High-precision inversion method and system for cloud base height of domestic passive meteorological satellite

Abstract

The invention provides a domestic passive meteorological satellite cloud base height high-precision inversion method and system, and belongs to the field of atmospheric cloud parameter remote sensing inversion and quantitative remote sensing mechanisms. According to the method, re-analysis data such as satellite multi-channel atmospheric top reflectivity observation, brightness temperature observation and atmospheric temperature profile are fused, a complex nonlinear feature extraction and modeling method is developed, and the nonlinear mapping capability among the atmospheric top reflectivity, the brightness temperature and the cloud base height is improved; the night and all-day inversion capability is enhanced by combining all-time infrared channel characteristics and layer junction information; through a multi-source sample expansion and stack learning strategy, the generalization adaptability of the model is improved; and a sample expansion and auxiliary training mechanism is designed, so that the problem of insufficient high-quality training data is effectively relieved, and finally, high-precision inversion of the all-day cloud base height of a domestic static or polar orbit meteorological satellite is realized.

Description

Technical Field

[0001] This invention relates to the field of remote sensing inversion of atmospheric cloud parameters and quantitative remote sensing mechanisms, and in particular to a high-precision inversion method and system for cloud base height from a domestically produced passive meteorological satellite. Background Technology

[0002] Cloud base height (CBH) is an important physical parameter describing the vertical structure of clouds, and it is of great significance for weather monitoring, climate change assessment, radiation budget calculation, and aviation safety.

[0003] Currently, there are various methods for retrieving cloud base height using passive satellite data internationally. The commonly used schemes are mainly divided into three categories: (1) indirect schemes based on cloud thickness; (2) schemes based on the ratio of cloud water path to liquid water path (ice water path to liquid water path); and (3) machine learning schemes. Among these three main schemes for cloud base height retrieval, the National Meteorological Satellite Center's website does not provide publicly available, free parameters such as effective cloud path, effective cloud emissivity, cloud optical thickness, cloud water path, and cloud liquid water content. Therefore, schemes based on cloud thickness and cloud water path cannot be used for retrieving cloud base height from my country's domestically developed meteorological satellites. Thus, machine learning schemes are expected to become a universal scheme for retrieving cloud base height from my country's domestically developed meteorological satellites around the clock.

[0004] In existing technologies, cloud base height retrieval methods based on meteorological satellites have been widely applied in meteorological monitoring and climate research. Among them, cloud base height retrieval technology based on machine learning methods has gradually become a research hotspot in recent years due to its strong nonlinear fitting ability and the advantages of multi-source data fusion. However, this technology still has several shortcomings in practical applications: 1. The multi-channel brightness temperature information provided by satellite passive remote sensing mainly reflects the radiation characteristics of the cloud top and its upper layers, while cloud base height is controlled by complex factors such as microphysical processes within the cloud, atmospheric stratification structure, and radiation transmission. There is no direct linear or unique mapping relationship between brightness temperature information and cloud base height; the two are highly nonlinearly coupled. This makes it difficult for traditional machine learning models to fully characterize the complex nonlinear relationship between brightness temperature and cloud base height during training, easily leading to fitting bias or model instability; 2. Most existing machine learning models rely on visible light and shortwave infrared channel information during the day, while the relevant channels are missing during nighttime observations, resulting in a significant reduction in the dimensionality of input features under nighttime or all-day conditions, affecting the accuracy and applicability of cloud base height inversion, and failing to meet the needs of all-weather observation; 3. Existing machine learning models generally suffer from strong regional dependence on training samples and poor climate adaptability. Most models rely on training samples concentrated in specific regions, seasons, or cloud types, lacking comprehensive sample support covering a wide range of climate zones, time periods, and cloud types. This results in limited generalization ability and insufficient stability of the models under cross-regional, cross-seasonal, or complex weather system conditions. 4. Although active remote sensing satellites can provide high-precision cloud vertical profile data, providing a reference data source for machine learning model training, their spatial and temporal coverage is limited, making it difficult to form a long-term, continuous, large-sample, and balanced training dataset. Especially under complex cloud system conditions such as extreme weather and overlapping multi-layered clouds, the training samples are severely insufficient, restricting the sufficiency of machine learning model training and its application expansion capabilities. Summary of the Invention

