Short-term and imminent forecasting method, device, equipment and medium

Abstract

The invention relates to a short-term and imminent forecasting method and device, equipment and a medium. The short-term and imminent forecasting method comprises the following steps: acquiring a radar reflectivity result of satellite inversion; performing pyramid layering on the radar reflectivity result to obtain a corresponding layering result; establishing a Gaussian pyramid model according to the layering result, and calculating a radar echo value based on the Gaussian pyramid model; and forecasting short temporary rainfall based on the radar echo value. According to the embodiment of the invention, the forecasting accuracy of extreme rainfall weather can be improved.

Description

Technical Field

[0001] The present disclosure relates to the technical application field of severe convective extreme weather forecasting, and particularly to a short-term and imminent forecasting method, device, equipment and medium. Background Art

[0002] With global warming, the frequency of extreme rainfall increases. Severe rainstorm weather is likely to cause a rain curtain to form on overhead lines, insulator strings and towers. In severe cases, a rain column bridging will occur, significantly reducing the electrical insulation performance of the conductor-tower gap, resulting in the tripping of overhead lines and power outages.

[0003] For precipitation forecasting, the following methods are usually adopted: methods based on satellite-inverted radar echo and precipitation, and methods improved based on the Genuine Progress Indicator (GPI). For example, the precipitation inversion method earliest used on the Geostationary Operational Environmental Satellite (GOES), that is, fitting the relationship between the cloud-top brightness temperature and ground precipitation through a simple function. Although this method is simple and easy to implement, it has great limitations in areas where the latitude is higher than 40°, and the inversion accuracy is relatively low. The method improved based on the Genuine Progress Indicator (GPI) uses the data of 5 visible light / infrared bands of GOES, estimates the precipitation intensity by using an exponential function and a quadratic curve, and corrects the precipitation intensity by using a humidity correction factor and a cloud growth rate correction factor. Although the accuracy of this method is improved compared with GPI, the functional relationship used in this method is more complex, and the relationship between infrared brightness temperature and precipitation is still preset, so there is an upper limit in terms of accuracy. Summary of the Invention

[0004] To solve the above technical problems, the present disclosure provides a short-term and imminent forecasting method, device, equipment and medium.

[0005] In a first aspect, the present disclosure provides a short-term and imminent forecasting method, including:

[0006] Obtaining the radar reflectivity result inverted by the satellite;

[0007] Performing pyramid layering on the radar reflectivity result to obtain a corresponding layering result;

[0008] Establishing a Gaussian pyramid model according to the layering result, and calculating the radar echo value based on the Gaussian pyramid model;

[0009] Performing short-term and imminent precipitation forecasting based on the radar echo value.

[0010] In some examples, the performing pyramid layering on the radar reflectivity result to obtain a corresponding layering result includes:

[0011] Based on the optical flow method, pyramid layering is performed on each frame image of the radar reflectivity result to obtain the layering result.

[0012] In some examples, the performing pyramid layering on each frame image of the radar reflectivity result based on the optical flow method to obtain the layering result includes:

[0013] Taking the picture with the largest scale among multiple frame images of the radar reflectivity result as the top layer and the original picture among multiple frame images as the bottom layer to obtain the corresponding layering result.

[0014] In some examples, the establishing a Gaussian pyramid model according to the layering result and calculating the radar echo value based on the Gaussian pyramid model includes:

[0015] Obtaining a historical training data set containing the generation and disappearance characteristics of single entities under various weather conditions;

[0016] Establishing a Gaussian pyramid model according to the layering result and performing calculations on the historical training data set based on the Gaussian pyramid model to obtain the radar echo value.

[0017] In some examples, the performing calculations on the historical training data set based on the Gaussian pyramid model to obtain the radar echo value includes:

[0018] Inputting the radar echo at a preset time in the historical training data set into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information;

[0019] Obtaining the radar reflectivity residual information and the corresponding picture color information between the radar echo at the preset time and the target radar echo at an adjacent time, and performing training through a pre-trained deep learning model based on the pixel-by-pixel optical flow field information, the radar reflectivity residual information, the picture color information, and the forecasting factors to obtain the radar echo value.

[0020] In some examples, the forecasting factors include at least one dynamic feature such as rotation, convergence, and divergence in the optical flow field.

[0021] In some examples, the method further includes:

[0022] Constructing the relationship between the satellite observation brightness temperature and the radar echo value.

