Artificial influence weather operation effect evaluation method and device based on deep learning, equipment and storage medium

Abstract

The invention provides an artificial influence weather operation effect evaluation method and device based on deep learning, equipment and a storage medium, relates to the field of artificial influence weather, and aims to conveniently and accurately analyze an artificial influence weather operation effect. The method comprises the following steps: acquiring historical data of a target area, identifying rainfall events, and dividing the rainfall events into artificial influence rainfall events and non-artificial influence rainfall events according to artificial influence weather operation records in the historical data; dividing the corresponding target data before the precipitation event without artificial influence into a training set, a test set and a verification set, training a deep learning model to obtain a trained precipitation prediction model, and calculating to obtain the target precipitation of the precipitation event without artificial influence corresponding to the precipitation event without artificial influence. And comparing the rainfall amount with the actually measured rainfall amount of the artificial influence rainfall event, and analyzing the artificial influence weather operation effect.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of artificial weather modification, and specifically, to a method, device, equipment and storage medium for evaluating the effect of artificial weather modification operations based on deep learning. Background Art

[0002] Weather modification is an important scientific and technological activity, which aims to transform weather phenomena in a predetermined direction through artificial means, so as to reduce or avoid meteorological disasters and make rational use of climate resources. Among them, effect evaluation is an important part of weather modification. Artificial rainfall is a widely carried out weather modification operation. In order to accurately describe the impact of artificial rainfall, in addition to qualitatively judging whether cloud seeding has a positive or negative effect, it is also necessary to answer from a quantitative perspective.

[0003] Traditional methods for evaluating the effectiveness of artificial weather modification operations include statistical comparison, controlled experiment, numerical simulation, physical process analysis and other methods, but these methods all have shortcomings. For example, the statistical comparison method is greatly affected by natural climate changes; the controlled experiment method is difficult to implement; the numerical simulation method has a large amount of calculation; and the physical process analysis method requires complex observation equipment and data, and the analysis process is complicated. Summary of the invention

[0004] The embodiments of the present application provide a method, device, equipment and storage medium for evaluating the effects of artificial weather modification operations based on deep learning, aiming to solve the problems that traditional methods for evaluating the effects of artificial weather modification operations are greatly affected by natural climate changes, are difficult to implement, have large amounts of calculations and complex analysis processes.

[0005] The first aspect of the embodiment of the present application provides a method for evaluating the effect of artificial weather modification operations based on deep learning, comprising:

[0006] Obtain historical data of the target area; the historical data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation and artificial weather modification operation records;

[0007] Preprocessing the historical data to identify precipitation events;

[0008] According to the artificial weather modification operation record, the precipitation event is divided into artificially influenced precipitation events and non-artificially influenced precipitation events;

[0009] The target data corresponding to the precipitation event before the event was not affected by human intervention is divided into a training set, a test set and a validation set, and a deep learning model is trained to obtain a trained precipitation prediction model; the target data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity and precipitation; the output data of the trained precipitation prediction model is the target precipitation of the precipitation event not affected by human intervention;

[0010] The data to be evaluated corresponding to before the artificially influenced precipitation event is input into the precipitation prediction model, the target precipitation of the precipitation event not affected by artificial influence corresponding to the artificially influenced precipitation event is calculated, and compared with the measured precipitation of the artificially influenced precipitation event, and the effect of the artificial weather influence operation of the artificially influenced precipitation event is analyzed.

[0011] In an optional implementation manner, the preprocessing of the historical data to identify precipitation events includes:

[0012] The period when the precipitation recorded in the target area exceeds 1 mm is regarded as a precipitation period;

[0013] Taking one or more consecutive precipitation periods as consecutive precipitation periods;

[0014] The time interval between a continuous precipitation period and the continuous precipitation period before or after the continuous precipitation period is greater than 6 hours, which is regarded as a precipitation event;

[0015] The time intervals between multiple continuous precipitation periods are all less than 6 hours, and the time intervals between the continuous precipitation periods before or after the multiple continuous precipitation periods are greater than 6 hours, which are regarded as a precipitation event.

[0016] In an optional implementation, the dividing the precipitation events into artificially influenced precipitation events and non-artificially influenced precipitation events according to the artificial weather modification operation records includes:

[0017] The precipitation events within 6 hours after the artificial weather modification operation record shall be regarded as artificially affected precipitation events; the precipitation events other than the artificially affected precipitation events shall be regarded as non-artificially affected precipitation events.

[0018] In an optional implementation, the target data corresponding to the precipitation event before the event not affected by human influence is divided into a training set, a test set and a validation set, and a deep learning model is trained to obtain a trained precipitation prediction model, including:

[0019] Using the training set to train the deep learning model, and adjusting the parameters of the deep learning model by a back propagation algorithm;

[0020] Using the validation set to evaluate the performance of the deep learning model, and adjusting the hyperparameters of the deep learning model according to the performance on the validation set;

[0021] Testing the deep learning model using the test set, calculating a relative error between the precipitation amount of each uninfluenced precipitation event in the test set and the target precipitation amount of each uninfluenced precipitation event corresponding to the output of the deep learning model;

[0022] When the relative error corresponding to each precipitation event not affected by human influence in the test set does not exceed 10%, the trained precipitation prediction model is obtained.