[0005] To address the problems in the existing technologies, this invention provides a high-precision cloud base height retrieval method and system for domestically produced passive meteorological satellites. This invention integrates reanalysis data from satellite multi-channel atmospheric top reflectivity observations, brightness temperature observations, and atmospheric temperature profiles. It develops complex nonlinear feature extraction and modeling methods, improving the nonlinear mapping capability between atmospheric top reflectivity, brightness temperature, and cloud base height. By combining all-time infrared channel features and stratification information, it enhances nighttime and all-day retrieval capabilities. Through multi-source sample expansion and stacked learning strategies, it improves the model's generalization adaptability. Furthermore, it designs sample expansion and auxiliary training mechanisms to effectively alleviate the problem of insufficient high-quality training data, ultimately achieving high-precision all-day cloud base height retrieval applicable to domestically produced geostationary or polar-orbiting meteorological satellites. To achieve the above objectives, the technical solution is as follows: On the one hand, this invention provides a high-precision inversion method for cloud base height from domestically produced passive meteorological satellites, the method comprising: S1. Based on the existing first geostationary meteorological satellite dataset and active remote sensing satellite cloud profile dataset, a multi-source database set is obtained through time matching and spatial matching; S2. Based on the multi-source database set, by comparing it with the second geostationary meteorological satellite dataset and performing multi-channel spectral function response, the top of the atmosphere brightness temperature dataset is obtained. The top of the atmosphere brightness temperature dataset includes the top of the atmosphere brightness temperature dataset of the first geostationary meteorological satellite and the top of the atmosphere brightness temperature dataset of the second geostationary meteorological satellite. S3. Based on the atmospheric top brightness temperature dataset, the second geostationary meteorological satellite dataset is corrected to obtain the inversion dataset of the first geostationary meteorological satellite. S4. Merge the multi-source database set with the inversion dataset of the first geostationary meteorological satellite to obtain the all-day inversion dataset; S5. Based on the all-day inversion dataset, obtain the all-day cloud base height using a stacked machine learning method.

[0006] Optionally, the multi-source database set includes: L1 level multi-channel data and solar zenith angle data, L2 level cloud cover data and cloud mask data, cloud base height, temperature profile and relative humidity profile data, total atmospheric water vapor content and elevation data.

[0007] Optionally, in S2, based on this multi-source database set, a top atmospheric brightness temperature dataset is obtained by comparing its multi-channel spectral response function with that of a second geostationary meteorological satellite dataset, including: S21. Based on the multi-source database set and the second geostationary meteorological satellite dataset, the brightness temperature channels with large differences in spectral response functions are obtained by comparing the spectral response functions. S22. Based on the brightness temperature channels with large differences in the spectral response function, the brightness temperature dataset of the top of the atmosphere is obtained through MODTRAN radiative transfer simulation.

[0008] Optionally, in S3, the second geostationary meteorological satellite dataset is corrected based on the atmospheric top brightness temperature dataset to obtain the inversion dataset of the first geostationary meteorological satellite, including: S31. Based on the brightness temperature dataset of the top of the atmosphere, establish a linear relationship model between the brightness temperature channels of the first geostationary meteorological satellite and the second geostationary meteorological satellite; S32. Based on the linear relationship model between the brightness and temperature channels of the first geostationary meteorological satellite and the second geostationary meteorological satellite, the dataset of the second geostationary meteorological satellite is corrected to obtain the inversion dataset of the first geostationary meteorological satellite.