[0023] In some examples, the constructing the relationship between the satellite observation brightness temperature and the radar echo value includes:

[0024] Based on a preset network, performing interval classification on the radar echo value to obtain multiple categories;

[0025] Construct a satellite-inverted radar echo classification model based on multiple categories, and the satellite-inverted radar echo classification model is used to simulate the relationship between the satellite-observed brightness temperature and the radar echo value.

[0026] In a second aspect, the present disclosure provides a short-term and imminent forecasting device, including:

[0027] A result acquisition module, configured to acquire the radar reflectivity result inverted by the satellite;

[0028] An image layering module, configured to perform pyramid layering on the radar reflectivity result to obtain a corresponding layering result;

[0029] A data calculation module, configured to establish a Gaussian pyramid model based on the layering result and calculate the radar echo value based on the Gaussian pyramid model;

[0030] A precipitation forecasting module, configured to perform short-term and imminent precipitation forecasting based on the radar echo value.

[0031] In some examples, the image layering module includes:

[0032] A first processing unit, configured to perform pyramid layering on each frame image of the radar reflectivity result based on the optical flow method to obtain the layering result.

[0033] In some examples, the image layering module includes:

[0034] A second processing unit, configured to use the largest-scale picture among multiple frame images of the radar reflectivity result as the top layer and the original picture among multiple frame images as the bottom layer to obtain a corresponding layering result.

[0035] In some examples, the data calculation module includes:

[0036] A data acquisition unit, configured to acquire a historical training data set including the generation and disappearance characteristics of monomers under various weather conditions;

[0037] A model calculation unit, configured to establish a Gaussian pyramid model based on the layering result and calculate the historical training data set based on the Gaussian pyramid model to obtain the radar echo value.

[0038] In some examples, the data calculation module includes:

[0039] A model processing unit, configured to input the radar echo at a preset time in the historical training data set into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information;

[0040] A model training unit, configured to obtain the radar reflectivity residual information and the corresponding picture color information between the radar echo at the preset time and the target radar echo at the adjacent time, and train through a pre-trained deep learning model based on the per-pixel optical flow field information, the radar reflectivity residual information, the picture color information, and the prediction factors, so as to obtain the radar echo value.

[0041] In some examples, the prediction factors include at least one dynamic feature such as rotation, convergence, and divergence in the optical flow field.

[0042] In some examples, the device further includes:

[0043] A relationship construction module, configured to construct a relationship between the satellite observed brightness temperature and the radar echo value.

[0044] In some examples, the relationship construction module includes:

[0045] An interval classification unit, configured to perform interval classification on the radar echo value based on a preset network to obtain multiple categories;

[0046] A relationship construction unit, configured to construct a satellite-inverted radar echo classification model based on multiple categories, and the satellite-inverted radar echo classification model is used to simulate the relationship between the satellite observed brightness temperature and the radar echo value.

[0047] In a third aspect, the present disclosure provides a short-term and imminent prediction device, including:

[0048] A processor;

[0049] A memory, configured to store executable instructions;

[0050] Wherein, the processor is configured to read the executable instructions from the memory and execute the executable instructions to implement the short-term and imminent prediction method of the first aspect.

[0051] In a fourth aspect, the present disclosure provides a computer-readable storage medium, and the storage medium stores a computer program, and when the computer program is executed by a processor, the processor is enabled to implement the short-term and imminent prediction method of the first aspect.

[0052] The technical solutions provided by the embodiments of the present disclosure have the following advantages compared with the prior art:

[0053] The short-term and impending forecast method, device, equipment and medium of the embodiments of the present disclosure can obtain the radar reflectivity results retrieved by satellites, then perform pyramid layering on the radar reflectivity results to obtain corresponding layering results, then establish a Gaussian pyramid model based on the layering results, calculate the radar echo value based on the Gaussian pyramid model, and finally perform short-term and impending precipitation forecast based on the radar echo value. Thus, a Gaussian pyramid model is established, and the radar echo value is calculated based on the Gaussian pyramid model for short-term and impending precipitation forecast, thereby improving the forecast accuracy of extreme precipitation weather. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In combination with the accompanying drawings and with reference to the following specific embodiments, the above and other features, advantages and aspects of the embodiments of the present disclosure will become more obvious. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic, and the original components and elements are not necessarily drawn to scale.

