A meteorological satellite data set generation method, device and terminal equipment thereof

Abstract

The application provides a meteorological satellite data set generation method, device and terminal equipment, which can improve the cloud block recognition accuracy and help simulate the clear sky low temperature condition after night radiation cooling. The meteorological satellite data set generation method comprises the following steps: calculating the cloud cover of each historical meteorological satellite data set through cloud grid points; the historical meteorological satellite data set is historical infrared data, historical visible light data and historical vertical temperature distribution data in the same observation time period of the same observation region; the cloud cover is evenly divided into a plurality of cloud cover intervals, and cloud cover distribution data is obtained according to the cloud cover intervals; the same number of data samples are extracted from each cloud cover interval to generate a meteorological satellite data set; and the daytime clear sky area brightness temperature in the meteorological satellite data set is adjusted downward to simulate the clear sky low temperature condition after night radiation cooling.

Description

Technical Field

[0001] This invention relates to the field of meteorological satellite data processing, and more specifically, to a method, apparatus, and terminal equipment for generating meteorological satellite datasets. Background Technology

[0002] Meteorological satellites are crucial tools for monitoring weather, helping weather forecasters and meteorologists analyze cloud changes to determine the evolution of different weather phenomena. They also serve as a source of cloud observation data for numerical weather prediction models. Accurately identifying cloud locations not only enhances the monitoring capabilities for severe weather (such as severe convection, typhoons, fog, and fronts) but also reduces errors in the initial field of weather forecast models, thereby improving the accuracy of numerical weather prediction models.

[0003] Meteorological satellites typically detect the intensity of visible and infrared light from the Earth's surface to identify clouds. Visible light, with wavelengths of 0.4-0.7 micrometers, is the electromagnetic spectrum visible to the human eye. When visible light (sunlight) from the sun illuminates the Earth during the day, meteorological satellites can generate true-color and black-and-white visible light satellite cloud images that are most intuitive for humans by detecting the reflectance of the visible light spectrum, effectively reflecting the location of liquid clouds in the lower and middle layers of the atmosphere. However, high-altitude solid clouds (ice clouds) are highly transparent, making them difficult to monitor in visible light satellite cloud images. Infrared light, with wavelengths of 0.9-13.4 micrometers, is the electromagnetic spectrum invisible to the naked eye, and its intensity reflects the surface of objects. Meteorological satellites can, day and night, invert the detected infrared light intensity into cloud brightness temperature, which helps identify the location of high-altitude ice clouds that are opaque to infrared light.

[0004] With the complementarity of visible and infrared light, cloud identification algorithms are relatively simple and direct during the day. However, after nightfall, visible light reflectance is lost, leaving only infrared data. When ground cooling and temperature inversion occur at night, the surface temperature can be similar to that of low clouds / fog, making it impossible to directly determine the location of low clouds. To address this issue, the industry has begun developing deep learning models that can invert nighttime infrared satellite data into visible light satellite reflectance for cloud block identification. However, because the datasets used to train these deep learning models only contain daytime satellite data, and land has a low specific heat capacity, the land cooling effect is significant under clear skies at night. This causes the models to tend to incorrectly invert the infrared topographic brightness and temperature texture of clear-sky areas at night into clouds, severely impacting the accuracy of cloud block identification. Summary of the Invention

[0005] The purpose of this invention is to provide a method, apparatus and terminal equipment for generating meteorological satellite datasets, which can improve the accuracy of cloud block identification and help simulate the low temperature conditions of clear sky after nighttime radiative cooling.

[0006] The embodiments of the present invention are implemented as follows:

[0007] A method for generating meteorological satellite datasets, comprising the following steps:

[0008] The cloud coverage rate of each historical meteorological satellite data set is calculated using cloud grid points; the historical meteorological satellite data set consists of historical infrared light data, historical visible light data, and historical vertical temperature distribution data within the same observation period in the same observation area.

[0009] The cloud coverage rate is divided into several cloud coverage rate intervals, and cloud coverage rate distribution data is obtained based on the cloud coverage rate intervals.