[0009] Optionally, in S5, based on the all-day inversion dataset, the all-day cloud base height is obtained through a stacked machine learning method, including: S51. Based on the all-day inversion dataset, filter the daytime inversion data and the nighttime inversion data to obtain the daytime inversion dataset and the nighttime inversion dataset; S52. Based on the daytime inversion dataset, multiple preliminary cloud base heights for each daytime are obtained using different machine learning models; S53. Based on the nighttime inversion dataset, preliminary cloud base heights for multiple nights are obtained using different machine learning models; S54. Based on the multiple preliminary daytime cloud base heights, the daytime cloud base height is obtained through a stacking model; S55. Based on the preliminary cloud base heights from multiple nights, the nighttime cloud base height is obtained by stacking models; S56. Based on the cloud base height during the day and the cloud base height at night, obtain the cloud base height for the entire day.

[0010] Optionally, the machine learning model includes: RF machine learning model, XGBOOST machine learning model and LightGBM machine learning model.

[0011] Optionally, the method for filtering daytime and nighttime inversion data includes: If the solar zenith angle in the all-day inversion dataset is less than 90 degrees, it is placed in the daytime inversion dataset; if the solar zenith angle in the all-day inversion dataset is not less than 90 degrees, it is placed in the nighttime inversion dataset.

[0012] On the other hand, this invention provides a high-precision cloud base height retrieval system for domestically produced passive meteorological satellites. This system is applied to a high-precision cloud base height retrieval method for domestically produced passive meteorological satellites. The system includes: The database set construction module is used to obtain a multi-source database set based on the existing first geostationary meteorological satellite dataset and active remote sensing satellite cloud profile dataset through time matching and spatial matching; The dataset comparison module is used to obtain the atmospheric top brightness temperature dataset by comparing it with the second geostationary meteorological satellite dataset and performing multi-channel spectral function response based on the multi-source database set. The atmospheric top brightness temperature dataset includes the atmospheric top brightness temperature dataset of the first geostationary meteorological satellite and the atmospheric top brightness temperature dataset of the second geostationary meteorological satellite. The dataset correction module is used to correct the second geostationary meteorological satellite dataset based on the atmospheric top brightness temperature dataset to obtain the inversion dataset of the first geostationary meteorological satellite. The inversion dataset construction module is used to merge the multi-source database set with the inversion dataset of the first geostationary meteorological satellite to obtain a full-time inversion dataset; The cloud base height inversion module is used to obtain the cloud base height for all time based on the all-time inversion dataset through a stacked machine learning method.

