Dust aerosol identification method, system, model training method, medium and equipment

Abstract

The present invention provides a dust aerosol identification method, system, model training method, medium, and equipment. The dust aerosol identification model training method includes obtaining training thermal infrared observation brightness temperature data, training simulated clear sky brightness temperature data, and training surface information data; obtaining clear sky labels, cloud labels, and dust labels corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data; and training the dust aerosol identification model based on the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, the training surface information data, and the clear sky labels, cloud labels, and dust labels. The present invention can eliminate the influence of clouds and the surface, accurately identify thin dust aerosols in all weather conditions, and does not require an additional cloud detection program as preprocessing, thereby reducing the steps in the dust identification process, improving the accuracy of dust identification, and effectively achieving the distinction between clear sky, clouds, and dust.

Description

Technical Field

[0001] The present invention relates to an aerosol identification method, and in particular to a dust aerosol identification method, system, model training method, medium and equipment. Background Art

[0002] Dust aerosols, a naturally occurring aerosol, influence global climate and ecosystems by impacting the atmospheric radiation budget, causing cloud contamination, and generating geochemical reactions during transport. Dust aerosols typically occur in arid and semi-arid regions, and are most common in spring in Asia. Satellite observations enable large-scale, high-temporal and spatial resolution monitoring of dust dynamics. Because dust aerosols share spectral signatures similar to those of other aerosols, thin clouds, and bright surfaces (barren land or deserts), their identification is challenging. Numerous dust aerosol identification algorithms have been proposed, generally categorized as physics-based and machine learning-based.

[0003] Physics-based dust detection algorithms typically use the brightness temperature difference of the thermal infrared channel or a specific dust index to identify dust. For example, dust exhibits a negative brightness temperature difference between 10μm and 12μm. A classic physics-based dust detection algorithm uses the brightness temperature difference between 10μm and 11μm to identify dust, assuming that the brightness temperature difference for dust in this channel combination is typically 1 to 5K, while negative values ​​are observed for clouds and clear skies. However, the setting of this threshold is uncertain and varies depending on atmospheric conditions and surface types. The threshold is generally not fixed and requires empirical judgment. Other physical methods also typically require threshold settings, making them difficult to apply to diverse and complex climatic and geographical conditions.

[0004] Machine learning-based algorithms are more flexible. Previous studies have compared the recognition performance of different machine learning models, and their results generally outperform physical algorithms. However, most of these machine learning algorithms require visible light bands as input, making them inoperable at night and limiting the all-weather recognition of dust. Secondly, these algorithms do not effectively address the difficulties in dust identification (easily confused features), and dust identification is still affected by these climatic and geographical characteristics. In addition, some algorithms use aerosol optical depth (AOD) products as labels for modeling, which can introduce errors in dust identification caused by the influence of other aerosols. Summary of the Invention

[0005] In view of the shortcomings of the prior art described above, the object of the present invention is to provide a dust aerosol identification method, system, model training method, medium and equipment to solve the technical problem that the prior art lacks a method that can accurately and automatically identify dust aerosols in all-weather conditions.

[0006] To achieve the above-mentioned objectives and other related objectives, the first aspect of the present invention provides a training method for a dust aerosol recognition model, comprising obtaining training thermal infrared observation brightness temperature data, training simulated clear sky brightness temperature data and training surface information data; obtaining clear sky labels, cloud labels and dust labels corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data and the training surface information data; and training a dust aerosol recognition model based on the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, the training surface information data and the clear sky labels, cloud labels and dust labels.

[0007] In an embodiment of the first aspect, the method of obtaining training thermal infrared observation brightness temperature data, training simulated clear sky brightness temperature data and training surface information data includes obtaining the training thermal infrared observation brightness temperature data based on satellite observation data; obtaining the training simulated clear sky brightness temperature data and the training surface information data based on satellite observation data, medium resolution imaging spectrometer data and atmospheric reanalysis data.

[0008] In an embodiment of the first aspect, the obtaining of the training simulated clear sky brightness temperature data based on satellite observation data, medium-resolution imaging spectrometer data and atmospheric reanalysis data includes: obtaining observation angle data based on satellite observation data; obtaining surface emissivity data based on medium-resolution imaging spectrometer data; obtaining atmospheric profile data, surface temperature and surface pressure data based on atmospheric reanalysis data; and obtaining the training simulated clear sky brightness temperature data based on the atmospheric profile data, the surface temperature, the surface pressure data, the observation angle data and the surface emissivity data.