[0023] In an optional implementation, the analyzing the weather modification effect of the target precipitation event includes:

[0024] The measured precipitation of the artificially influenced precipitation event is calculated by difference with the target precipitation of the non-artificially influenced precipitation event corresponding to the artificially influenced precipitation event to obtain the precipitation increase;

[0025] The effect of the artificial weather modification operation is judged based on the positivity or negativity of the precipitation increase.

[0026] In an optional implementation manner, judging the effect of the artificial weather modification operation according to the precipitation includes:

[0027] When the increase in precipitation is greater than zero, the effect of the artificial weather modification operation is a positive effect;

[0028] When the increase in precipitation is less than zero, the effect of the artificial weather modification operation is a negative effect;

[0029] When the increase in precipitation is equal to zero, the effect of the artificial weather modification operation is invalid.

[0030] In an optional embodiment, when a meteorological disaster occurs in the target area, artificial intervention is performed on the target area based on the artificial weather modification operation with the positive effect.

[0031] A second aspect of the embodiment of the present application provides a device for evaluating the effect of artificial weather modification operations based on deep learning, comprising:

[0032] An acquisition module is used to acquire historical data of the target area; the historical data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation and artificial weather modification operation records;

[0033] An identification module, used for preprocessing the historical data and identifying precipitation events;

[0034] A classification module, used for classifying the precipitation event into artificially influenced precipitation events and non-artificially influenced precipitation events according to the artificial weather modification operation record;

[0035] The precipitation prediction model module is used to divide the target data corresponding to the precipitation event not affected by human intervention into a training set, a test set and a validation set, train a deep learning model, and obtain a trained precipitation prediction model; the target data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity and precipitation; the output data of the trained precipitation prediction model is the target precipitation of the precipitation event not affected by human intervention;

[0036] The analysis module is used to input the data to be evaluated corresponding to the artificially influenced precipitation event before the artificially influenced precipitation event into the precipitation prediction model, calculate the target precipitation of the precipitation event not affected by artificial influence corresponding to the artificially influenced precipitation event, and compare it with the measured precipitation of the artificially influenced precipitation event to analyze the effect of the artificial weather influence operation of the artificially influenced precipitation event.

[0037] In an optional implementation, the identification module is specifically used to:

[0038] The period when the precipitation recorded in the target area exceeds 1 mm is regarded as a precipitation period;

[0039] Taking one or more consecutive precipitation periods as consecutive precipitation periods;

[0040] The time interval between a continuous precipitation period and the continuous precipitation period before or after the continuous precipitation period is greater than 6 hours, which is regarded as a precipitation event;

[0041] The time intervals between multiple continuous precipitation periods are all less than 6 hours, and the time intervals between the continuous precipitation periods before or after the multiple continuous precipitation periods are greater than 6 hours, which are regarded as a precipitation event.

[0042] In an optional implementation, the division module is specifically used to:

[0043] The precipitation events within 6 hours after the artificial weather modification operation record shall be regarded as artificially affected precipitation events; the precipitation events other than the artificially affected precipitation events shall be regarded as non-artificially affected precipitation events.

[0044] In an optional implementation, the precipitation prediction model module is specifically used for:

[0045] Using the training set to train the deep learning model, and adjusting the parameters of the deep learning model by a back propagation algorithm;

[0046] Using the validation set to evaluate the performance of the deep learning model, and adjusting the hyperparameters of the deep learning model according to the performance on the validation set;

[0047] Testing the deep learning model using the test set, calculating a relative error between the precipitation amount of each uninfluenced precipitation event in the test set and the target precipitation amount of each uninfluenced precipitation event corresponding to the output of the deep learning model;

[0048] When the relative error corresponding to each precipitation event not affected by human influence in the test set does not exceed 10%, the trained precipitation prediction model is obtained.

[0049] In an optional implementation, the analysis module is specifically used for:

[0050] The measured precipitation of the artificially influenced precipitation event is calculated by difference with the target precipitation of the non-artificially influenced precipitation event corresponding to the artificially influenced precipitation event to obtain the precipitation increase;

[0051] The effect of the artificial weather modification operation is judged based on the positivity or negativity of the precipitation increase.

[0052] In an optional implementation, the analysis module is specifically used for:

[0053] When the increase in precipitation is greater than zero, the effect of the artificial weather modification operation is a positive effect;

[0054] When the increase in precipitation is less than zero, the effect of the artificial weather modification operation is a negative effect;

[0055] When the increase in precipitation is equal to zero, the effect of the artificial weather modification operation is invalid.

[0056] In an optional implementation, the analysis module is specifically used for:

[0057] In the event of a meteorological disaster occurring in the target area, artificial intervention is performed on the target area based on the artificial weather modification operation with the positive effect.