[0013] Compared with the prior art, the technical solution of the present invention has at least the following beneficial effects: The above-mentioned scheme, on the one hand, develops complex nonlinear feature extraction and modeling methods by integrating reanalysis data such as satellite multi-channel atmospheric top reflectivity observations, brightness temperature observations, and atmospheric temperature profiles, thereby improving the nonlinear mapping capability between atmospheric top reflectivity, brightness temperature, and cloud base height; on the other hand, it enhances the nighttime and all-day inversion capability by combining all-time infrared channel features and stratification information; and on the third hand, it improves the generalization adaptability of the model through multi-source sample expansion and stacked learning strategies, and designs sample expansion and auxiliary training mechanisms to effectively alleviate the problem of insufficient high-quality training data, ultimately achieving high-precision inversion of cloud base height for all-day conditions applicable to domestic geostationary or polar-orbiting meteorological satellites. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart of an embodiment of the domestically developed passive meteorological satellite cloud base height high-precision inversion method of the present invention; Figure 2 This is a flowchart illustrating the process of obtaining the atmospheric top brightness temperature dataset in an embodiment of the high-precision inversion method for cloud base height of domestically produced passive meteorological satellites of the present invention. Figure 3 This is a flowchart illustrating the high-precision inversion method for cloud base height of domestically produced passive meteorological satellites according to the present invention, which obtains the inversion dataset of the first geostationary meteorological satellite. Figure 4 This is a flowchart illustrating the method for obtaining the cloud base height under all weather conditions in an embodiment of the domestically developed passive meteorological satellite cloud base height high-precision inversion method of the present invention; Figure 5 The images shown are daytime cloud base height training results in an embodiment of the domestic passive meteorological satellite cloud base height high-precision inversion method of the present invention, wherein a) is a daytime preliminary cloud base height training result obtained by the RF machine learning model, b) is a daytime preliminary cloud base height training result obtained by the XGBOOST machine learning model, c) is a daytime preliminary cloud base height training result obtained by the LightGBM machine learning model, and d) is a daytime cloud base height training result obtained by the stacking model. Figure 6 The images shown are daytime cloud base height verification results in an embodiment of the domestic passive meteorological satellite cloud base height high-precision inversion method of the present invention. Among them, a) is the daytime preliminary cloud base height verification result obtained by the RF machine learning model, b) is the daytime preliminary cloud base height verification result obtained by the XGBOOST machine learning model, c) is the daytime preliminary cloud base height verification result obtained by the LightGBM machine learning model, and d) is the daytime cloud base height verification result obtained by the stacking model. Figure 7 The images shown are nighttime cloud base height training results in an embodiment of the domestic passive meteorological satellite cloud base height high-precision inversion method of the present invention, wherein a) is the nighttime preliminary cloud base height training result obtained by the RF machine learning model, b) is the nighttime preliminary cloud base height training result obtained by the XGBOOST machine learning model, c) is the nighttime preliminary cloud base height training result obtained by the LightGBM machine learning model, and d) is the nighttime cloud base height training result obtained by the stacking model. Figure 8 The images shown are nighttime cloud base height verification results in an embodiment of the domestic passive meteorological satellite cloud base height high-precision inversion method of the present invention, wherein a) is the nighttime preliminary cloud base height verification result obtained by the RF machine learning model, b) is the nighttime preliminary cloud base height verification result obtained by the XGBOOST machine learning model, c) is the nighttime preliminary cloud base height verification result obtained by the LightGBM machine learning model, and d) is the nighttime cloud base height verification result obtained by the stacking model. Figure 9 This is a training result diagram of cloud base height under all weather conditions in an embodiment of the high-precision inversion method for cloud base height of domestically produced passive meteorological satellites of the present invention; Figure 10 This is a diagram showing the all-day cloud base height verification results in an embodiment of the domestically developed passive meteorological satellite cloud base height high-precision inversion method of the present invention; Figure 11 This is a system block diagram of an embodiment of the domestically developed passive meteorological satellite cloud base height high-precision inversion system of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0017] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] like Figure 1 The flowchart shown is an embodiment of the high-precision cloud base height retrieval method for domestically produced passive meteorological satellites according to the present invention. The present invention provides a high-precision cloud base height retrieval method for domestically produced passive meteorological satellites, which is implemented by a high-precision cloud base height retrieval system for domestically produced passive meteorological satellites. The method includes: S1. Based on the existing first geostationary meteorological satellite dataset and active remote sensing satellite cloud profile dataset, a multi-source database set is obtained through time matching and spatial matching; Specifically, this multi-source database set includes: L1 level multi-channel data and solar zenith angle data, L2 level cloud cover data and cloud mask data, cloud base height, temperature profile and relative humidity profile data, total atmospheric water vapor content and elevation data.

[0020] Furthermore, in this embodiment, the available time of the FY-4B dataset from the first geostationary meteorological satellite and the CALIOP cloud profile data from the active remote sensing satellite is consistent, from June 1, 2022 to June 30, 2023. Cloud and sky pixels are identified using FY-4B cloud mask data. Through temporal and spatial matching of the FY-4B dataset and CALIOP, as well as spatial matching of the DEM, a multi-source database set is constructed, which includes FY-4B L1 multi-channel data, L2 cloud cover, cloud type and cloud mask, CALOP cloud base height, ERA5 temperature profile and relative humidity profile, elevation, etc.