[0009] In an embodiment of the first aspect, the clear sky tag, cloud tag, and dust tag correspond to cloud-aerosol lidar and infrared sounder satellite data products.

[0010] In an embodiment of the first aspect, the obtaining of clear sky labels, cloud labels, and dust labels corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data includes: establishing a correspondence between the satellite observation data, the medium-resolution imaging spectrometer data, and the atmospheric reanalysis data and the cloud-aerosol lidar and infrared detector satellite data products based on the principle of time and space matching; and obtaining clear sky labels, cloud labels, and dust labels corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data based on the correspondence.

[0011] In one embodiment of the first aspect, the cloud-aerosol lidar and infrared sounder satellite data products include vertical feature distribution data products and aerosol profile data products.

[0012] The second aspect of the present invention provides a method for identifying dust aerosols, comprising obtaining thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified; inputting the thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified into the dust aerosol identification model described in the first aspect; and obtaining the identification result of the dust aerosol in the area to be identified output by the dust aerosol identification model.

[0013] The third aspect of the present invention provides a dust aerosol identification system, comprising a first acquisition module for acquiring thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified; an identification module for inputting the thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified into the dust aerosol identification model described in the first aspect; and a second acquisition module for acquiring the identification result of the dust aerosol in the area to be identified output by the dust aerosol identification model.

[0014] The fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the training method for the dust aerosol recognition model described in the first aspect of the present invention and / or the dust aerosol recognition method described in the second aspect of the present invention are implemented.

[0015] The fifth aspect of the present invention provides an electronic device, comprising: a memory storing a computer program; a processor communicatively connected to the memory, and configured to execute the dust aerosol recognition model training method described in the first aspect of the present invention and / or the dust aerosol recognition method described in the second aspect of the present invention when calling the computer program.

[0016] As described above, the dust aerosol identification method, system, model training method, medium, and equipment provided by the embodiments of the present invention have the following beneficial effects: they can address the influence of complex climatic and geographical factors, eliminate the influence of clouds and the ground surface, extract atmospheric information in all-weather conditions, accurately identify rarefied dust aerosols, and achieve range detection of dust areas. Furthermore, the present invention does not require additional cloud detection preprocessing, reducing the number of steps in the dust identification process, significantly improving dust identification accuracy, and effectively distinguishing between clear sky, clouds, and dust. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 Shown is a flowchart of a method for training a dust aerosol recognition model in an embodiment of the present invention.

[0018] Figure 2 Shown is a flowchart of a method for training a dust aerosol recognition model in an embodiment of the present invention.

[0019] Figure 3 Shown is a schematic diagram comparing the daytime identification results of the dust aerosol identification model in an embodiment of the present invention and the CALIPSO vertical profile.

[0020] Figure 4 Shown is a schematic diagram comparing the daytime identification results of the dust aerosol identification model in an embodiment of the present invention and the CALIPSO vertical profile.

[0021] Figure 5 Shown is a schematic diagram comparing the nighttime identification results of the dust aerosol identification model in an embodiment of the present invention and the CALIPSO vertical profile.

[0022] Figure 6 Shown is a schematic diagram comparing the nighttime identification results of the dust aerosol identification model in an embodiment of the present invention and the CALIPSO vertical profile.

[0023] Figure 7 Shown is a schematic diagram of the daytime identification results of the sand and dust product and the daytime observation results of CALIPSO in an embodiment of the present invention.

[0024] Figure 8 Shown is a schematic diagram of the daytime recognition results of the dust aerosol recognition model in an embodiment of the present invention.

[0025] Figure 9 Shown is a schematic diagram of the nighttime identification results of the sand and dust product and the nighttime observation results of CALIPSO in an embodiment of the present invention.

[0026] Figure 10 Shown is a schematic diagram of the nighttime recognition results of the dust aerosol recognition model according to an embodiment of the present invention.

[0027] Figure 11 Shown is a schematic flow chart of a dust aerosol identification method according to an embodiment of the present invention.

[0028] Figure 12 Shown is a schematic structural diagram of a dust aerosol identification system according to an embodiment of the present invention.

[0029] Figure 13 Shown is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

[0030] Figure 14 Shown is a schematic diagram of observation data in an embodiment of the present invention.

[0031] Component number description

[0032] 20 First acquisition module

[0033] 30 Identification Module

[0034] 40 Second acquisition module

[0035] 90 Electronic devices

[0036] 901 Memory

[0037] 902 processor

[0038] 903 Display

[0039] Steps S1 to S3

[0040] Steps S21-S22

[0041] Steps S4 to S6 DETAILED DESCRIPTION

[0042] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless they conflict.