[0058] According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method for evaluating the effect of artificial weather modification operations based on deep learning as described in the first aspect.

[0059] In a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for evaluating the effect of artificial weather modification operations based on deep learning described in the first aspect are implemented.

[0060] In the disclosed embodiment, precipitation events are identified by acquiring historical data of the target area, and precipitation events are divided into artificially affected precipitation events and non-artificially affected precipitation events according to the artificial weather modification operation records in the historical data; according to the target data corresponding to the non-artificially affected precipitation event, a deep learning model is trained to obtain a trained precipitation prediction model, which is used to calculate the target precipitation of the non-artificially affected precipitation event corresponding to the artificially affected precipitation event, and then compared with the measured precipitation to analyze the effect of the artificial weather modification operation. The operation is simple, and the required data are all conventional observation data, which improves the scientific nature of the evaluation of the effect of the artificial weather modification operation and helps to more accurately analyze the effect of the artificial weather modification operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0062] Figure 1 It is a flowchart of the steps of a method for evaluating the effect of artificial weather modification operations based on deep learning proposed in one embodiment of the present application;

[0063] Figure 2 It is a structural schematic diagram of a weather modification operation effect evaluation device based on deep learning proposed in one embodiment of the present application;

[0064] Figure 3 is a schematic diagram of an electronic device in the present application. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0066] In the drawings, the size of the constituent elements, the thickness of the layer or the area may be exaggerated for the sake of clarity. Therefore, any implementation of the present disclosure is not necessarily limited to the size shown in the drawings, and the shapes and sizes of the components in the drawings do not reflect the true proportions. In addition, the drawings schematically show ideal examples, and any implementation of the present disclosure is not limited to the shapes or values ​​shown in the drawings.

[0067] Weather modification operations are an important scientific and technological activity that aims to transform weather phenomena in a predetermined direction through artificial means, so as to reduce or avoid meteorological disasters and make rational use of climate resources. Effect evaluation is an important part of the scientific research and business work of weather modification. The main difficulty in effect evaluation stems from the huge natural variation of the cloud-precipitation process. The impact of artificial operations is often small and weak compared to natural changes. In order to accurately describe the impact of artificial rainfall, in addition to qualitatively judging whether cloud seeding has a positive or negative effect, it is also necessary to answer from a quantitative perspective. The existing effect evaluation methods include statistical comparison method, control experiment method, numerical simulation method, physical process analysis method and other methods.

[0068] The principle of the statistical comparison method is to analyze the differences by comparing the meteorological data before and after artificial weather modification. The disadvantage is that it is greatly affected by natural climate changes, and it is difficult to determine whether the changes are entirely caused by artificial measures; the principle of the control experiment method is to set up test areas and control areas, and evaluate the effects of artificial influence by comparing the meteorological data of the two. The disadvantage is that it requires large test areas and control areas, which is difficult to implement and it is difficult to eliminate the natural differences between different regions; the principle of the numerical simulation method is to use meteorological models to simulate meteorological changes with and without artificial influence. The disadvantage is that it has high requirements on the accuracy of the model and a large amount of calculation; the principle of the physical process analysis method is to determine the mechanism of action of artificial influence measures on clouds and precipitation through analysis of cloud physical processes. The disadvantage is that it requires complex observation equipment and data, and the analysis process is complex.

[0069] Reference Figure 1 , Figure 1 1 is a flowchart of the steps of a method for evaluating the effect of artificial weather modification operations based on deep learning proposed in one embodiment of the present application. Figure 1 As shown, the method includes the following steps S11 to S15:

[0070] Step S11: Acquire historical data of the target area; the historical data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation and artificial weather modification operation records.

[0071] Historical data such as historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation, and weather modification operation records of the target area can be obtained through the Meteorological Bureau, Satellite Remote Sensing Center, scientific research institutions, etc. For example, historical satellite cloud images come from meteorological satellites, which can be obtained by visiting the official website of the Meteorological Bureau or data center and using the data download service. This provides strong data support for subsequent analysis.

[0072] Step S12: pre-process the historical data to identify precipitation events.

[0073] Preprocess historical data such as historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation, and weather modification operation records to identify precipitation events. Preprocessing at least includes data cleaning. Data cleaning can remove or correct errors, outliers, and missing values ​​in the data to ensure the accuracy and completeness of the data. Accurate precipitation event identification results are of great significance for the study of weather modification operations.

[0074] Furthermore, the preprocessing of the historical data to identify precipitation events includes:

[0075] The period when the precipitation recorded in the target area exceeds 1 mm is regarded as a precipitation period;

[0076] Taking one or more consecutive precipitation periods as consecutive precipitation periods;

[0077] The time interval between a continuous precipitation period and the continuous precipitation period before or after the continuous precipitation period is greater than 6 hours, which is regarded as a precipitation event;

[0078] The time intervals between multiple continuous precipitation periods are all less than 6 hours, and the time intervals between the continuous precipitation periods before or after the multiple continuous precipitation periods are greater than 6 hours, which are regarded as a precipitation event.