[0021] S2. Based on the multi-source database set, by comparing it with the second geostationary meteorological satellite dataset and performing multi-channel spectral function response, the top of the atmosphere brightness temperature dataset is obtained. The top of the atmosphere brightness temperature dataset includes the top of the atmosphere brightness temperature dataset of the first geostationary meteorological satellite and the top of the atmosphere brightness temperature dataset of the second geostationary meteorological satellite. Specifically, such as Figure 2 The flowchart shown in this embodiment of the high-precision inversion method for cloud base height from domestically produced passive meteorological satellites of the present invention illustrates the process of obtaining the atmospheric top brightness temperature dataset. In step S2, based on this multi-source database set, the atmospheric top brightness temperature dataset is obtained by comparing its multi-channel spectral response function with that of a second geostationary meteorological satellite dataset. This dataset includes: S21. Based on the multi-source database set and the second geostationary meteorological satellite dataset, the brightness temperature channels with large differences in spectral response functions are obtained by comparing the spectral response functions. Furthermore, the differences between the multi-source database sets of the second geostationary meteorological satellite FY-4A and FY-4B were compared. The results showed that the spectral response functions of FY-4A and FY-4B had small differences in reflectance channels 1-6, while the spectral response functions of FY-4A in channels 7-14 and FY-4B in channels 7-15 (brightness temperature) had larger differences.

[0022] S22. Based on the brightness temperature channels with large differences in the spectral response function, the brightness temperature dataset of the top of the atmosphere is obtained through MODTRAN radiative transfer simulation.

[0023] Furthermore, in this embodiment, 2284 clear-sky profiles were selected from the atmospheric profiles. First, the top atmospheric brightness temperature datasets of channels FY-4A 7~14 and FY4B 7~15 were constructed using MODTRAN radiative transfer simulations.

[0024] S3. Based on the atmospheric top brightness temperature dataset, the second geostationary meteorological satellite dataset is corrected to obtain the inversion dataset of the first geostationary meteorological satellite. Specifically, such as Figure 3 The flowchart shown in this embodiment of the high-precision inversion method for cloud base height of domestically produced passive meteorological satellites of the present invention is used to obtain the inversion dataset of the first geostationary meteorological satellite. In step S3, the second geostationary meteorological satellite dataset is corrected based on the atmospheric top brightness temperature dataset to obtain the inversion dataset of the first geostationary meteorological satellite, including: S31. Based on the brightness temperature dataset of the top of the atmosphere, establish a linear relationship model between the brightness temperature channels of the first geostationary meteorological satellite and the second geostationary meteorological satellite; S32. Based on the linear relationship model between the brightness and temperature channels of the first geostationary meteorological satellite and the second geostationary meteorological satellite, the dataset of the second geostationary meteorological satellite is corrected to obtain the inversion dataset of the first geostationary meteorological satellite.

[0025] Furthermore, a linear relationship model between the brightness temperature channels of FY-4A and FY-4B was established, and the brightness temperature channels 7-14 of FY-4A were corrected to channels 7-15 of FY-4B. This enabled the previously constructed 2019-2020 FY-4A cloud base height inversion database to be converted into the dataset required for FY-4B cloud base height inversion, namely the inversion dataset of the first geostationary meteorological satellite.

[0026] S4. Merge the multi-source database set with the inversion dataset of the first geostationary meteorological satellite to obtain the all-day inversion dataset; S5. Based on the all-day inversion dataset, obtain the all-day cloud base height using a stacked machine learning method.

[0027] Specifically, such as Figure 4The flowchart shown in this embodiment of the high-precision cloud base height retrieval method for domestically produced passive meteorological satellites of the present invention illustrates the process of obtaining the all-day cloud base height. In step S5, based on the all-day inversion dataset, the all-day cloud base height is obtained through a stacked machine learning method, including: S51. Based on the all-day inversion dataset, filter the daytime inversion data and the nighttime inversion data to obtain the daytime inversion dataset and the nighttime inversion dataset; Furthermore, the method for filtering daytime and nighttime inversion data includes: If the solar zenith angle in the all-day inversion dataset is less than 90 degrees, it is placed in the daytime inversion dataset; if the solar zenith angle in the all-day inversion dataset is not less than 90 degrees, it is placed in the nighttime inversion dataset.