[0010] The same number of data samples are extracted from each cloud coverage interval to generate a meteorological satellite dataset.

[0011] In a preferred embodiment of the present invention, the specific method for acquiring the above-mentioned historical satellite data set includes the following steps:

[0012] Acquire infrared brightness temperature data and visible reflectance data for a specified time period during the day, and project the acquired data onto a map coordinate system; combine infrared and visible light data from the same observation time into a satellite data set;

[0013] Based on the observation area and time of the satellite data set, the vertical temperature distribution of the corresponding area and time is obtained. The vertical temperature distribution is then projected onto the map coordinate system to obtain the historical meteorological satellite data set.

[0014] In a preferred embodiment of the present invention, the method for calculating the cloud coverage rate includes:

[0015] Obtain cloud grid points for the corresponding observation area and time period;

[0016] The proportion of cloud grid points to all grid points in the observation area is calculated and used as the cloud coverage rate for the corresponding observation time and region. The location of cloud grid points and cloud coverage rate are stored as new dimensions in the meteorological satellite data set.

[0017] In a preferred embodiment of the present invention, the above-mentioned cloud grid acquisition method includes:

[0018] Data grid points with a visible light red band reflectance greater than 0.15 within the historical meteorological satellite data set are designated as cloud grid points;

[0019] Data points with a brightness-temperature difference of Tb(8.6μm)-Tb(10.8μm)≥2K are designated as cloud grid points;

[0020] From the atmospheric environmental data, the temperature T2m at a distance of nearly 2 meters was read, and the data grid points where T2m-Tb(10.4μm)>20K were designated as cloud grid points.

[0021] In a preferred embodiment of the present invention, the cloud coverage intervals are continuous and have the same length.

[0022] In a preferred embodiment of the present invention, the above-mentioned data sample extraction method includes: calculating the number of data samples within the cloud coverage interval, using the minimum number of samples S as the extraction number, extracting S data samples from each cloud coverage interval and storing them in the historical meteorological satellite data set to generate a meteorological satellite dataset.

[0023] In a preferred embodiment of the present invention, the above-mentioned meteorological satellite dataset is used to simulate the low temperature conditions under clear skies after surface radiative cooling in clear skies at night. The simulation method includes:

[0024] Obtain cloud grid points and treat non-cloud grid points as clear sky grid points;

[0025] The infrared light channel brightness temperature value of each clear sky point on land was lowered, and the adjusted data was saved as a separate nighttime clear sky low temperature dataset.

[0026] A meteorological satellite dataset generation device, the device comprising:

[0027] The cloud coverage calculation module is used to calculate the cloud coverage of each historical meteorological satellite data group by using cloud grid points.

[0028] The cloud coverage balancing module is used to divide the cloud coverage into several cloud coverage intervals on an average basis, and obtain cloud coverage distribution data based on the cloud coverage intervals.

[0029] The meteorological satellite dataset generation module is used to generate meteorological satellite datasets from the data processed by the cloud coverage balancing module.

[0030] A terminal device includes a memory and a processor. The memory stores a computer program, which executes the above-described meteorological satellite dataset generation method when run on the processor.

[0031] A readable storage medium storing a computer program that runs the above-described meteorological satellite dataset generation method on a processor.

[0032] The beneficial effects of this invention are as follows: By balancing cloud coverage data, this invention avoids machine learning or deep learning models from focusing on cloud coverage during training. Simultaneously, the processed meteorological satellite dataset can be used to simulate clear-sky low-temperature conditions after nighttime radiative cooling. The daytime clear-sky brightness temperature of the meteorological satellite dataset is lowered and added as artificial data to the nighttime clear-sky low-temperature dataset. This allows machine learning or deep learning models to simultaneously access data samples from both daytime and nighttime clear-sky areas, reducing the chance of the model misinterpreting low surface brightness temperature signals in the nighttime clear-sky area as clouds, thus improving the model's accuracy. Attached Figure Description

[0033] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a flowchart illustrating the meteorological satellite dataset generation method according to the first embodiment of the present invention;

[0035] Figure 2 This is a schematic diagram of the brightness temperature correction process for clear-sky infrared satellite data according to the second embodiment of the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0037] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0038] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0039] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.