[0079] The period when the precipitation recorded in the target area exceeds 1 mm is defined as the precipitation period. This step is the basis for identifying precipitation events, which ensures that only periods where precipitation actually occurs are considered. The precipitation threshold is set to 1 mm because such precipitation is usually considered to be effective precipitation in actual meteorological observations.

[0080] One or more consecutive precipitation periods are combined into one continuous precipitation period. Here, "continuous" means that the time interval between precipitation periods does not exceed a certain threshold (here, it is instantaneous, that is, without interruption).

[0081] If the interval between a continuous precipitation period and the previous or subsequent continuous precipitation period is greater than 6 hours, then the continuous precipitation period is considered as a precipitation event. This rule is used to distinguish different precipitation events and ensure that there is enough non-precipitation time interval between them.

[0082] If the time interval between multiple consecutive precipitation periods is less than 6 hours, and the time interval between these consecutive precipitation periods and the previous or next consecutive precipitation periods is greater than 6 hours, then these consecutive precipitation periods are combined into one precipitation event. This reflects the continuity and integrity of precipitation activities, even if there may be short breaks in between.

[0083] Step S13: according to the weather modification operation record, the precipitation events are divided into precipitation events affected by artificial weather and precipitation events not affected by artificial weather.

[0084] According to the implementation time in the weather modification operation records, precipitation events are divided into artificially affected precipitation events and non-artificially affected precipitation events. Different types of precipitation events are distinguished, and then the non-artificially affected precipitation events are identified, providing a clear data basis for subsequent model training and effect evaluation.

[0085] Furthermore, the dividing of the precipitation events into artificially influenced precipitation events and non-artificially influenced precipitation events according to the artificial weather modification operation records includes:

[0086] The precipitation events within 6 hours after the artificial weather modification operation record shall be regarded as artificially affected precipitation events; the precipitation events other than the artificially affected precipitation events shall be regarded as non-artificially affected precipitation events.

[0087] Precipitation events that occur within 6 hours after the implementation of weather modification operations are specially marked as artificially affected precipitation events. This is because weather modification operations (such as artificial rainfall) may have an impact on subsequent precipitation activities, and the 6-hour time window is used as a reasonable period to evaluate such impacts, which can also be adjusted according to specific meteorological conditions and operation types. The remaining precipitation events, that is, precipitation events that do not meet the above conditions, are recorded as non-artificially affected precipitation events.

[0088] Step S14: Divide the target data corresponding to the precipitation event before the event was not affected by human intervention into a training set, a test set and a validation set, train the deep learning model, and obtain a trained precipitation prediction model; the target data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity and precipitation; the output data of the trained precipitation prediction model is the target precipitation of the precipitation event not affected by human intervention.

[0089] The target data such as historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity and precipitation corresponding to the precipitation event before the artificial influence are divided into training set, test set and verification set. The division ratio can be adjusted according to the actual situation. In this embodiment, the ratio is 70% training set, 15% verification set, and 15% test set. According to the characteristics of the data and the needs of the prediction task, a suitable deep learning model can be selected, and a hybrid model (such as a combination of a convolutional neural network and a recurrent neural network) can be constructed.

[0090] The deep learning model is trained using satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation and other data at several moments before the uninfluenced precipitation event as independent variables (model input). The "several moments" here refer to the time points or time periods of various meteorological data related to the target precipitation event collected within a period of time before the precipitation event occurs. The selection of these time points or time periods depends on the prediction requirements of the model, the availability of data and the characteristics of the meteorological system. The output data of the trained precipitation prediction model is the target precipitation of the uninfluenced precipitation event. A precipitation prediction model that can accurately predict the uninfluenced precipitation is constructed, providing a reliable prediction tool for the subsequent evaluation of the effectiveness of artificial weather modification operations.

[0091] Furthermore, the target data corresponding to the precipitation event not affected by human intervention is divided into a training set, a test set and a validation set, and a deep learning model is trained to obtain a trained precipitation prediction model, including:

[0092] Using the training set to train the deep learning model, and adjusting the parameters of the deep learning model by a back propagation algorithm;

[0093] Using the validation set to evaluate the performance of the deep learning model, and adjusting the hyperparameters of the deep learning model according to the performance on the validation set;

[0094] Testing the deep learning model using the test set, calculating a relative error between the precipitation amount of each uninfluenced precipitation event in the test set and the target precipitation amount of each uninfluenced precipitation event corresponding to the output of the deep learning model;

[0095] When the relative error corresponding to each precipitation event not affected by human influence in the test set does not exceed 10%, the trained precipitation prediction model is obtained.

[0096] The data in the divided training set, test set, and validation set are preprocessed, including normalization, cropping, scaling, etc. of image data, and normalization or standardization of numerical data. The preprocessed training set data is input into the deep learning model, and the back propagation algorithm is used to adjust the model parameters (such as weights and biases) to minimize the prediction error.