[0028] S52. Based on the daytime inversion dataset, multiple preliminary cloud base heights for each daytime are obtained using different machine learning models; Furthermore, the machine learning models, including the RF, XGBOOST, and LightGBM models, are used to estimate the daytime cloud base height of FY-4B. First, these models are used to estimate the daytime cloud base height of FY-4B. The 29 atmospheric temperature and relative humidity layers of ERA5, surface features (ERA5 LST and CWV, DEM, FY-4B cloud cover, longitude, and latitude), FY-4B satellite observation angle, reflectivity of channels 1-6, and brightness temperature of channels 7-15 are used as feature parameters, with CALIOP CBH as the label (true value). These parameters are then used to train the three machine learning models to obtain the estimated preliminary daytime cloud base height of FY-4B. Next, the preliminary daytime cloud base heights predicted by these three models are used to train a stacked model using LightGBM to obtain the daytime cloud base height. The training and validation results are as follows: Figure 5 The image shown is a daytime cloud base height training result image from an embodiment of the domestically developed passive meteorological satellite cloud base height high-precision inversion method of the present invention. Specifically, a) is a daytime preliminary cloud base height training result image obtained through the RF machine learning model; b) is a daytime preliminary cloud base height training result image obtained through the XGBOOST machine learning model; c) is a daytime preliminary cloud base height training result image obtained through the LightGBM machine learning model; and d) is a daytime cloud base height training result image obtained through a stacking model. Figure 6The following are daytime cloud base height verification results in an embodiment of the domestic passive meteorological satellite cloud base height high-precision inversion method of the present invention: a) is the daytime preliminary cloud base height verification result obtained by the RF machine learning model, b) is the daytime preliminary cloud base height verification result obtained by the XGBOOST machine learning model, c) is the daytime preliminary cloud base height verification result obtained by the LightGBM machine learning model, and d) is the daytime cloud base height verification result obtained by the stacking model.

[0029] S53. Based on the nighttime inversion dataset, preliminary cloud base heights for multiple nights are obtained using different machine learning models; Furthermore, the machine learning model includes: an RF machine learning model, an XGBOOST machine learning model, and a LightGBM machine learning model. In this embodiment, the model input data is basically the same as during the day, except for the FY-4B 1~6 reflectivity band. First, the RF machine learning model, XGBOOST machine learning model, and LightGBM machine learning model are used to estimate the preliminary cloud base height at night. Then, the LightGBM model is used to train a stacked model to estimate the final cloud base height at night. The training results and validation results are as follows. Figure 7 The image shown is an example of nighttime cloud base height training results in an embodiment of the domestically developed passive meteorological satellite cloud base height high-precision inversion method of the present invention. Specifically, a) is a preliminary nighttime cloud base height training result obtained through the RF machine learning model; b) is a preliminary nighttime cloud base height training result obtained through the XGBOOST machine learning model; c) is a preliminary nighttime cloud base height training result obtained through the LightGBM machine learning model; and d) is a nighttime cloud base height training result obtained through a stacking model. Figure 8 The nighttime cloud base height verification results shown in the embodiment of the domestic passive meteorological satellite cloud base height high-precision inversion method of the present invention are as follows: a) is the nighttime preliminary cloud base height verification result obtained by the RF machine learning model, b) is the nighttime preliminary cloud base height verification result obtained by the XGBOOST machine learning model, c) is the nighttime preliminary cloud base height verification result obtained by the LightGBM machine learning model, and d) is the nighttime cloud base height verification result obtained by the stacking model. S54. Based on the multiple preliminary daytime cloud base heights, the daytime cloud base height is obtained through a stacking model; S55. Based on the preliminary cloud base heights from multiple nights, the nighttime cloud base height is obtained by stacking models; S56. Based on the cloud base height during the day and the cloud base height at night, obtain the cloud base height for the entire day.