[0055] Figure 1 It is a schematic flowchart of a short-term and impending forecast method provided by an embodiment of the present disclosure;

[0056] Figure 2 It is a schematic flowchart of another short-term and impending forecast method provided by an embodiment of the present disclosure;

[0057] Figure 3 It is a schematic structural diagram of a short-term and impending forecast device provided by an embodiment of the present disclosure;

[0058] Figure 4 It is a schematic structural diagram of a short-term and impending forecast equipment provided by an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0060] It should be understood that the various steps recorded in the method embodiments of the present disclosure can be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0061] As used herein, the term "including" and its variations are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0062] It should be noted that the concepts such as "first", "second", etc. mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0063] It should be noted that the modifications of "one" and "multiple" mentioned in this disclosure are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly specified in the context, it should be understood as "one or more".

[0064] The names of the messages or information exchanged between multiple devices in the embodiments of this disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0065] To solve the above problems, the embodiments of this disclosure provide a short-term and impending forecast method, apparatus, device and medium. The following will be combined with Figures 1 to 2 to describe in detail the short-term and impending forecast method provided by the embodiments of this disclosure.

[0066] Figure 1 The flowchart of a short-term and impending forecast method provided by the embodiments of this disclosure is shown.

[0067] In the embodiments of this disclosure, the short-term and impending forecast method can be executed by an electronic device. Among them, the electronic device can include but is not limited to devices such as computer devices, cloud servers or cloud server clusters.

[0068] As Figure 1 shown, the short-term and impending forecast method can include the following steps.

[0069] S110. Obtain the radar reflectivity result retrieved by the satellite.

[0070] In the embodiments of this disclosure, the electronic device can obtain the radar reflectivity result retrieved by the satellite.

[0071] Optionally, the radar reflectivity result can refer to the total backscattering cross section of the atmospheric substances in the spatial unit volume for the microwave emitted by the radar. It can indicate the occurrence location of precipitation and severe weather.

[0072] Specifically, the electronic device can obtain the radar reflectivity result retrieved by the satellite.

[0073] S120. Perform pyramid layering on the radar reflectivity result to obtain the corresponding layering result.

[0074] In the embodiment of the present disclosure, the electronic device can perform pyramid layering on the radar reflectivity result to obtain the corresponding layering result.

[0075] Specifically, the electronic device can perform pyramid layering on the radar reflectivity result, that is, perform pyramid layering on the image in the radar reflectivity result, so as to obtain the corresponding layering result.

[0076] S130. Establish a Gaussian pyramid model according to the layering result, and calculate the radar echo value based on the Gaussian pyramid model.

[0077] In the embodiment of the present disclosure, the electronic device can establish a Gaussian pyramid model according to the layering result, and calculate the radar echo value based on the Gaussian pyramid model.

[0078] Optionally, the Gaussian pyramid model is a commonly used multi-scale representation method in image processing, mainly used for downsampling and upsampling of images.

[0079] Optionally, the radar echo value refers to the energy intensity of the electromagnetic wave emitted by the radar reflected back after encountering the target during propagation. The radar detects and locates the target by receiving these reflected signals.

[0080] Specifically, the electronic device can establish a Gaussian pyramid model according to the layering result, and calculate the corresponding radar echo value through the Gaussian pyramid model.

[0081] S140. Perform short-term and nowcasting precipitation forecasting based on the radar echo value.

[0082] In the embodiment of the present disclosure, the electronic device can perform short-term and nowcasting precipitation forecasting based on the radar echo value.

[0083] Specifically, the electronic device can perform short-term and nowcasting precipitation forecasting according to the radar echo value, such as realizing the nowcasting extrapolation of the radar reflectivity factor within 0 - 2h, that is, for weather forecasting within 0 - 2h.

[0084] Thus, in the embodiment of the present disclosure, it is possible to obtain the radar reflectivity result retrieved by the satellite, then perform pyramid layering on the radar reflectivity result to obtain the corresponding layering result, then establish a Gaussian pyramid model according to the layering result, calculate the radar echo value based on the Gaussian pyramid model, and finally perform short-term and nowcasting precipitation forecasting based on the radar echo value. Thus, a Gaussian pyramid model is established, and the radar echo value is calculated based on the Gaussian pyramid model for short-term and nowcasting precipitation forecasting, thereby improving the forecasting accuracy of extreme precipitation weather.