[0043] It should be noted that the diagrams provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. The diagrams only show components relevant to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be arbitrarily varied, and the component layout may also be more complex. Furthermore, in this document, relational terms such as "first," "second," and the like are used solely to distinguish one entity or operation from another, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0044] The present invention provides a dust aerosol identification method, system, model training method, medium, and equipment. These methods eliminate the influence of clouds and the ground surface, extract atmospheric information in all-weather conditions, accurately identify dilute dust aerosols, and achieve range detection of dust areas. Furthermore, the present invention eliminates the need for additional cloud detection preprocessing, reduces the number of steps in the dust identification process, significantly improves dust identification accuracy, and effectively distinguishes between clear sky, clouds, and dust.

[0045] Next, the dust aerosol identification method, system, model training method, medium and equipment provided by the present invention will be described through specific embodiments and drawings.

[0046] like Figure 1As shown, in one embodiment, the training method of the dust aerosol recognition model of the present invention includes steps S1 to S4:

[0047] S1: Obtain training thermal infrared observation brightness temperature data, training simulated clear sky brightness temperature data, and training surface information data.

[0048] Wherein, the training thermal infrared observation brightness temperature data is obtained based on satellite observation data.

[0049] Specifically, in one embodiment, the training thermal infrared observation brightness temperature data is obtained based on the brightness temperature observed by the 9th to 16th channels of the thermal infrared satellite of Japan, Sunflower-8.

[0050] The training simulated clear sky brightness temperature data and the surface information data are obtained based on satellite observation data, medium-resolution imaging spectrometer data and atmospheric reanalysis data.

[0051] Specifically, in one embodiment, the simulated clear-sky brightness temperature data is calculated using the Rapid Radiative Transfer Model (ERTM). This model uses atmospheric profiles (absolute humidity, temperature, and ozone fraction), surface temperature, and surface pressure data from the ERA-5 reanalysis data, observation angle data from the Japan Himawari-8 satellite (AHI), and surface emissivity product data from the Moderate Resolution Imaging Spectroradiometer (MODIS) product as input to simulate the AHI-observed brightness temperature under cloud-free conditions.

[0052] Specifically, the training surface information data is obtained based on medium-resolution imaging spectrometer data and atmospheric reanalysis data.

[0053] In one embodiment, surface pressure and surface temperature from ERA-5 reanalysis data and land cover types from MODIS products are used.

[0054] S2: Obtain clear sky labels, cloud labels, and dust labels corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data.

[0055] Specifically, the clear sky, cloud, and dust tags correspond to Cloud-Aerosol Lidar and Infrared Sounder Satellite (CALIPSO) data products. Among the CALIPSO secondary products, the vertical feature distribution data product can be used to determine cloud / aerosol types, and the aerosol profile data product can be used to determine aerosol optical depth.

[0056] Therefore, the clear sky label, cloud label, and dust label are mapped to the vertical feature distribution data product and the aerosol profile data product, respectively. The clear sky category is defined as cloud-free pixels with an aerosol optical depth of 0.1 or less, the cloud category is defined as cloud pixels with a cloud optical depth of 0.1 or less, and the dust category is defined as cloud-free pixels with an aerosol optical depth of 0.1 or greater.

[0057] Specifically, such as Figure 2 As shown, step S2 includes step S21 and step S22:

[0058] S21: establishing a correspondence between the satellite observation data, the moderate resolution imaging spectrometer data, the atmospheric reanalysis data, and the cloud-aerosol lidar and infrared sounder satellite data products based on a spatiotemporal matching principle.

[0059] Specifically, since the thermal infrared observation brightness temperature data, simulated clear-sky brightness temperature data, and surface information data are obtained in step S1 based on satellite observation data, moderate-resolution imaging spectrometer data, and atmospheric reanalysis data, and the three category labels are obtained based on CALIPSO level 2 products, in order to establish the relationship between the three category labels and the training thermal infrared observation brightness temperature data, training simulated clear-sky brightness temperature data, and training surface information data, step S21 first establishes the corresponding relationship between the satellite observation data, moderate-resolution imaging spectrometer data, atmospheric reanalysis data, and vertical characteristic distribution data products and aerosol profile data products based on the principle of spatiotemporal matching.