[0040] First Embodiment

[0041] This embodiment addresses existing satellite data combination methods, which only use geographical location and solar time as screening criteria to collect infrared and visible light data without considering data balance. This causes machine learning or deep learning models to focus on cloud coverage during training.

[0042] Based on this, this embodiment provides a method for generating meteorological satellite datasets, which has practical applications in meteorological monitoring, weather forecasting, aviation weather forecasting, marine weather forecasting, and numerical weather prediction data assimilation. The generation of this meteorological satellite dataset includes the following steps:

[0043] 1. Acquisition and processing of infrared and visible light data

[0044] Historical meteorological satellite data files are decoded according to their data format using conventional methods to obtain infrared brightness temperature data and visible light reflectance data for a specified daytime period. The infrared data uses electromagnetic bands with wavelengths greater than 0.86 micrometers, and the visible light data uses electromagnetic bands with wavelengths between 0.4 and 0.86 micrometers. In actual use, meteorological satellites generally detect near-infrared data at 0.86 micrometers, and some people use it as visible light. Therefore, the visible light data in this embodiment actually includes near-infrared light data.

[0045] The acquired infrared and visible light data are projected onto a specified map coordinate system, using geometric equations and interpolation projection methods.

[0046] Infrared and visible light data from the same observation time are combined into a satellite data set.

[0047] 2. Acquisition and processing of vertical temperature distribution within the observed area and time period.

[0048] Based on the observation area and time of each satellite data set, the historical three-dimensional atmospheric environment field data file for the corresponding time is read using conventional methods and decoded according to its data format;

[0049] Obtain the vertical temperature distribution for the corresponding observation area and time;

[0050] The data will be projected onto the map coordinate system using conventional geometric equations and interpolation methods;

[0051] The historical meteorological satellite data set was obtained, which includes historical infrared light data, historical visible light data, and historical vertical temperature distribution data for the same observation period in the same observation area.

[0052] 3. Calculate the cloud cover for each set of historical meteorological satellite data:

[0053] The data grid points within the historical observation area whose visible light red band reflectance is greater than a specified value X% are designated as cloud grid points. In this embodiment, X is 0.15. The wavelength of the visible light red band is approximately 0.64 micrometers.

[0054] Then, the brightness temperature Tb(8.6μm) of the meteorological satellite data set with a wavelength of approximately 8.6 micrometers and the brightness temperature Tb(10.8μm) with a wavelength of approximately 10.8 micrometers were calculated, and the data grid points with a difference Tb(8.6μm)-Tb(10.8μm)≥2K were marked as cloud grid points;

[0055] Then, from the atmospheric environmental data, the temperature T2m at a distance of nearly 2 meters was read, and the data grid points where T2m-Tb(10.4μm)>20K were designated as cloud grid points;

[0056] The proportion of cloud grid points to all grid points in the observation area is used as the cloud coverage rate for that observation period and area, and the location of the cloud grid points is stored as a new data dimension in the meteorological satellite data set.

[0057] 4. Cloud coverage data balancing processing

[0058] The cloud coverage rate is divided into several cloud coverage rate intervals, and cloud coverage rate distribution data is obtained based on these intervals. Specifically, the cloud coverage rate from 0% to 100% is divided into several cloud coverage rate intervals, each interval accounting for a specified percentage Y. In this embodiment, Y is 10%, meaning there are a total of 10 data intervals (0-10%, 11%-20%, 21%-30%, ..., 91%-100%).

[0059] The data distribution of cloud coverage is created according to the cloud coverage interval, and the number of data samples within the cloud coverage interval is calculated. The minimum number of samples S is used as the number of samples extracted from each cloud coverage interval.

[0060] 5. Extract the same number of data samples from each cloud coverage interval to generate a meteorological satellite dataset.

[0061] The observation data sets for this cloud coverage range are labeled with integers;

[0062] Using a random integer generator, generate S integers and extract the corresponding observation data sets into the final meteorological satellite dataset.