[0097] During the training process, the validation set data is used regularly to evaluate the performance of the model. This can be done by calculating performance metrics such as loss function value, accuracy, recall, F1 score, etc. Based on the performance metrics on the validation set, the hyperparameters of the model are adjusted. Hyperparameters are those parameters that are set before the training starts and do not change during the training process, such as learning rate, batch size, number of network layers, activation function, etc.

[0098] After the model training and optimization are completed, the model is tested using the test set data. The relative error between the precipitation of each unaffected precipitation event in the test set and the target precipitation of each unaffected precipitation event corresponding to the output of the deep learning model is calculated. If the relative error of all unaffected precipitation events in the test set does not exceed 10%, the model is considered to have been trained and has achieved the expected performance standard. This indicates that the trained precipitation prediction model has high accuracy and reliability. The relative error is calculated according to the following formula (1):

[0099]

[0100] Among them, E 相对 Represents relative error; P 计算 represents the target precipitation of precipitation events without artificial influence output by the deep learning model; P 实测 Represents the precipitation amount of precipitation events in the test set that are not affected by human intervention.

[0101] Step S15: input the data to be evaluated corresponding to the artificially influenced precipitation event before the artificially influenced precipitation event into the precipitation prediction model, calculate the target precipitation of the precipitation event not affected by artificial influence corresponding to the artificially influenced precipitation event, compare it with the measured precipitation of the artificially influenced precipitation event, and analyze the effect of the artificial weather influence operation of the artificially influenced precipitation event.

[0102] The corresponding data to be evaluated before the artificially affected precipitation event to be analyzed is input into the precipitation prediction model, and the target precipitation of the non-artificially affected precipitation event corresponding to the artificially affected precipitation event is calculated. The target precipitation of the non-artificially affected precipitation event here is predicted by the precipitation prediction model. The evaluation data refers to the satellite cloud map, radar echo map, temperature, air pressure, wind speed, wind direction, humidity, and precipitation at several moments of the event. These data can be obtained through the Meteorological Bureau, Satellite Remote Sensing Center, scientific research institutions, etc.

[0103] The measured precipitation of the artificially affected precipitation event to be analyzed is the measured precipitation after the artificial weather modification operation, which can be obtained through scientific research institutions related to artificial weather modification operations. The measured precipitation of the artificially affected precipitation event is compared with the target precipitation of the corresponding non-artificially affected precipitation event to analyze the effect of the artificial weather modification operation of the artificially affected precipitation event. The target data obtained by model prediction is compared with the measured precipitation to quantify the increase or decrease in precipitation caused by the artificial weather modification operation, so as to judge whether the operation is successful and the degree of success, which improves the scientific nature of the evaluation of the effect of artificial weather modification operations and helps to more accurately analyze the effect of artificial weather modification operations.

[0104] Furthermore, the analyzing the weather modification effect of the target precipitation event includes:

[0105] The measured precipitation of the artificially influenced precipitation event is calculated by difference with the target precipitation of the non-artificially influenced precipitation event corresponding to the artificially influenced precipitation event to obtain the precipitation increase;

[0106] The effect of the artificial weather modification operation is judged based on the positivity or negativity of the precipitation increase.

[0107] The difference between the measured precipitation of the artificially affected precipitation event and the target precipitation of the non-artificially affected precipitation event corresponding to the artificially affected precipitation event is calculated to obtain the precipitation increase; then the effect of the artificial weather modification operation is judged based on the positive or negative sign of the precipitation increase. The precipitation increase obtained by the difference calculation provides a direct basis for evaluating the effect of artificial weather operations. The precipitation increase is calculated according to the following formula (2):

[0108] P Δ =P1-P0 (2)

[0109] Among them, P Δ represents the increase in precipitation; P1 represents the measured precipitation of the artificially influenced precipitation event; P0 represents the target precipitation of the non-artificially influenced precipitation event corresponding to the artificially influenced precipitation event.

[0110] Furthermore, judging the effect of the artificial weather modification operation according to the precipitation includes:

[0111] When the increase in precipitation is greater than zero, the effect of the artificial weather modification operation is a positive effect;

[0112] When the increase in precipitation is less than zero, the effect of the artificial weather modification operation is a negative effect;

[0113] When the increase in precipitation is equal to zero, the effect of the artificial weather modification operation is invalid.

[0114] In this embodiment, the artificial weather modification operation can conduct a specific effect evaluation on the artificial rainmaking operation. If the increase in precipitation is greater than zero, that is, the measured precipitation is greater than the target precipitation, it means that the artificial weather modification operation has produced a positive effect and increased precipitation; if the increase in precipitation is negative, that is, the measured precipitation is less than the target precipitation, it means that the artificial weather modification operation may have a counter-effect or fail to achieve the expected goal; if the increase in precipitation is zero, that is, the measured precipitation is equal to the target precipitation, it means that the artificial weather modification operation has no obvious effect. According to the results of the effect judgment, the future practice of artificial weather modification operations can be guided. For example, if a certain operation produces a significant positive effect, the experience can be summarized and promoted under similar conditions; if the effect is not significant or produces a counter-effect, it is necessary to analyze the cause and adjust the operation strategy.