[0085] Optionally, S120 may specifically include: performing pyramid layering on each frame of the radar reflectivity result based on the optical flow method to obtain the layering result.

[0086] In the embodiments of the present disclosure, the electronic device may perform pyramid layering on each frame of the radar reflectivity result based on the optical flow method to obtain the layering result.

[0087] Optionally, the optical flow method may be a method that utilizes the change of pixels in the time domain in an image sequence and the correlation between adjacent frames to find the corresponding relationship between the previous frame and the current frame, thereby calculating the motion information of an object between adjacent frames.

[0088] Specifically, the electronic device may use the optical flow method to perform pyramid layering on each frame of the radar reflectivity result, that is, for each frame of image, an image set with different resolutions from fine to coarse generated according to certain rules, to obtain the layering result.

[0089] Optionally, S120 may specifically include: using the largest-scale picture among multiple frames of the radar reflectivity result as the top layer and the original picture among multiple frames as the bottom layer to obtain the corresponding layering result.

[0090] Furthermore, the electronic device may use the largest-scale picture among multiple frames of the radar reflectivity result as the top layer and the original picture among multiple frames as the bottom layer to obtain the corresponding layering result.

[0091] Specifically, the electronic device may use the largest-scale picture among multiple frames of the radar reflectivity result as the top layer and the original picture as the bottom layer, that is, the resolution of the bottom layer is the highest and the data volume is the largest. As the number of layers increases, its resolution gradually decreases and the data volume also decreases proportionally, thereby obtaining an image set with gradually decreasing resolution arranged in a pyramid shape to obtain the corresponding layering result.

[0092] Next, start estimating the position of the next frame from the top layer as the initial position of the next layer, search downward along the pyramid, and repeat the estimation action until reaching the bottom layer of the pyramid. Such a search can not only solve the tracking of large-motion targets, but also solve the aperture problem to a certain extent (the same-sized window can cover as many corner points as possible on a large-scale picture, and these corner points cannot be covered on the original picture); it will also make the iterative process converge quickly, reduce the number of iterations, and reduce the risk of directly falling into a local optimum in the iteration.

[0093] Optionally, S130 may specifically include: obtaining a historical training data set containing the generation and disappearance characteristics of single entities under various weather conditions; establishing a Gaussian pyramid model according to the stratification result, and calculating the historical training data set based on the Gaussian pyramid model to obtain the radar echo value.

[0094] In the embodiments of the present disclosure, the electronic device may obtain a historical training data set containing the generation and disappearance characteristics of single entities under various weather conditions.

[0095] Specifically, the electronic device may obtain a historical training data set containing the generation and disappearance characteristics of single entities under various weather conditions based on multi-source meteorological data such as satellite data, radar data, ground automatic stations, and radiosondes.

[0096] Further, the electronic device may establish a Gaussian pyramid model according to the stratification result, and calculate the historical training data set based on the Gaussian pyramid model to obtain the radar echo value.

[0097] Specifically, the electronic device may establish a Gaussian pyramid model according to the stratification result, that is, a series of images obtained by successively downsampling the image with the highest resolution at the bottom. Before each sampling, the image is subjected to Gaussian filtering (Gaussian blur), and then the even rows and even columns of the image are removed, thereby reducing the resolution. This process is repeated until a certain termination condition is reached. The bottom of the Gaussian pyramid is the high-resolution representation of the original image, and the top is the low-resolution approximation. Then, the electronic device may calculate the historical training data set based on the Gaussian pyramid model to obtain the radar echo value.

[0098] Optionally, calculating the historical training data set based on the Gaussian pyramid model to obtain the radar echo value may include: inputting the radar echo at a preset time in the historical training data set into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information; obtaining the radar reflectivity residual information and the corresponding picture color information between the radar echo at the preset time and the target radar echo at an adjacent time, and training through a pre-trained deep learning model based on the pixel-by-pixel optical flow field information, the radar reflectivity residual information, the picture color information, and the prediction factors to obtain the radar echo value.

[0099] In the embodiments of the present disclosure, the electronic device may input the radar echo at a preset time in the historical training data set into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information.

[0100] Specifically, the electronic device may input the radar echo at a preset time in the historical training data set into the Gaussian pyramid model. For example, the radar echo of the previous 5 frames (the previous 30 minutes) is used as input data and input into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information.