[0060] Specifically, the spatiotemporal matching principle is to establish a relationship between the closest time, the closest longitude, and the closest latitude. The following will further illustrate step S21 in an embodiment.

[0061] First, step S21 determines the aerosol type determined by the CALIPSO vertical characteristic distribution data product at a certain time, a certain longitude, and a certain latitude, and the optical thickness obtained by the aerosol profile data product, which corresponds to the clear sky label, cloud label, or sand and dust label.

[0062] Secondly, step 21 obtains the AHI 9-16 channel observation brightness temperature at the nearest time, nearest longitude, and nearest latitude, AHI observation angle data, surface emissivity product data and surface cover type product of MODIS products, atmospheric profile (absolute humidity, temperature, and ozone fraction), surface temperature, and surface pressure data of ERA-5 reanalysis data, thereby obtaining thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data, and surface information data.

[0063] When performing spatiotemporal matching in step S21, the temporal and spatial resolution of the observation data must be considered. When matching AHI observation data, MODIS data products, and ERA-5 data with tags, the time difference and position distance difference between their observation time and observation location and those of CALIPSO observations should be less than the minimum temporal and spatial resolution of the observation data. Figure 14 As shown,

[0064] Specifically, the AHI observations have a temporal resolution of 10 minutes and a spatial resolution of 2 km. When matching AHI observations with category labels, the difference between the AHI observation time and the CALIPSO observation time is less than 5 minutes, and the distance difference between the observation locations is less than 2 km in both longitude and latitude.

[0065] Specifically, the MODIS surface emissivity product used is an 8-day average product with a spatial resolution of 0.05° in both longitude and latitude. When matching MODIS surface emissivity data with category labels, the time difference between the observation time of the surface emissivity data and the CALIPSO observation time is less than eight days, and the distance difference between the observation locations is less than 0.05° in both longitude and latitude.

[0066] Specifically, the MODIS land cover classification product is annual data with a spatial resolution of 500 meters. When matching the MODIS land cover classification product with the category labels, the time difference between the land cover classification observation and the CALIPSO observation is less than one year, and the distance difference between the observation locations is less than 500 meters in both longitude and latitude.

[0067] Specifically, the ERA-5 reanalysis data has a temporal resolution of 1 hour, with atmospheric profile data having a spatial resolution of 0.25° in both longitude and latitude. The surface temperature and surface pressure data have a spatial resolution of 0.25° on the ocean surface and a higher spatial resolution of 0.1° on the land surface. When matching the ERA-5 reanalysis data with the category labels, the time difference between the observation time and the CALIPSO observation time is less than 30 minutes. The distance difference between the observation location of the atmospheric profile data is less than 0.25° in both longitude and latitude, the distance difference between the observation location of the surface temperature and surface pressure data on the ocean surface is less than 0.25° in both longitude and latitude, and the distance difference between the observation location of the surface temperature and surface pressure data on the land surface is less than 0.1° in both longitude and latitude.

[0068] S22: Based on the corresponding relationship, obtain a clear sky label, a cloud label, and a dust label corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data.

[0069] Specifically, step S21 establishes a correspondence between satellite observation data, medium-resolution imaging spectrometer data, atmospheric reanalysis data, vertical feature distribution data products, and aerosol profile data products. Therefore, step S22 establishes a correspondence between thermal infrared observation brightness temperature data, simulated clear-sky brightness temperature data, and surface information data at the nearest time, nearest longitude, and nearest latitude, and the three category labels.

[0070] S3: Training a dust aerosol recognition model based on the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, the training surface information data, and the clear sky label, cloud label, and dust label.

[0071] Specifically, in one embodiment, the period from 2017 to March-May 2020 was selected. Based on step S2, thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data, and surface information data were matched with clear sky labels, cloud labels, and dust labels to create a machine learning dataset. This dataset was then used to train a dust aerosol recognition model, establishing a relationship between input and labels. 10% of the data was used for validation, and model parameters were adjusted to achieve optimal performance. The model's performance was then evaluated on an independent test set (March 2021).

[0072] like Figure 3 、 4 As shown in Figures 5 and 6, the dust aerosol recognition model has a high accuracy in identifying dust aerosols on the test set. Figure 3 and Figure 4 The dust aerosol identification model is used to classify daytime cases on CALIPSO trajectories (March 23, 2021 and March 28, 2021). Figure 5 and Figure 6 Night case results are shown (the night of March 11 and March 16, 2021, respectively).