[0115] Furthermore, in the event of a meteorological disaster occurring in the target area, artificial intervention is performed on the target area based on the artificial weather modification operation with the positive effect.

[0116] When a meteorological disaster, such as drought, occurs in the target area, we can choose those operation plans that have been proven to have positive effects and are suitable for current meteorological conditions based on the previous evaluation results of artificial weather modification operations to conduct artificial intervention in the target area. This may include the use of different types of catalysts (such as silver iodide, dry ice, etc.), different operation methods (such as aircraft seeding, rocket launches, etc.), and different operation times and frequencies to mitigate the impact of the disaster.

[0117] Based on the same inventive concept, the present application embodiment discloses a weather modification operation effect evaluation device based on deep learning, such as Figure 2 As shown, the device comprises:

[0118] An acquisition module is used to acquire historical data of the target area; the historical data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation and artificial weather modification operation records;

[0119] An identification module, used for preprocessing the historical data and identifying precipitation events;

[0120] A classification module, used for classifying the precipitation event into artificially influenced precipitation events and non-artificially influenced precipitation events according to the artificial weather modification operation record;

[0121] The precipitation prediction model module is used to divide the target data corresponding to the precipitation event not affected by human intervention into a training set, a test set and a validation set, train a deep learning model, and obtain a trained precipitation prediction model; the target data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity and precipitation; the output data of the trained precipitation prediction model is the target precipitation of the precipitation event not affected by human intervention;

[0122] The analysis module is used to input the data to be evaluated corresponding to the artificially influenced precipitation event before the artificially influenced precipitation event into the precipitation prediction model, calculate the target precipitation of the precipitation event not affected by artificial influence corresponding to the artificially influenced precipitation event, and compare it with the measured precipitation of the artificially influenced precipitation event to analyze the effect of the artificial weather influence operation of the artificially influenced precipitation event.

[0123] In an optional implementation, the identification module is specifically used to:

[0124] The period when the precipitation recorded in the target area exceeds 1 mm is regarded as a precipitation period;

[0125] Taking one or more consecutive precipitation periods as consecutive precipitation periods;

[0126] The time interval between a continuous precipitation period and the continuous precipitation period before or after the continuous precipitation period is greater than 6 hours, which is regarded as a precipitation event;

[0127] The time intervals between multiple continuous precipitation periods are all less than 6 hours, and the time intervals between the continuous precipitation periods before or after the multiple continuous precipitation periods are greater than 6 hours, which are regarded as a precipitation event.

[0128] In an optional implementation, the division module is specifically used to:

[0129] The precipitation events within 6 hours after the artificial weather modification operation record shall be regarded as artificially affected precipitation events; the precipitation events other than the artificially affected precipitation events shall be regarded as non-artificially affected precipitation events.

[0130] In an optional implementation, the precipitation prediction model module is specifically used for:

[0131] Using the training set to train the deep learning model, and adjusting the parameters of the deep learning model by a back propagation algorithm;

[0132] Using the validation set to evaluate the performance of the deep learning model, and adjusting the hyperparameters of the deep learning model according to the performance on the validation set;

[0133] Testing the deep learning model using the test set, calculating a relative error between the precipitation amount of each uninfluenced precipitation event in the test set and the target precipitation amount of each uninfluenced precipitation event corresponding to the output of the deep learning model;

[0134] When the relative error corresponding to each precipitation event not affected by human influence in the test set does not exceed 10%, the trained precipitation prediction model is obtained.

[0135] In an optional implementation, the analysis module is specifically used for:

[0136] The measured precipitation of the artificially influenced precipitation event is calculated by difference with the target precipitation of the non-artificially influenced precipitation event corresponding to the artificially influenced precipitation event to obtain the precipitation increase;

[0137] The effect of the artificial weather modification operation is judged based on the positivity or negativity of the precipitation increase.

[0138] In an optional implementation, the analysis module is specifically used for:

[0139] When the increase in precipitation is greater than zero, the effect of the artificial weather modification operation is a positive effect;

[0140] When the increase in precipitation is less than zero, the effect of the artificial weather modification operation is a negative effect;

[0141] When the increase in precipitation is equal to zero, the effect of the artificial weather modification operation is invalid.

[0142] In an optional implementation, the analysis module is specifically used for:

[0143] In the event of a meteorological disaster occurring in the target area, artificial intervention is performed on the target area based on the artificial weather modification operation with the positive effect.