[0101] Further, obtain the radar reflectivity residual information and the corresponding picture color information between the radar echo at the preset time and the target radar echo at the adjacent time, and train through a pre-trained deep learning model based on the per-pixel optical flow field information, the radar reflectivity residual information, the picture color information, and the prediction factors to obtain the radar echo value.

[0102] Optionally, the prediction factors include at least one dynamic feature such as rotation, convergence, and divergence in the optical flow field.

[0103] Specifically, the electronic device can obtain the radar reflectivity residual information and the corresponding picture color information between the radar echo at the preset time and the target radar echo at the adjacent time, and use the generated per-pixel optical flow field information, the radar reflectivity residual information, and the RGB picture information as machine learning input quantities to input into the ConvLSTM deep learning model for training, and add dynamic features such as rotation, convergence, and divergence in the optical flow field as prediction factors during the extrapolation process to obtain the radar echo value, realizing the nowcasting extrapolation of the 0-2h radar reflectivity factor.

[0104] Since there is no direct physical relationship between satellite observation brightness temperature and information such as radar echo value, a relationship needs to be fitted through statistical methods.

[0105] Optionally, the nowcasting and short-term forecasting method may further include: constructing a relationship between satellite observation brightness temperature and the radar echo value.

[0106] In the embodiments of the present disclosure, the electronic device can construct a relationship between satellite observation brightness temperature and the radar echo value.

[0107] Optionally, constructing a relationship between satellite observation brightness temperature and the radar echo value may include: based on a preset network, performing interval classification on the radar echo value to obtain multiple categories; constructing a satellite-inverted radar echo classification model based on the multiple categories, and constructing a relationship between satellite observation brightness temperature and the radar echo value.

[0108] In the embodiments of the present disclosure, the electronic device can perform interval classification on the radar echo value based on a preset network to obtain multiple categories.

[0109] Specifically, the electronic device uses the Unet network as the basic framework, and divides the radar echo samples into 16 categories at intervals of 5 dbz from 0 to 80 dbz. Among them, the Unet network is mainly divided into an encoding layer and a decoding layer: the encoding layer reduces the image size through convolution and downsampling to extract some low-order features in the image; the decoding layer performs convolution and upsampling on the image to obtain some high-order features in the image, and performs interval classification based on the low-order features and high-order features.

[0110] Furthermore, the electronic device can construct a satellite-inverted radar echo classification model based on multiple categories to establish the relationship between the satellite-observed brightness temperature and the radar echo value.

[0111] Specifically, the electronic device can construct a satellite-inverted radar echo classification model based on multiple categories. The satellite-inverted radar echo classification model can be used to simulate the relationship between the satellite-observed brightness temperature and the radar echo value. In the overall model structure of the satellite-inverted radar echo classification model, by releasing the information of the encoding layer and the pooling layer into the decoding layer, the deep and shallow features can be effectively combined to refine the image.

[0112] In the satellite-inverted radar echo classification model, Unet++ is used as the basic structure, and the pre-trained parameters of ResNet34 are loaded as the initialization parameters of the model encoding layer. Unet++ solves the information fusion gap between high and low orders; the pre-trained parameters accelerate the model convergence and enhance the model generalization ability; the symmetry of parameter initialization is broken to facilitate the extraction of different features by the hidden layer. And the terrain and underlying surface types are added to strengthen the physical influence mechanism of the model; finally, the differences in different infrared channels are added to the training data, which is beneficial to enhancing the learning ability of cloud thickness, reducing false alarms, and having a stronger fitting ability for strong echo signals, that is, enabling the satellite-inverted radar echo classification model to better simulate the relationship between the satellite-observed brightness temperature and the radar echo value.

[0113] Figure 2 The flowchart of a short-term and imminent prediction method provided by an embodiment of the present disclosure is shown.

[0114] As Figure 2 shown, the electronic device can obtain the radar reflectivity result inverted by the satellite, and then use the optical flow method to perform pyramid layering on each frame image of the radar reflectivity result. The largest-scale picture is used as the top layer, and the original picture is used as the bottom layer. Starting from the top layer, the position of the next frame is estimated as the initial position of the next layer, and the search is performed downward along the pyramid, repeating the estimation action until the bottom layer of the pyramid is reached. Such a search can not only solve the tracking of large moving targets but also, to a certain extent, solve the aperture problem (the same-sized window can cover as many corner points as possible on the large-scale picture, and these corner points cannot be covered on the original picture); it also enables the iterative process to converge quickly, reduces the number of iterations, and reduces the risk of directly falling into a local optimum in the iteration.