[0030] Furthermore, the training and validation results for cloud base height under all-day conditions are as follows: Figure 9The image shown is an example of the high-precision inversion method for cloud base height from domestically produced passive meteorological satellites according to the present invention, along with the training results of cloud base height training under all-day conditions. Figure 10 The figure shown is a diagram illustrating the all-day cloud base height verification results in an embodiment of the domestically developed passive meteorological satellite cloud base height high-precision inversion method of the present invention. like Figure 11 The diagram shown is a system block diagram of an embodiment of the domestically produced passive meteorological satellite cloud base height high-precision inversion system of the present invention. The present invention provides a domestically produced passive meteorological satellite cloud base height high-precision inversion system, which is applied to a domestically produced passive meteorological satellite cloud base height high-precision inversion method. The system includes: a database set construction module, a dataset comparison module, a dataset correction module, an inversion dataset construction module, and a cloud base height inversion module. Specifically, The database set construction module is used to obtain a multi-source database set based on the existing first geostationary meteorological satellite dataset and active remote sensing satellite cloud profile dataset through time matching and spatial matching; The dataset comparison module is used to obtain the atmospheric top brightness temperature dataset by comparing it with the second geostationary meteorological satellite dataset and performing multi-channel spectral function response based on the multi-source database set; The dataset correction module is used to correct the second geostationary meteorological satellite dataset based on the atmospheric top brightness temperature dataset to obtain the inversion dataset of the second geostationary meteorological satellite. The inversion dataset construction module is used to merge the multi-source database set with the inversion dataset of the second geostationary meteorological satellite to obtain a full-time inversion dataset; The cloud base height inversion module is used to obtain the cloud base height for all time based on the all-time inversion dataset through a stacked machine learning method.

[0031] This invention provides a high-precision inversion method and system for cloud base height from domestically produced passive meteorological satellites. By fusing reanalysis data from multi-channel atmospheric top reflectivity observations, brightness temperature observations, and atmospheric temperature profiles, this invention develops complex nonlinear feature extraction and modeling methods, enhancing the nonlinear mapping capability between atmospheric top reflectivity, brightness temperature, and cloud base height. Combining all-time infrared channel features and stratification information strengthens the inversion capability for both nighttime and all-day scenarios. Through multi-source sample expansion and stacked learning strategies, the generalization adaptability of the model is improved. Furthermore, a sample expansion and auxiliary training mechanism is designed to effectively alleviate the problem of insufficient high-quality training data, ultimately achieving high-precision inversion of cloud base height for all-day scenarios applicable to domestically produced geostationary or polar-orbiting meteorological satellites.

[0032] It is understood that the present invention has been described through the above embodiments and should not be construed as limiting the implementation and scope of the present invention. Those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the present invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. A high-precision inversion method for cloud base height from domestically produced passive meteorological satellites, characterized in that, The method includes: S1. Based on the existing first geostationary meteorological satellite dataset and active remote sensing satellite cloud profile dataset, a multi-source database set is obtained through time matching and spatial matching; S2. Based on the multi-source database set, by comparing it with the second geostationary meteorological satellite dataset and performing multi-channel spectral function response, an atmospheric top brightness temperature dataset is obtained. The atmospheric top brightness temperature dataset includes the atmospheric top brightness temperature dataset of the first geostationary meteorological satellite and the atmospheric top brightness temperature dataset of the second geostationary meteorological satellite. S3. Based on the atmospheric top brightness temperature dataset, the second geostationary meteorological satellite dataset is corrected to obtain the inversion dataset of the first geostationary meteorological satellite; S4. Merge the multi-source database set with the inversion dataset of the first geostationary meteorological satellite to obtain a full-day inversion dataset; S5. Based on the all-day inversion dataset, obtain the all-day cloud base height using a stacked machine learning method.

2. The high-precision inversion method for cloud base height from domestically produced passive meteorological satellites according to claim 1, characterized in that, The multi-source database set includes: L1 level multi-channel data and solar zenith angle data, L2 level cloud cover data and cloud mask data, cloud base height, temperature profile and relative humidity profile data, total atmospheric water vapor content and elevation data.