[0063] Second Embodiment

[0064] With the complementarity of visible and infrared light, cloud identification algorithms are relatively simple and direct during the day. However, after nightfall, visible light reflectance is lost, leaving only infrared data. When ground cooling and temperature inversion occurs at night, the surface temperature can be similar to that of low clouds / fog, making it impossible to directly determine the location of low clouds.

[0065] Currently, the industry can use deep learning models to invert infrared satellite data at night into visible satellite reflectance for cloud block identification. However, because the datasets used to train these deep learning models only contain daytime satellite data, and land has a low specific heat capacity, the cooling effect on land under clear skies at night is significant. This causes the models to tend to incorrectly invert the infrared topographic brightness temperature texture of clear-sky areas at night into clouds, severely affecting the accuracy of cloud block identification.

[0066] Based on this, this embodiment uses the final meteorological satellite dataset obtained in the first embodiment to simulate the interday minimum brightness temperature value after surface radiative cooling in clear sky areas at night. The specific operation method is as follows:

[0067] Read the cloud grid locations calculated in step 3 from the meteorological satellite data set, and mark the non-cloud grid locations as the clear sky grid locations;

[0068] The surface type data is decoded according to the relevant data format using conventional methods, the distribution of surface types in the observation area is read, and the data is projected onto the specified map coordinate system using conventional geometric equations and interpolation methods.

[0069] Based on the land surface type data, mark the locations of clear, empty points above non-water bodies as clear, empty points on land for that observation time and region;

[0070] The brightness temperature values ​​of each infrared channel at each clear-sky land location are adjusted downwards. The specific adjustment range can be obtained by freely selecting a value, using a boundary layer temperature change model in atmospheric boundary layer meteorology, or comparing it with the brightness temperature distribution closest to the observation time before sunrise. Alternatively, a reasonable number can be chosen, calculated using boundary layer theory, or compared with data from the moment before sunrise as the adjustment range. Afterwards, the adjusted brightness temperature values ​​are stored as new data (without overwriting the original data).

[0071] Third Embodiment

[0072] This embodiment provides a meteorological satellite dataset generation device, which includes:

[0073] The cloud coverage calculation module is used to calculate the cloud coverage of each historical meteorological satellite data group by using cloud grid points.

[0074] The cloud coverage balancing module is used to divide the cloud coverage into several cloud coverage intervals on an average basis, and obtain cloud coverage distribution data based on the cloud coverage intervals.

[0075] The meteorological satellite dataset generation module is used to generate meteorological satellite datasets from the data processed by the cloud coverage balancing module.

[0076] The meteorological satellite dataset generation device disclosed in this embodiment is used in conjunction with a cloud coverage calculation module, a cloud coverage balancing module, and a meteorological satellite dataset generation module to execute the meteorological satellite dataset generation method described in the above embodiment. The implementation schemes and beneficial effects involved in the above embodiment are also applicable in this embodiment, and will not be repeated here.

[0077] It is understood that this application relates to a terminal device, which includes a memory and a processor, wherein the memory stores a computer program, and the computer program executes the meteorological satellite dataset generation method when it runs on the processor.

[0078] It is understood that this application relates to a readable storage medium storing a computer program that, when run on a processor, executes the meteorological satellite dataset generation method described in this application.

[0079] In summary, the dataset generation method provided in this embodiment differs from the current methods for composing meteorological satellite observation datasets. This invention achieves a balance between clear sky and cloud cover in the data samples, avoiding situations where machine learning or deep learning models focus on cloud cover during training. Simultaneously, this invention also lowers the brightness temperature of clear sky areas during the daytime in the data samples to simulate the low temperature of clear sky after radiative cooling at night, and adds this artificial data to the final dataset. This allows machine learning or deep learning models to simultaneously grasp data samples of clear sky areas during both day and night, reducing the chance of the model misinterpreting low surface brightness temperature signals in clear sky areas at night as clouds, thus improving the model's accuracy.

