Atmospheric water vapor total amount reconstruction model training method, reconstruction and calibration method and device of MERSI imager on FY-3A satellite

Abstract

The invention provides an atmospheric water vapor total amount reconstruction model training method, reconstruction and calibration method and device for an MERSI imager on an FY-3A satellite, and the training method comprises the steps: taking multi-source observation data corresponding to a target railway in each historical unit time as sample data corresponding to each historical unit time; taking the PWV data of the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite as the label of each piece of sample data; and training the generalized regression neural network model by adopting the sample data with the labels to obtain an atmospheric water vapor total amount reconstruction model of the MERSI imager on the FY-3A satellite. The invention designs a training mode specially suitable for the atmospheric water vapor total amount reconstruction model of the MERSI imager on the FY-3A satellite, and the reliability of the reconstruction result of the atmospheric water vapor total amount PWV data of the MERSI imager on the FY-3A satellite can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of railway meteorological remote sensing technology, and in particular to a total atmospheric water vapor reconstruction model training method, reconstruction and calibration method and device for the MERSI imager on the FY-3A satellite. Background Art

[0002] In order to ensure the reliable operation of the railway, it is necessary to conduct various environmental monitoring of the railway area and its surrounding areas. The various environmental monitoring data include the total atmospheric water vapor PWV (Precipitable Water Vapor), which refers to the amount of precipitation that can be generated when all the water vapor contained in the atmospheric column per unit area from the ground to the top of the atmosphere condenses and falls to the ground. It is also called atmospheric precipitable water.

[0003] Specifically, the amount of precipitable water in the atmosphere is an important factor in meteorology and climatology. The spatiotemporal transport of water vapor is the main driving force of weather and climate change, and plays an important role in the study of climate change, hydrological cycle, energy budget, and so on at global and regional scales. Accurate measurement of water vapor can not only increase the understanding of the above fields, but also calibrate the wet delay of many geodetic observations. Traditional water vapor monitoring technologies such as water vapor radiometers (WVR) and radiosonde instruments have low spatiotemporal resolution of water vapor and cannot accurately grasp the changes in the spatiotemporal distribution of water vapor. However, in recent decades, the remarkable progress made in global navigation satellite system (GNSS) meteorology has made the global navigation satellite system (GNSS) a powerful means to observe water vapor with high temporal resolution.

[0004] At present, the China National Satellite Meteorological Center has inverted the atmospheric precipitable water product with a time resolution of 5 minutes for the 15-17th near-infrared channel of the MERSI imager (FY-3C medium-resolution spectroradiometer). However, due to the satellite orbit and inversion algorithm, the PWV data based on the MERSI imager on the FY-3A satellite has inherent defects, including regional data missing due to cloud cover, poor accuracy, and other problems, resulting in incomplete data coverage for areas such as railway areas. Therefore, it is urgent to design a feasible method to reconstruct the total atmospheric water vapor (FY-3A / MERSIPWV) of the MERSI imager on the FY-3A satellite. Summary of the invention

[0005] In view of this, the embodiments of the present application provide a total atmospheric water vapor reconstruction model training method, a reconstruction and calibration method and a device for the MERSI imager on the FY-3A satellite to eliminate or improve one or more defects in the prior art.

[0006] One aspect of the present application provides a method for training a total atmospheric water vapor reconstruction model of a MERSI imager on a FY-3A satellite, comprising:

[0007] The multi-source observation data corresponding to each historical unit time of the target railway are respectively used as the sample data corresponding to each historical unit time, and the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time are respectively used as the labels of the sample data corresponding to each historical unit time;

[0008] The preset generalized regression neural network model is trained using the sample data each provided with the label, so as to train the generalized regression neural network model into a total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, which is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

[0009] In some embodiments of the present application, the multi-source observation data includes: environmental meteorological data, normalized vegetation index data, digital elevation model data, and longitude and latitude information;

[0010] The environmental meteorological data include: temperature value, pressure value, specific humidity value, wind speed value, downward shortwave radiation value, downward longwave radiation value and precipitation.

[0011] In some embodiments of the present application, before taking the multi-source observation data corresponding to each historical unit time of the target railway as the sample data corresponding to each historical unit time, the method further includes:

[0012] Collecting the environmental meteorological data, the normalized vegetation index data, the digital elevation model data and the longitude and latitude information corresponding to each historical unit time of the target railway;

[0013] And, obtaining the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time of the target railway;

[0014] The environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data corresponding to each historical unit time are preprocessed respectively to obtain the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data corresponding to each historical unit time and meeting the preset time resolution and the preset first spatial resolution.