[0144] The method for evaluating the effect of artificial weather modification operations based on deep learning provided by the disclosed embodiment identifies precipitation events by acquiring historical data of the target area, and divides precipitation events into artificially affected precipitation events and non-artificially affected precipitation events according to the artificial weather modification operation records in the historical data; trains a deep learning model according to the target data corresponding to the non-artificially affected precipitation event, obtains a trained precipitation prediction model, and calculates the target precipitation of the non-artificially affected precipitation event corresponding to the artificially affected precipitation event, and then compares it with the measured precipitation to analyze the effect of artificial weather modification operations. The operation is simple, and the required data are all conventional observation data, which improves the scientific nature of the evaluation of the effect of artificial weather modification operations and helps to more accurately analyze the effect of artificial weather modification operations.

[0145] It should be noted that, for the method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present disclosure are not limited by the order of the actions described, because according to the embodiments of the present disclosure, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present disclosure.

[0146] The present disclosure also provides an electronic device, referring to Figure 3 , Figure 3 Schematic diagram of an electronic device shown in an embodiment of the present disclosure. Figure 3 As shown, the electronic device 100 includes: a memory 110 and a processor 120. The memory 110 and the processor 120 are connected via a bus communication. The memory 110 stores a computer program, which can be executed on the processor 120 to implement the steps of the method for evaluating the effect of artificial weather modification operations based on deep learning disclosed in the embodiment of the present disclosure.

[0147] The embodiments of the present disclosure also provide a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by a processor of a computer device, the computer device is enabled to execute the steps in the method for evaluating the effect of artificial weather modification operations based on deep learning as disclosed in the embodiments of the present disclosure.

[0148] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0149] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, apparatuses, electronic devices, and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0150] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0151] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0152] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present application.

[0153] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0154] The above is a detailed introduction to the method, device, equipment and storage medium for evaluating the effect of artificial weather modification operations based on deep learning provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for evaluating the effect of artificial weather modification operations based on deep learning, characterized in that: include: Obtain historical data of the target area; the historical data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation and artificial weather modification operation records; Preprocessing the historical data to identify precipitation events; According to the artificial weather modification operation record, the precipitation event is divided into artificially influenced precipitation events and non-artificially influenced precipitation events; The target data corresponding to the precipitation event before the event was not affected by human intervention is divided into a training set, a test set and a validation set, and a deep learning model is trained to obtain a trained precipitation prediction model; the target data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity and precipitation; the output data of the trained precipitation prediction model is the target precipitation of the precipitation event not affected by human intervention; The data to be evaluated corresponding to before the artificially influenced precipitation event is input into the precipitation prediction model, the target precipitation of the precipitation event not affected by artificial influence corresponding to the artificially influenced precipitation event is calculated, and compared with the measured precipitation of the artificially influenced precipitation event, and the effect of the artificial weather influence operation of the artificially influenced precipitation event is analyzed.

2. The method for evaluating the effect of artificial weather modification operations based on deep learning according to claim 1, characterized in that: The preprocessing of the historical data to identify precipitation events includes: The period when the precipitation recorded in the target area exceeds 1 mm is regarded as a precipitation period; Taking one or more consecutive precipitation periods as consecutive precipitation periods; The time interval between a continuous precipitation period and the continuous precipitation period before or after the continuous precipitation period is greater than 6 hours, which is regarded as a precipitation event; The time intervals between multiple continuous precipitation periods are all less than 6 hours, and the time intervals between the continuous precipitation periods before or after the multiple continuous precipitation periods are greater than 6 hours, which are regarded as a precipitation event.

3. The method for evaluating the effect of artificial weather modification operations based on deep learning according to claim 2, characterized in that: The step of dividing the precipitation events into artificially influenced precipitation events and non-artificially influenced precipitation events according to the artificial weather influence operation records includes: The precipitation events within 6 hours after the artificial weather modification operation record shall be regarded as artificially affected precipitation events; the precipitation events other than the artificially affected precipitation events shall be regarded as non-artificially affected precipitation events.

4. The method for evaluating the effect of artificial weather modification operations based on deep learning according to claim 1, characterized in that: The target data corresponding to the precipitation event before the artificial influence is divided into a training set, a test set and a validation set, and a deep learning model is trained to obtain a trained precipitation prediction model, including: Using the training set to train the deep learning model, and adjusting the parameters of the deep learning model by a back propagation algorithm; Using the validation set to evaluate the performance of the deep learning model, and adjusting the hyperparameters of the deep learning model according to the performance on the validation set; Testing the deep learning model using the test set, calculating a relative error between the precipitation amount of each uninfluenced precipitation event in the test set and the target precipitation amount of each uninfluenced precipitation event corresponding to the output of the deep learning model; When the relative error corresponding to each precipitation event not affected by human influence in the test set does not exceed 10%, the trained precipitation prediction model is obtained.

5. The method for evaluating the effect of artificial weather modification operations based on deep learning according to claim 1, characterized in that: The analyzing the weather modification effect of the target precipitation event includes: The measured precipitation of the artificially influenced precipitation event is calculated by difference with the target precipitation of the non-artificially influenced precipitation event corresponding to the artificially influenced precipitation event to obtain the precipitation increase; The effect of the artificial weather modification operation is judged based on the positive or negative value of the precipitation increase.