[0115] Further, based on multi-source meteorological data such as satellite data, radar data, surface automatic weather stations, and radiosondes, a historical training dataset containing the generation and dissipation characteristics of individual cells under various weather conditions is obtained. According to the stratification result, a Gaussian pyramid model is established. The radar echo of the first 5 frames (the first 30 minutes) is used as input data and input into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information. The generated pixel-by-pixel optical flow field information, together with the radar reflectivity residual information and RGB image information at adjacent times, is used as input for machine learning and input into the ConvLSTM deep learning model for training. During the extrapolation process, dynamic features such as rotation, convergence, and divergence in the optical flow field are added as forecasting factors to obtain the radar echo value, realizing the nowcasting extrapolation of the radar reflectivity factor within 0-2 hours.

[0116] Thus, using the convolutional neural network method, research and development on satellite observation data inversion technology is carried out. A nowcasting prediction model for radar echo is developed by combining the optical flow method and ConvLSTM deep learning. The impact of cloud microphysical parameterization schemes and land surface parameterization schemes on heavy rainfall is studied, and a combined parameterization scheme suitable for the study area is proposed. The short-term and nowcasting technology is coupled to form a 6-hour precipitation forecast.

[0117] Figure 3 The structural schematic diagram of a short-term and nowcasting prediction device provided by an embodiment of the present disclosure is shown.

[0118] As Figure 3 shown, the short-term and nowcasting prediction device 300 may include a result acquisition module 310, an image stratification module 320, a data calculation module 330, and a precipitation forecast module 340.

[0119] The result acquisition module 310 may be used to acquire the radar reflectivity result retrieved from the satellite.

[0120] The image stratification module 320 may be used to perform pyramid stratification on the radar reflectivity result based on the optical flow method to obtain the corresponding stratification result.

[0121] The data calculation module 330 may be used to establish a Gaussian pyramid model according to the stratification result and calculate the radar echo value based on the Gaussian pyramid model.

[0122] The precipitation forecast module 340 may be used to perform short-term and nowcasting precipitation forecasting based on the radar echo value.

[0123] Thus, in the embodiments of the present disclosure, the radar reflectivity results retrieved by the satellite can be obtained. Then, the radar reflectivity results are subjected to pyramid layering to obtain corresponding layering results. Next, a Gaussian pyramid model is established based on the layering results, and the radar echo value is calculated based on the Gaussian pyramid model. Finally, short-term and nowcasting precipitation forecasts are made based on the radar echo value. Thus, a Gaussian pyramid model is established, and the radar echo value is calculated based on the Gaussian pyramid model to perform short-term and nowcasting precipitation forecasts, thereby improving the forecast accuracy of extreme precipitation weather.

[0124] In some embodiments of the present disclosure, the image layering module 320 may specifically include a first processing unit.

[0125] The first processing unit may be configured to perform pyramid layering on each frame image of the radar reflectivity results based on the optical flow method to obtain the layering results.

[0126] In some embodiments of the present disclosure, the image layering module 320 may specifically include a second processing unit.

[0127] The second processing unit may be configured to use the largest-scale picture among the multi-frame images of the radar reflectivity results as the top layer and the original picture among the multi-frame images as the bottom layer to obtain corresponding layering results.

[0128] In some embodiments of the present disclosure, the data calculation module 330 may specifically include a data acquisition unit and a model calculation unit.

[0129] The data acquisition unit may be configured to acquire a historical training data set including the generation and disappearance characteristics of single entities under various weather conditions.

[0130] The model calculation unit may be configured to establish a Gaussian pyramid model based on the layering results and calculate the historical training data set based on the Gaussian pyramid model to obtain the radar echo value.

[0131] In some embodiments of the present disclosure, the data calculation module 330 may specifically include a model processing unit and a model training unit.

[0132] The model processing unit may be configured to input the radar echo of a preset time in the historical training data set into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information.

[0133] The model training unit may be configured to obtain the radar reflectivity residual information and the corresponding picture color information between the radar echo of the preset time and the target radar echo of the adjacent time, and train through a pre-trained deep learning model based on the pixel-by-pixel optical flow field information, the radar reflectivity residual information, the picture color information, and the forecast factors to obtain the radar echo value.