[0015] In some embodiments of the present application, before taking the multi-source observation data corresponding to each historical unit time of the target railway as the sample data corresponding to each historical unit time, the method further includes:

[0016] Based on the NDVI time series harmonic analysis method, the normalized difference vegetation index data satisfying the preset time resolution and the preset first spatial resolution corresponding to each historical unit time is reconstructed in time series.

[0017] In some embodiments of the present application, the multi-source observation data corresponding to each historical unit time of the target railway are used as sample data corresponding to each historical unit time, and the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time are used as labels of the sample data corresponding to each historical unit time, including:

[0018] Sampling the preprocessed environmental meteorological data, the normalized difference vegetation index data, the digital elevation model data, the longitude and latitude information, and the total atmospheric water vapor PWV data to a preset second spatial resolution, wherein the preset second spatial resolution is greater than the preset first spatial resolution;

[0019] Grid point spatiotemporal matching is performed on the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data that have been sampled to the second spatial resolution to select sample data and labels corresponding to each historical unit time, wherein the sample data includes the environmental meteorological data, the normalized vegetation index data, the digital elevation model data and the longitude and latitude information that have been sampled to the second spatial resolution corresponding to a historical unit time, and the label includes: the total atmospheric water vapor PWV data corresponding to a historical unit time.

[0020] In some embodiments of the present application, the preset generalized regression neural network model is trained using the sample data with each label to train the generalized regression neural network model into a reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite for outputting the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time, including:

[0021] The sample data with the labels are used to perform multiple rounds of iterative training on the preset generalized regression neural network model, and in each round of iterative training, the ten-fold verification method is used to obtain the optimal standard deviation of the trained generalized regression neural network model in the current iterative training round, and the generalized regression neural network model is trained in the next iterative training round based on the optimal standard deviation until the model converges or a preset number of iterative training rounds are completed, and the generalized regression neural network model is determined as the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, wherein the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

[0022] In some embodiments of the present application, the input layer of the generalized regression neural network model includes: neurons corresponding to the temperature value, the pressure value, the specific humidity value, the wind speed value, the downlink shortwave radiation value, the downlink longwave radiation value, the precipitation, the normalized vegetation index data, the digital elevation model data, and the longitude, latitude, altitude and time in the longitude and latitude information, respectively, and each of the neurons in the input layer is connected to each of the neurons in the pattern layer of the generalized regression neural network model, wherein the number of neurons in the pattern layer is the same as that in the input layer;

[0023] The output layer of the generalized regression neural network model includes a neuron for outputting the reconstruction result of the total atmospheric water vapor PWV data.

[0024] Another aspect of the present application provides a method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, comprising:

[0025] Acquire multi-source observation data of the target railway corresponding to a target unit time and satisfying a preset time resolution and a preset first spatial resolution;

[0026] Input the multi-source observation data satisfying the preset time resolution and the preset first spatial resolution corresponding to the target unit time into the atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, so that the atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite outputs the reconstruction result of the atmospheric water vapor PWV data corresponding to the target railway at the target unit time, wherein the atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite is pre-trained based on the atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite;

[0027] Based on the pre-acquired total atmospheric water vapor PWV value of the global navigation satellite system corresponding to the target railway at the target unit time, the reconstruction result of the total atmospheric water vapor PWV data is calibrated to obtain the calibrated total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to the target railway at the target unit time.

[0028] The third aspect of the present application provides a training device for the atmospheric water vapor total reconstruction model of the MERSI imager on the FY-3A satellite, comprising:

[0029] A sample setting module is used to use the multi-source observation data corresponding to each historical unit time of the target railway as the sample data corresponding to each historical unit time, and use the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time as the label of the sample data corresponding to each historical unit time;

[0030] The model training module is used to train a preset generalized regression neural network model using the sample data each provided with the label, so as to train the generalized regression neural network model into a total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, which is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

[0031] The fourth aspect of the present application provides a reconstruction and calibration device for the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, comprising:

[0032] A data acquisition module, used to acquire multi-source observation data of the target railway corresponding to a target unit time and satisfying a preset time resolution and a preset first spatial resolution;

[0033] A model reconstruction module is used to input the multi-source observation data satisfying a preset time resolution and a preset first spatial resolution corresponding to the target unit time into a reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite, so that the reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite outputs a reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at the target unit time, wherein the reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite is pre-trained based on a training method for the reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite;

[0034] The calibration module is used to calibrate the reconstruction result of the total atmospheric water vapor PWV data based on the pre-acquired total atmospheric water vapor PWV value of the global navigation satellite system corresponding to the target railway at the target unit time, so as to obtain the calibrated total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to the target railway at the target unit time.