[0080] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, as an alternative implementation, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0081] In addition, the functional modules or units in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0082] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0083] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for generating meteorological satellite datasets, characterized in that, The method includes the following steps: The cloud coverage rate of each historical meteorological satellite data set is calculated using cloud grid points; the historical meteorological satellite data set consists of historical infrared light data, historical visible light data, and historical vertical temperature distribution data within the same observation period in the same observation area. The cloud coverage rate is divided into several cloud coverage rate intervals, and cloud coverage rate distribution data is obtained based on the cloud coverage rate intervals. The same number of data samples are extracted from each of the cloud coverage intervals to generate a meteorological satellite dataset; The meteorological satellite dataset is used to simulate the low temperatures in clear-sky conditions after surface radiative cooling in clear-sky areas at night. The simulation method includes: Obtain the cloud grid points, and use the non-cloud grid points as clear sky grid points; The infrared light channel brightness temperature value of each clear sky point on land is lowered, and the adjusted data is saved as a separate nighttime clear sky low temperature dataset.

2. The meteorological satellite dataset generation method according to claim 1, characterized in that, The specific method for acquiring the historical meteorological satellite data set includes the following steps: Acquire infrared brightness temperature data and visible reflectance data for a specified time period during the day, and project the acquired data onto a map coordinate system; combine infrared and visible light data from the same observation time into a satellite data set; Based on the observation area and time of the satellite data set, the vertical temperature distribution of the corresponding area and time is obtained, and the vertical temperature distribution is projected onto the map coordinate system to obtain the historical meteorological satellite data set.

3. The meteorological satellite dataset generation method according to claim 1, characterized in that, The method for calculating the cloud coverage rate includes: Obtain cloud grid points for the corresponding observation area and time period; The proportion of the cloud grid points to all grid points in the observation area is calculated and used as the cloud coverage rate for the corresponding observation time and area. The location of the cloud grid points and the cloud coverage rate are stored as new dimensions in the meteorological satellite data set.

4. The meteorological satellite dataset generation method according to claim 1, characterized in that, Methods for obtaining cloud grid points include: Data grid points with a visible light red band reflectance greater than 0.15 within the aforementioned historical meteorological satellite data set are designated as cloud grid points; The brightness temperature difference Tb (8.6µm) Data grid points with Tb(10.8µm)≥2K are cloud grid points; From the atmospheric environmental data, the temperature T2m at a distance of nearly 2 meters was read, and T2m was... Data grid points with Tb(10.4µm)>20K are cloud grid points.

5. The meteorological satellite dataset generation method according to claim 1, characterized in that, The cloud coverage intervals are continuous and have the same length.

6. The meteorological satellite dataset generation method according to claim 1, characterized in that, The data sample extraction method includes: calculating the number of data samples within the cloud coverage interval, using the minimum number of samples S as the extraction number, extracting S data samples from each cloud coverage interval and storing them in the historical meteorological satellite data set to generate a meteorological satellite dataset.

7. A meteorological satellite dataset generation apparatus, comprising the meteorological satellite dataset generation method according to any one of claims 1 to 6, characterized in that, The device includes: The cloud coverage calculation module is used to calculate the cloud coverage of each historical meteorological satellite data group by using cloud grid points. The cloud coverage balancing module is used to divide the cloud coverage into several cloud coverage intervals on an average basis, and obtain cloud coverage distribution data based on the cloud coverage intervals. The meteorological satellite dataset generation module is used to generate meteorological satellite datasets from the data processed by the cloud coverage balancing module. The meteorological satellite dataset is used to simulate the low temperatures in clear-sky conditions after surface radiative cooling in clear-sky areas at night. The simulation method includes: Obtain the cloud grid points, and use the non-cloud grid points as clear sky grid points; The infrared light channel brightness temperature value of each clear sky point on land is lowered, and the adjusted data is saved as a separate nighttime clear sky low temperature dataset.

8. A terminal device, characterized in that, The system includes a memory and a processor, wherein the memory stores a computer program that, when executed on the processor, performs the meteorological satellite dataset generation method according to any one of claims 1 to 6.

9. A readable storage medium, characterized in that, It stores a computer program that, when run on a processor, executes the meteorological satellite dataset generation method according to any one of claims 1 to 6.

Images