[0134] In some embodiments of the present disclosure, the predictors include dynamic features such as rotation, convergence, and divergence in the optical flow field.

[0135] In some embodiments of the present disclosure, the short-term and nowcasting device 300 may further include a relationship construction module.

[0136] The relationship construction module may be used to construct the relationship between the satellite observed brightness temperature and the radar echo value.

[0137] In some embodiments of the present disclosure, the relationship construction module may specifically include an interval classification unit and a relationship construction unit.

[0138] The interval classification unit may be used to perform interval classification on the radar echo value based on a preset network to obtain multiple categories.

[0139] The relationship construction unit may be used to construct a satellite-inverted radar echo classification model based on multiple categories, and the satellite-inverted radar echo classification model is used to simulate the relationship between the satellite observed brightness temperature and the radar echo value.

[0140] It should be noted that Figure 3 the short-term and nowcasting device 300 shown may execute Figures 1 to 2 each step in the method embodiments shown, and achieve Figures 1 to 2 each process and effect in the method embodiments shown, which will not be elaborated here.

[0141] Figure 4 FIG. shows a schematic structural diagram of a short-term and nowcasting device provided by an embodiment of the present disclosure.

[0142] In some embodiments of the present disclosure, Figure 4 the short-term and nowcasting device shown may be an electronic device. Specifically, the electronic device may include, but is not limited to, devices such as computer devices, cloud servers, or cloud server clusters.

[0143] As Figure 4 shown, the short-term and nowcasting device may include a processor 401 and a memory 402 storing computer program instructions.

[0144] Specifically, the above-mentioned processor 401 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0145] The memory 402 may include a mass memory for information or instructions. By way of example and not limitation, the memory 402 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 402 may include removable or non-removable (or fixed) media. Where appropriate, the memory 402 may be internal or external to the integrated gateway device. In a particular embodiment, the memory 402 is a non-volatile solid-state memory. In a particular embodiment, the memory 402 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM), or a flash memory, or a combination of two or more of these.

[0146] The processor 401 reads and executes the computer program instructions stored in the memory 402 to perform the steps of the short-term prediction method provided by the embodiments of the present disclosure.

[0147] In one example, the short-term prediction device may further include a transceiver 403 and a bus 404. Among them, as Figure 4 shown, the processor 401, the memory 402, and the transceiver 403 are connected through the bus 404 and complete communication with each other.

[0148] The bus 404 includes hardware, software, or both. By way of example and not limitation, the bus can include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side BUS (FSB), a Hyper Transport (HT) interconnect, an Industrial Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable bus or a combination of two or more of these. Where appropriate, the bus 404 can include one or more buses. Although embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect.

[0149] Embodiments of the present disclosure also provide a computer-readable storage medium that can store a computer program, which, when executed by a processor, causes the processor to implement the short-term and imminent forecasting method provided by the embodiments of the present disclosure.

[0150] The above storage medium can, for example, include a memory 402 storing computer program instructions, and the above instructions can be executed by a processor 401 of the short-term and imminent forecasting device to complete the short-term and imminent forecasting method provided by the embodiments of the present disclosure. Optionally, the storage medium can be a non-transitory computer-readable storage medium. For example, the non-transitory computer-readable storage medium can be a ROM, a Random Access Memory (RAM), a Compact Disc ROM (CD-ROM), a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0151] It should be noted that in this document, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article, or device that comprises a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or device.

[0152] The above are only specific embodiments of the present disclosure, enabling those skilled in the art to understand or implement the present disclosure. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to these embodiments described herein, but rather will conform to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A short-term forecast method, characterized in that: include: Obtain radar reflectivity results from satellite inversion; Performing pyramid stratification on the radar reflectivity result to obtain corresponding stratification results; Establishing a Gaussian pyramid model according to the stratification result, and calculating a radar echo value based on the Gaussian pyramid model; A short-term precipitation forecast is performed based on the radar echo value.

2. The method according to claim 1, characterized in that The performing pyramid stratification on the radar reflectivity result to obtain the corresponding stratified result includes: Based on the optical flow method, pyramid layering is performed for each frame image of the radar reflectivity result to obtain the layered result.

3. The method according to claim 2, characterized in that The step of performing pyramid layering on each frame of the radar reflectivity result based on the optical flow method to obtain the layered result includes: The largest scale image in the multi-frame images of the radar reflectivity result is used as the top layer, and the original image in the multi-frame images is used as the bottom layer to obtain the corresponding layered results.