[0035] The fifth aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for reconstructing a model of the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, and / or implements a method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite.

[0036] The sixth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements a method for reconstructing a model of the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, and / or implements a method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite.

[0037] The seventh aspect of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements a method for training a model for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, and / or implements a method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite.

[0038] The present application provides a method for reconstructing a model of total atmospheric water vapor of the MERSI imager on the FY-3A satellite. The method comprises the following steps: using the multi-source observation data corresponding to each historical unit time of the target railway as the sample data corresponding to each historical unit time, and using the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time as the label of the sample data corresponding to each historical unit time; using the sample data with the labels to train a preset generalized regression neural network model, so as to train the generalized regression neural network model into a generalized regression neural network model for reconstructing a model of total atmospheric water vapor of the target railway according to the multi-source observation data corresponding to each historical unit time of the target railway. The observation data corresponds to the reconstruction result of the total atmospheric water vapor PWV data of the target railway corresponding to the target unit time of the MERSI imager on the FY-3A satellite. A training method specifically suitable for the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite can be designed, which can effectively improve the training efficiency, reduce the calculation cost and improve the robustness of the model, thereby realizing the intelligent and automatic reconstruction of the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite, and can effectively improve the reliability of the reconstruction results of the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite.

[0039] Additional advantages, purposes, and features of the present application will be partially described in the following description, and will become partially apparent to those skilled in the art after studying the following, or may be learned from the practice of the present application. The purposes and other advantages of the present application can be achieved and obtained by the structures specifically pointed out in the specification and the drawings.

[0040] Those skilled in the art will understand that the purposes and advantages that can be achieved by the present application are not limited to the above specific description, and the above and other purposes that can be achieved by the present application will be more clearly understood based on the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings described herein are used to provide a further understanding of the present application, constitute a part of the present application, and do not constitute a limitation of the present application. The components in the drawings are not drawn to scale, but are only for the purpose of illustrating the principles of the present application. In order to facilitate the illustration and description of some parts of the present application, the corresponding parts in the drawings may be enlarged, that is, they may become larger relative to other components in the exemplary device actually manufactured according to the present application. In the drawings:

[0042] Figure 1 This is a schematic diagram of a first flow chart of a method for training a total atmospheric water vapor reconstruction model for the MERSI imager on the FY-3A satellite in one embodiment of the present application.

[0043] Figure 2 Schematic diagram of a second flow chart of a method for training a total atmospheric water vapor reconstruction model for the MERSI imager on the FY-3A satellite in an embodiment of the present application.

[0044] Figure 3 This is a third flow chart of a method for training a total atmospheric water vapor reconstruction model for the MERSI imager on the FY-3A satellite in one embodiment of the present application.

[0045] Figure 4 Schematic diagram of a flow chart of a method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite in one embodiment of the present application.

[0046] Figure 5 This is a flow chart of a method for reconstructing and calibrating the PWV of the MERSI imager (MERSI / FY-3A) on the FY-3A satellite based on multi-source observation data and a generalized regression neural network (GRNN) model in an application example of the present application.

[0047] Figure 6 Schematic diagram of the structure of a total atmospheric water vapor reconstruction model training device for the MERSI imager on the FY-3A satellite in one embodiment of the present application.

[0048] Figure 7 Schematic diagram of the structure of a device for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite in one embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the implementation modes and the accompanying drawings. Here, the illustrative implementation modes and descriptions of the present application are used to explain the present application, but are not intended to limit the present application.

[0050] It should also be noted here that in order to avoid obscuring the present application due to unnecessary details, only the structures and / or processing steps closely related to the scheme according to the present application are shown in the accompanying drawings, while other details that are not very relevant to the present application are omitted.

[0051] It should be emphasized that the term “include / comprises” when used herein refers to the presence of features, elements, steps or components, but does not exclude the presence or addition of one or more other features, elements, steps or components.