[0073] It can be seen that the dust aerosol recognition model constructed by the present invention performs equally well during the day and at night, can effectively distinguish between clouds and dust, and is also very effective in identifying thin dust.

[0074] In order to further verify the recognition accuracy of the dust aerosol recognition model, the model was further applied to the regional dust aerosol range detection. The results are as follows Figure 7 、 8 , 9, and 10.

[0075] in, Figure 7These are the daytime dust product and CALIPSO observation results on March 30, 2021. The yellow dotted line in the figure is the dust range identified by the dust product, and the white dots are the pixels where CALIPSO observed dust. Figure 8 This is the recognition result of the dust aerosol recognition model of the present invention during the daytime on March 30, 2021. Figure 9 These are the dust product and CALIPSO observation results for the night of March 15, 2021. The yellow dotted line range is the dust range identified by the dust product, and the white dots are the pixels where CALIPSO observed dust. Figure 10 This is the recognition result of the dust aerosol recognition model of the present invention on the night of March 15, 2021.

[0076] It can be seen that compared with the dust product, the dust aerosol identification model can more clearly display the dust identification results, clearly showing the range of clear sky, clouds, and dust. In addition, the dust area it detects is generally larger than the dust product and closer to the CALIPSO observation results. This is because the algorithm based on the brightness temperature of the thermal infrared channel used in the dust product is not sensitive to thin dust and cannot accurately identify dust aerosols with a small optical thickness. However, the dust aerosol identification algorithm proposed in this invention, due to the introduction of information such as clear sky brightness temperature, is more sensitive to thin dust and can better identify thinner dust, significantly improving the accuracy of dust identification.

[0077] It should be noted that strict quality control is also required when constructing the dataset and training the dust aerosol identification model. The quality control methods can be referred to existing methods and will not be elaborated here.

[0078] like Figure 11 As shown, the present invention also provides a method for identifying dust aerosols, comprising steps S4 to S6:

[0079] S4: Obtain thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified.

[0080] S5: Inputting the thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified into the dust aerosol identification model.

[0081] S6: Obtaining the dust aerosol recognition result in the area to be identified output by the dust aerosol recognition model.

[0082] Specifically, in one embodiment, thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data, surface pressure, surface temperature and surface cover type in the area to be identified can be obtained based on AHI satellite observation data, MODIS data products and ERA-5 reanalysis data, and this information can be input into the dust aerosol identification model, which will automatically identify dust aerosols in the area to be identified.

[0083] like Figure 12 As shown, the present invention further provides a dust aerosol identification system, which includes a first acquisition module 20 , an identification module 30 and a second acquisition module 40 .

[0084] Among them, the first acquisition module 20 is used to obtain thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified.

[0085] The identification module 30 is used to input the thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified into the dust aerosol identification model.

[0086] The second acquisition module 40 is used to obtain the dust aerosol recognition result in the area to be identified output by the dust aerosol recognition model.

[0087] Specifically, in one embodiment, the first acquisition module 20 obtains thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data, surface pressure, surface temperature and surface cover type in the area to be identified based on AHI satellite observation data, MODIS data products and ERA-5 reanalysis data. The identification module 30 inputs this information into the dust aerosol identification model, and the model will automatically identify the dust aerosol in the area to be identified. The second acquisition module 40 will obtain the identification result output by the model.

[0088] The present invention further provides a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program implements the dust aerosol identification model training method provided in the embodiments of the present invention and / or the dust aerosol identification method provided in the embodiments of the present invention.

[0089] In the present invention, any combination of one or more storage media may be used. The storage medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, RAM, ROM, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or device.

[0090] The present invention also provides an electronic device. Figure 13 The diagram shows the structure of an electronic device 90 according to an embodiment of the present invention. Figure 4 As shown, in this embodiment, the electronic device 90 includes a memory 901 and a processor 902 .

[0091] The memory 901 is used to store computer programs; preferably, the memory 901 includes: ROM, RAM, disk, USB flash drive, memory card or optical disk, etc., various media that can store program codes.

[0092] Specifically, the memory 901 may include a computer system readable medium in the form of a volatile memory, such as a random access memory (RAM) and / or a cache memory. The electronic device 90 may further include other removable / non-removable, volatile / non-volatile computer system storage media. The memory 901 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0093] The processor 902 is connected to the memory 901 and is used to execute the computer program stored in the memory 901 so that the electronic device 90 performs the training method of the dust aerosol recognition model provided in the embodiment of the present invention and / or implements the dust aerosol recognition method provided in the embodiment of the present invention.