4. The method according to claim 1, characterized in that: The step of establishing a Gaussian pyramid model according to the stratification result and calculating a radar echo value based on the Gaussian pyramid model includes: Obtain a historical training dataset containing the characteristics of monomer growth and disappearance under various weather conditions; A Gaussian pyramid model is established according to the stratification result, and the historical training data set is calculated based on the Gaussian pyramid model to obtain the radar echo value.

5. The method according to claim 4, characterized in that The calculating the historical training data set based on the Gaussian pyramid model to obtain the radar echo value includes: Inputting the radar echo of the preset time in the historical training data set into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information; The radar reflectivity residual information and the corresponding image color information between the radar echo at the preset time and the target radar echo at the adjacent time are obtained, and based on the pixel-by-pixel optical flow field information, the radar reflectivity residual information, the image color information and the prediction factor, a pre-trained deep learning model is used for training to obtain the radar echo value.

6. The method according to claim 5, characterized in that The predictor includes at least one dynamic feature of rotation, convergence and divergence in the optical flow field.

7. The method according to claim 1, characterized in that The method further comprises: A relationship between the satellite observed brightness temperature and the radar echo value is constructed.

8. The method according to claim 7, characterized in that The step of constructing the relationship between the satellite observed brightness temperature and the radar echo value includes: Based on a preset network, the radar echo values ​​are classified into multiple categories by interval classification; A satellite inversion radar echo classification model is constructed based on multiple categories, and the satellite inversion radar echo classification model is used to simulate the relationship between the satellite observed brightness temperature and the radar echo value.

9. A short-term forecasting device, characterized in that: include: The result acquisition module is used to obtain the radar reflectivity results inverted by satellite; An image stratification module, used for performing pyramid stratification on the radar reflectivity result to obtain a corresponding stratification result; A data calculation module, used for establishing a Gaussian pyramid model according to the stratification result, and calculating the radar echo value based on the Gaussian pyramid model; The precipitation forecast module is used to make a short-term precipitation forecast based on the radar echo value.

10. The device according to claim 9, characterized in that The image layering module comprises: The first processing unit is used to perform pyramid layering on each frame image of the radar reflectivity result based on the optical flow method to obtain the layered result.

11. The device according to claim 10, characterized in that The image layering module comprises: The second processing unit is used to use the largest scale image in the multi-frame images of the radar reflectivity result as the top layer, and use the original image in the multi-frame images as the bottom layer to obtain corresponding layered results.

12. The device according to claim 9, characterized in that The data calculation module includes: A data acquisition unit, used to acquire a historical training data set containing the monomer generation and disappearance characteristics under various weather conditions; A model calculation unit is used to establish a Gaussian pyramid model according to the stratification result, and calculate the historical training data set based on the Gaussian pyramid model to obtain the radar echo value.

13. The device according to claim 12, characterized in that The data calculation module includes: A model processing unit, used for inputting radar echoes at a preset time in the historical training data set into the Gaussian pyramid model to generate pixel-by-pixel optical flow field information; The model training unit is used to obtain the radar reflectivity residual information and the corresponding image color information between the radar echo at the preset time and the target radar echo at the adjacent time, and to obtain the radar echo value by training through a pre-trained deep learning model based on the pixel-by-pixel optical flow field information, the radar reflectivity residual information, the image color information and the prediction factor.

14. The device according to claim 13, characterized in that The predictor includes at least one dynamic feature of rotation, convergence and divergence in the optical flow field.

15. The device according to claim 9, characterized in that The device also includes: The relationship building module is used to build the relationship between the satellite observed brightness temperature and the radar echo value.

16. The device according to claim 15, characterized in that The relationship building module includes: An interval classification unit, used for classifying the radar echo values ​​into multiple categories based on a preset network; The relationship building unit is used to build a satellite inversion radar echo classification model based on multiple categories, and the satellite inversion radar echo classification model is used to simulate the relationship between the satellite observed brightness temperature and the radar echo value.

17. A short-term forecast device, characterized in that: include: processor; A memory for storing executable instructions; The processor is used to read the executable instructions from the memory and execute the executable instructions to implement the short-term forecast method described in any one of claims 1 to 8.

18. A non-volatile computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the short-term and impending forecast method according to any one of claims 1 to 8.

Images