[0052] It should also be noted that, unless otherwise specified, the term “connection” herein may refer not only to a direct connection but also to an indirect connection involving an intermediate.

[0053] Hereinafter, embodiments of the present application will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0054] In order to design a feasible method for reconstructing the total atmospheric water vapor amount (FY-3A / MERSIPWV) of the MERSI imager on the FY-3A satellite, the embodiments of the present application respectively provide a total atmospheric water vapor amount reconstruction model training method for the MERSI imager on the FY-3A satellite, a total atmospheric water vapor amount reconstruction and calibration method for the MERSI imager on the FY-3A satellite, a total atmospheric water vapor amount reconstruction model training device for the MERSI imager on the FY-3A satellite for executing the total atmospheric water vapor amount reconstruction model training method for the MERSI imager on the FY-3A satellite, a total atmospheric water vapor amount reconstruction and calibration device for the MERSI imager on the FY-3A satellite for executing the total atmospheric water vapor amount reconstruction and calibration method for the MERSI imager on the FY-3A satellite, an electronic device, a computer-readable storage medium, and a computer program product.

[0055] The details are described in detail through the following examples.

[0056] Based on this, the embodiment of the present application provides a method for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, which can be implemented by the training device for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite. Figure 1 The training method of the atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite specifically includes the following contents:

[0057] Step 100: The multi-source observation data corresponding to the target railway at each historical unit time are respectively used as the sample data corresponding to each historical unit time, and the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time are respectively used as the labels of the sample data corresponding to each historical unit time.

[0058] In one or more embodiments of the present application, the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time used in the training process refers to the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite with complete data and no missing data.

[0059] It can be understood that the historical unit time refers to a historical unit time, and the length of the unit time can be selected according to actual application requirements, for example, it can be set to several minutes, one hour, several hours, one day, and several days.

[0060] In one or more embodiments of the present application, the target railway may refer to an area covered by a railway line or an area covered by a section of a railway line. The area may include the railway area and the surrounding area. The specific area range may be set according to the actual application scenario.

[0061] Step 200: A preset generalized regression neural network model is trained using the sample data each provided with the label, so as to train the generalized regression neural network model into a total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite for outputting a reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

[0062] In step 200, the generalized regression neural network model refers to the GRNN model. The generalized regression neural network (GRNN) is based on the maximum likelihood algorithm and is a special form of the radial basis function neural network. The model has the advantages of relatively simple structure, fast training speed, low computational cost and global convergence. Compared with the commonly used back-propagation neural network, it can overcome the shortcomings of slow convergence and easy to fall into local minima.

[0063] Specifically, the GRNN model includes an input layer, a pattern layer, a summation layer, and an output side connected in sequence, where:

[0064] (1) Input layer: The input layer is the first layer of the GRNN network, responsible for receiving input data and passing the data to the pattern layer. The number of nodes in the input layer is equal to the number of features of the input data. Each node corresponds to a feature in the input data. The nodes in the input layer do not perform any processing and only pass the original input data to the pattern layer.

[0065] (2) Pattern layer: The pattern layer is the second layer of the GRNN network. The number of its nodes is usually equal to the number of samples in the training data set. Each neuron node in the pattern layer represents a sample in the training data set and takes the distance between the input data and the sample as input. The pattern layer calculates the Euclidean distance between the input vector and each sample and uses the distance as the input of the next summation layer.

[0066] (3) Summation layer: The summation layer is the third layer of the GRNN network. The function of the summation layer is to perform weighted calculation on the distance transmitted from the pattern layer and calculate the weighted average. The weight calculation is achieved by transforming the distance using a Gaussian function. The input of the summation layer is the distance vector transmitted from the pattern layer. Each distance vector is transformed using a Gaussian function and the transformed value is used as the weight. Finally, all weights are weighted averaged with the output of the pattern layer to obtain the output value of the GRNN.

[0067] (4) Output layer: The output layer is the last layer of the GRNN network. The number of its nodes is equal to the dimension of the output data. Each node in the output layer corresponds to a dimension of the output data and is responsible for converting and outputting the output value of the summation layer.

[0068] From the above description, it can be seen that the training method for the reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite provided in the embodiment of the present application can design a training method specifically suitable for the reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite, which can effectively improve the training efficiency, reduce the computational cost and improve the robustness of the model, thereby realizing the intelligent and automatic reconstruction of the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite, and can effectively improve the reliability of the reconstruction results of the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite.