3. The high-precision inversion method for cloud base height from domestically produced passive meteorological satellites according to claim 1, characterized in that, In step S2, based on the multi-source database set, a multi-channel spectral response function comparison is performed with the second geostationary meteorological satellite dataset to obtain the atmospheric top brightness temperature dataset, including: S21. Based on the multi-source database set and the second geostationary meteorological satellite dataset, the brightness temperature channels with large differences in spectral response functions are obtained by comparing the spectral response functions. S22. Based on the brightness temperature channels with large differences in the spectral response function, the brightness temperature dataset of the top of the atmosphere is obtained through MODTRAN radiative transfer simulation.

4. The high-precision inversion method for cloud base height of domestically produced passive meteorological satellites according to claim 1, characterized in that, In step S3, the second geostationary meteorological satellite dataset is corrected based on the atmospheric top brightness temperature dataset to obtain the inversion dataset of the first geostationary meteorological satellite, including: S31. Based on the brightness temperature dataset of the top of the atmosphere, establish a linear relationship model between the brightness temperature channels of the first geostationary meteorological satellite and the second geostationary meteorological satellite; S32. Based on the linear relationship model between the brightness and temperature channels of the first geostationary meteorological satellite and the second geostationary meteorological satellite, the dataset of the second geostationary meteorological satellite is corrected to obtain the inversion dataset of the first geostationary meteorological satellite.

5. The high-precision inversion method for cloud base height from domestically produced passive meteorological satellites according to claim 1, characterized in that, In step S5, based on the all-day inversion dataset, the all-day cloud base height is obtained through a stacked machine learning method, including: S51. Based on the all-day inversion dataset, filter the daytime inversion data and the nighttime inversion data to obtain the daytime inversion dataset and the nighttime inversion dataset; S52. Based on the daytime inversion dataset, multiple preliminary daytime cloud base heights are obtained using different machine learning models; S53. Based on the nighttime inversion dataset, obtain the preliminary cloud base heights for multiple nights using different machine learning models; S54. Based on the multiple preliminary daytime cloud base heights, the daytime cloud base height is obtained through a stacking model; S55. Based on the preliminary cloud base heights of the multiple nights, the nighttime cloud base height is obtained through a stacking model; S56. Based on the daytime cloud base height and the nighttime cloud base height, obtain the all-day cloud base height.

6. The high-precision inversion method for cloud base height of domestically produced passive meteorological satellites according to claim 5, characterized in that, The machine learning models include: RF machine learning model, XGBOOST machine learning model and LightGBM machine learning model.

7. The high-precision inversion method for cloud base height of domestically produced passive meteorological satellites according to claim 5, characterized in that, The method for filtering daytime and nighttime inversion data includes: If the solar zenith angle in the all-day inversion dataset is less than 90 degrees, it is placed in the daytime inversion dataset; if the solar zenith angle in the all-day inversion dataset is not less than 90 degrees, it is placed in the nighttime inversion dataset.

8. A high-precision cloud base height retrieval system for domestically produced passive meteorological satellites, used to implement the high-precision cloud base height retrieval method for domestically produced passive meteorological satellites as described in any one of claims 1-7, characterized in that, The system includes: The database set construction module is used to obtain a multi-source database set based on the existing first geostationary meteorological satellite dataset and active remote sensing satellite cloud profile dataset through time matching and spatial matching; The dataset comparison module is used to obtain the atmospheric top brightness temperature dataset by comparing it with the second geostationary meteorological satellite dataset and performing multi-channel spectral function response based on the multi-source database set. The atmospheric top brightness temperature dataset includes the atmospheric top brightness temperature dataset of the first geostationary meteorological satellite and the atmospheric top brightness temperature dataset of the second geostationary meteorological satellite. The dataset correction module is used to correct the second geostationary meteorological satellite dataset based on the atmospheric top brightness temperature dataset to obtain the inversion dataset of the first geostationary meteorological satellite. The inversion dataset construction module is used to merge the multi-source database set with the inversion dataset of the first geostationary meteorological satellite to obtain a full-day inversion dataset; The cloud base height inversion module is used to obtain the all-day cloud base height based on the all-day inversion dataset through a stacked machine learning method.

Images