6. The method for evaluating the effect of artificial weather modification operations based on deep learning according to claim 5, characterized in that: The judging of the effect of the artificial weather modification operation according to the precipitation includes: When the increase in precipitation is greater than zero, the effect of the artificial weather modification operation is a positive effect; When the increase in precipitation is less than zero, the effect of the artificial weather modification operation is a negative effect; When the increase in precipitation is equal to zero, the effect of the artificial weather modification operation is invalid.

7. The method for evaluating the effect of artificial weather modification operations based on deep learning according to claim 6, characterized in that: In the event of a meteorological disaster occurring in the target area, artificial intervention is performed on the target area based on the artificial weather modification operation with the positive effect.

8. A device for evaluating the effect of artificial weather modification operations based on deep learning, characterized in that: The device comprises: An acquisition module is used to acquire historical data of the target area; the historical data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity, precipitation and artificial weather modification operation records; An identification module, used for preprocessing the historical data and identifying precipitation events; A classification module, used for classifying the precipitation event into artificially influenced precipitation events and non-artificially influenced precipitation events according to the artificial weather modification operation record; The precipitation prediction model module is used to divide the target data corresponding to the precipitation event not affected by human intervention into a training set, a test set and a validation set, train a deep learning model, and obtain a trained precipitation prediction model; the target data includes: historical satellite cloud images, radar echo images, temperature, air pressure, wind speed, wind direction, humidity and precipitation; the output data of the trained precipitation prediction model is the target precipitation of the precipitation event not affected by human intervention; The analysis module is used to input the data to be evaluated corresponding to the artificially influenced precipitation event before the artificially influenced precipitation event into the precipitation prediction model, calculate the target precipitation of the precipitation event not affected by artificial influence corresponding to the artificially influenced precipitation event, and compare it with the measured precipitation of the artificially influenced precipitation event to analyze the effect of the artificial weather influence operation of the artificially influenced precipitation event.

9. The deep learning-based weather modification operation effect evaluation device according to claim 8, characterized in that: The identification module is specifically used for: The period when the precipitation recorded in the target area exceeds 1 mm is regarded as a precipitation period; Taking one or more consecutive precipitation periods as consecutive precipitation periods; The time interval between a continuous precipitation period and the continuous precipitation period before or after the continuous precipitation period is greater than 6 hours, which is regarded as a precipitation event; The time intervals between multiple continuous precipitation periods are all less than 6 hours, and the time intervals between the continuous precipitation periods before or after the multiple continuous precipitation periods are greater than 6 hours, which are regarded as a precipitation event.

10. The deep learning-based artificial weather modification operation effect evaluation device according to claim 8, characterized in that: The partitioning module is specifically used for: The precipitation events within 6 hours after the artificial weather modification operation record shall be regarded as artificially affected precipitation events; the precipitation events other than the artificially affected precipitation events shall be regarded as non-artificially affected precipitation events.

11. The deep learning-based weather modification operation effect evaluation device according to claim 8, characterized in that: The precipitation prediction model module is specifically used for: Using the training set to train the deep learning model, and adjusting the parameters of the deep learning model by a back propagation algorithm; Using the validation set to evaluate the performance of the deep learning model, and adjusting the hyperparameters of the deep learning model according to the performance on the validation set; Testing the deep learning model using the test set, calculating a relative error between the precipitation amount of each uninfluenced precipitation event in the test set and the target precipitation amount of each uninfluenced precipitation event corresponding to the output of the deep learning model; When the relative error corresponding to each precipitation event not affected by human influence in the test set does not exceed 10%, the trained precipitation prediction model is obtained.

12. The deep learning-based artificial weather modification operation effect evaluation device according to claim 8, characterized in that: The analysis module is specifically used for: The measured precipitation of the artificially influenced precipitation event is calculated by difference with the target precipitation of the non-artificially influenced precipitation event corresponding to the artificially influenced precipitation event to obtain the precipitation increase; The effect of the artificial weather modification operation is judged based on the positive or negative value of the precipitation increase.

13. The deep learning-based weather modification operation effect evaluation device according to claim 12, characterized in that: The analysis module is specifically used for: When the increase in precipitation is greater than zero, the effect of the artificial weather modification operation is a positive effect; When the increase in precipitation is less than zero, the effect of the artificial weather modification operation is a negative effect; When the increase in precipitation is equal to zero, the effect of the artificial weather modification operation is invalid.

14. The deep learning-based weather modification operation effect evaluation device according to claim 13, characterized in that: The analysis module is specifically used for: In the event of a meteorological disaster occurring in the target area, artificial intervention is performed on the target area based on the artificial weather modification operation with the positive effect.

15. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for evaluating the effect of artificial weather modification operations based on deep learning as described in any one of claims 1 to 7 is implemented.

16. A computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method for evaluating the effect of artificial weather modification operations based on deep learning as described in any one of claims 1 to 7.

Images