameters, soil data and field management data of crops planted in the field within the target station range; The target site range is obtained based on the preset climate conditions; Determine a climate characteristic vector within the range of the target site according to the meteorological data, and randomly permute each climate characteristic in the climate characteristic vector a preset number of times to determine a plurality of climate characteristic vectors; Determining the biomass of crops planted in the field according to the variety parameters, soil data and field management data; the biomass represents an indicator of crop productivity; Inputting a plurality of climate characteristic vectors and the biomass into a crop yield prediction model respectively, and the crop yield prediction model outputs a predicted yield; Based on the predicted unit yield and the corresponding actual unit yield, determining a prediction error, and determining a risk factor for each climate feature based on the prediction error; The crop yield prediction model is a model for predicting crop yields, which is obtained through machine learning training by taking the climate characteristic vector and biomass determined according to the sample data, and the actual yield corresponding to the sample data as input.

2. The method for processing the impact of meteorological events on crops according to claim 1, characterized in that: The method further comprises: a step of obtaining the crop yield prediction model, comprising: Obtain a sample data set, and divide the sample data set into a training set and a test set; Configure the parameter grid for the initial random forest model. The parameters to be tuned include the number of trees, the maximum depth of the tree, the minimum number of samples required to split an internal node, the minimum number of samples on a leaf node, and the maximum number of features used for splitting. Perform parameter grid search based on sample data of the training set, determine the optimal parameter combination through cross-validation and specified scoring indicators, and obtain the optimal random forest model; The optimal random forest model is verified based on the sample data of the validation set, and the performance value of the model on the validation set is evaluated by calculating the determination coefficient and the root mean square error. When it is determined that the performance value meets the preset conditions, the optimal random forest model is used as the crop yield prediction model.

3. The method for processing the impact of meteorological events on crop yield according to claim 2, characterized in that: The obtaining of the sample data set comprises: Collecting meteorological data, soil data, field management data of each site within the target site range during the second time period and variety parameters of crops planted in the field within the target site range; Simulate crop growth based on meteorological data, soil data, field management data and variety parameters to obtain information on biomass, potential yield per unit area and growth stage division; Based on the division information of growth stages and the meteorological data of each station within the target station, the climate characteristics within the target station are obtained.

4. The method for processing the impact of meteorological events on crop yield according to claim 3, characterized in that: The method further comprises a step of acquiring variety parameters of crops, including: Collecting meteorological data of each site within the target site range in a third time period, and soil data and field management data of the fields within the target site range; Set different variety parameters, input the meteorological data, soil data, and field management data of each site into the crop growth simulation model, and the crop growth simulation model outputs the yield simulation value; Based on the simulated yield value and the measured yield value, a simulation evaluation index is determined, and based on the simulation evaluation index, variety parameters of the crop variety are determined; the variety parameters are growth parameters of the planted crop variety.

5. The method for processing the impact of meteorological events on crop yield according to claim 3, characterized in that: The method further comprises: Identify climate features to be analyzed that have a risk factor higher than a preset factor; Obtaining meteorological information of each station within the target station range in the fourth time period and the actual yield of crops planted in the field within the target station range; Collect sliding windows with different preset window sizes and step lengths to analyze the positive and negative correlations between different climate characteristics to be analyzed and actual yields in each growth stage; Based on the positive and negative correlations between different climate characteristics to be analyzed and the actual unit yield, the impact of the meteorological events corresponding to the climate characteristics to be analyzed on the actual unit yield of crops is determined.

6. The method for processing the impact of meteorological events on crop yield according to claim 1, characterized in that: The method further comprises: The actual unit yield of the crops planted in the field within the scope of the first station is obtained, the predicted unit yield is compared with the actual unit yield, and the comparison result is determined and displayed.

7. A device for processing the impact of meteorological events on crop yield, characterized in that: include: An acquisition module, used to acquire meteorological data of each station within the target station range in the first time period, as well as variety parameters, soil data and field management data of crops planted in the field within the target station range; The target site range is obtained based on the preset climate conditions; A first determination module is used to determine a climate characteristic vector within the range of the target site according to the meteorological data, and to perform a preset number of random permutations on each climate characteristic in the climate characteristic vector to determine a plurality of climate characteristic vectors; A second determination module is used to determine the biomass of crops planted in the field according to the variety parameters, soil data and field management data; the biomass represents an indicator of crop productivity; A prediction module, used for inputting a plurality of climate characteristic vectors and the biomass into a crop yield prediction model respectively, and the crop yield prediction model outputs a predicted yield; A processing module, configured to determine a prediction error based on the predicted unit yield and the corresponding actual unit yield, and determine a risk coefficient of each climate feature based on the prediction error; The crop yield prediction model is a model for predicting crop yields, which is obtained through machine learning training by taking the climate characteristic vector and biomass determined according to the sample data, and the actual yield corresponding to the sample data as input.

8. The device for processing the impact of meteorological events on crop yield according to claim 7, characterized in that: The device also includes a prediction module, which is used to: Identify climate features to be analyzed that have a risk factor higher than a preset factor; Obtaining meteorological information of each station within the target station range in the fourth time period and the actual yield of crops planted in the field within the target station range; Collect sliding windows with different preset window sizes and step lengths to analyze the positive and negative correlations between different climate characteristics to be analyzed and actual yields in each growth stage; Based on the positive and negative correlations between different climate characteristics to be analyzed and the actual unit yield, the impact of the meteorological events corresponding to the climate characteristics to be analyzed on the actual unit yield of crops is determined.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, it implements the method for processing the impact of meteorological events on crop yield as described in any one of claims 1-6.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the method for processing the impact of meteorological events on crop yield as described in any one of claims 1 to 6.

Images