[0094] Preferably, the processor 902 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0095] Preferably, the electronic device 90 in this embodiment may further include a display 903. The display 903 is communicatively connected to the memory 901 and the processor 902, and is used to display the training method of the dust aerosol identification model and / or a GUI interaction interface related to the dust aerosol identification method.

[0096] The scope of protection of the dust aerosol identification model training method and / or dust aerosol identification method described in the present invention is not limited to the order of execution of the steps listed in this embodiment. All solutions implemented by adding, subtracting, or replacing steps in the prior art based on the principles of the present invention are included in the scope of protection of the present invention.

[0097] In summary, the embodiments of the present invention provide a dust aerosol identification method, system, model training method, medium, and equipment that can address the impact of complex climatic and geographical factors, eliminate the influence of clouds and the ground surface, extract atmospheric information in all-weather conditions, accurately identify rarefied dust aerosols, and achieve range detection of dust areas. Furthermore, the present invention eliminates the need for additional cloud detection preprocessing, reduces the number of steps in the dust identification process, significantly improves dust identification accuracy, and effectively distinguishes between clear sky, clouds, and dust.

[0098] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. A training method for a dust aerosol recognition model, characterized in that: include: Obtain training thermal infrared observation brightness temperature data, training simulated clear sky brightness temperature data and training surface information data; Obtaining clear sky labels, cloud labels, and dust labels corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data; Training a dust aerosol recognition model based on the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, the training surface information data, and the clear sky label, cloud label, and dust label; in The methods for obtaining training thermal infrared observation brightness temperature data, training simulated clear sky brightness temperature data, and training surface information data include: Acquiring the training thermal infrared observation brightness temperature data based on satellite observation data; Acquire the training simulated clear sky brightness temperature data and the training surface information data based on satellite observation data, medium resolution imaging spectrometer data and atmospheric reanalysis data; The step of obtaining the training simulated clear sky brightness temperature data based on satellite observation data, medium resolution imaging spectrometer data, and atmospheric reanalysis data includes: Obtaining observation angle data based on satellite observation data; Obtain surface emissivity data based on Moderate Resolution Imaging Spectroradiometer data; Obtain atmospheric profile data, surface temperature and surface pressure data based on atmospheric reanalysis data; The training simulated clear sky brightness temperature data is obtained based on the atmospheric profile data, the surface temperature, the surface pressure data, the observation angle data and the surface emissivity data.

2. The dust aerosol identification model training method according to claim 1, characterized in that: The clear sky tag, cloud tag, and dust tag correspond to cloud-aerosol lidar and infrared sounder satellite data products.

3. The training method for the dust aerosol recognition model according to claim 2, characterized in that: The acquiring of clear sky labels, cloud labels, and dust labels corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data comprises: Establishing a correspondence between the satellite observation data, the moderate resolution imaging spectrometer data, the atmospheric reanalysis data, and the cloud-aerosol lidar and infrared sounder satellite data products based on a spatiotemporal matching principle; Based on the corresponding relationship, a clear sky label, a cloud label, and a dust label corresponding to the training thermal infrared observation brightness temperature data, the training simulated clear sky brightness temperature data, and the training surface information data are obtained.

4. The training method for the dust aerosol identification model according to claim 2, characterized in that: The cloud-aerosol lidar and infrared sounder satellite data products include vertical feature distribution data products and aerosol profile data products.

5. A method for identifying dust aerosols, characterized in that: include: Obtain thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified; Inputting the thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified into the dust aerosol identification model according to any one of claims 1 to 4; Obtain the dust aerosol identification result in the area to be identified output by the dust aerosol identification model.

6. A dust aerosol identification system, characterized in that: include: The first acquisition module is used to obtain thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data and surface information data in the area to be identified; An identification module, configured to input thermal infrared observation brightness temperature data, simulated clear sky brightness temperature data, and surface information data in the area to be identified into the dust aerosol identification model according to any one of claims 1 to 4; The second acquisition module is used to obtain the recognition result of the dust aerosol in the to-be-identified area output by the dust aerosol recognition model.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the method for training a dust aerosol recognition model according to any one of claims 1 to 4 or the method for identifying a dust aerosol according to claim 5.

8. An electronic device, characterized in that: The electronic device comprises: a memory storing a computer program; A processor is communicatively connected to the memory, and executes the dust aerosol identification model training method described in any one of claims 1 to 4 or the dust aerosol identification method described in claim 5 when calling the computer program.

Images