[0069] In order to further improve the effectiveness and reliability of the results of the reconstruction of the total amount of atmospheric water vapor by the MERSI imager on the FY-3A satellite, in a method for reconstructing a model of the total amount of atmospheric water vapor by the MERSI imager on the FY-3A satellite provided in an embodiment of the present application, the multi-source observation data may specifically include: environmental meteorological data (i.e., meteorological data of the area surrounding the railway), normalized difference vegetation index (NDVI) data, digital elevation model (DEM) data, and longitude and latitude information. The longitude and latitude information may include latitude (Latitude) and longitude (Longitude), and may also include altitude and time, etc.

[0070] It can be understood that the environmental meteorological data can specifically include: temperature (TEM), pressure (PRES), specific humidity (SHUM), wind speed (WIND), downward shortwave radiation (SRAD), downward longwave radiation (LRAD) and precipitation (PREC), all of which can be obtained from the China Meteorological Forcing Dataset (CMFD).

[0071] In order to further improve the effectiveness and reliability of the total atmospheric water vapor reconstruction model training of the MERSI imager on the FY-3A satellite, in a method for reconstructing the total atmospheric water vapor of the MERSI imager on the FY-3A satellite provided in an embodiment of the present application, see Figure 2 The training method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite further specifically includes the following contents before step 100:

[0072] Step 010: Collect the environmental meteorological data, the normalized vegetation index data, the digital elevation model data and the longitude and latitude information corresponding to each historical unit time of the target railway.

[0073] In step 010, the digital elevation model (DEM) data may be provided by the Shuttle Radar Topography Mission (SRTM), and the original resolution of the product is 30m. The MOD13Q1 Normalized Difference Vegetation Index (NDVI) data provided by the Moderate-resolution Imaging Spectroradiometer (MODIS) may also be used, and the temporal and spatial resolutions thereof may be 250m and 16 days (16 days).

[0074] And, step 020: obtaining the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time of the target railway.

[0075] Specifically, according to the scope of the area surrounding the railway, the 5-min PWV data of land water vapor from the MERSI imager (MERSI / FY-3A) on the FY-3A satellite can be obtained from the Fengyun satellite remote sensing data service network. The spatial resolution of the data is 1 km.

[0076] Step 030: Preprocess the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data corresponding to each historical unit time respectively to obtain the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data corresponding to each historical unit time and meeting the preset time resolution and the preset first spatial resolution.

[0077] In order to further improve the application effectiveness and reliability of normalized vegetation index data, in a method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite provided in an embodiment of the present application, see Figure 3 The method for reconstructing the total atmospheric water vapor volume model of the MERSI imager on the FY-3A satellite further specifically includes the following contents after step 010 and before step 100:

[0078] Step 040: Based on the NDVI time series harmonic analysis method, the normalized vegetation index data corresponding to each historical unit time and meeting the preset time resolution and the preset first spatial resolution are reconstructed in time series.

[0079] In order to further improve the effectiveness and robustness of the total atmospheric water vapor reconstruction model training of the MERSI imager on the FY-3A satellite, in a method for reconstructing the total atmospheric water vapor of the MERSI imager on the FY-3A satellite provided in an embodiment of the present application, see Figure 3 Step 100 of the method for training the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite specifically includes the following contents:

[0080] Step 110: Sample the preprocessed environmental meteorological data, the normalized difference vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data to a preset second spatial resolution, wherein the preset second spatial resolution is greater than the preset first spatial resolution.

[0081] The preset first spatial resolution may be 0.01°, and the preset second spatial resolution may be 0.25°.

[0082] Step 120: Perform grid point spatiotemporal matching on the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information, and the total atmospheric water vapor PWV data that have been sampled to the second spatial resolution, so as to select sample data and labels corresponding to each historical unit time, wherein the sample data includes the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, and the longitude and latitude information that have been sampled to the second spatial resolution corresponding to a historical unit time, and the label includes: the total atmospheric water vapor PWV data corresponding to a historical unit time.

[0083] In order to further improve the effectiveness and robustness of the total atmospheric water vapor reconstruction model training of the MERSI imager on the FY-3A satellite, in a method for reconstructing the total atmospheric water vapor of the MERSI imager on the FY-3A satellite provided in an embodiment of the present application, see Figure 3 Step 200 of the method for training the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite specifically includes the following contents:

[0084] Step 210: Perform multiple rounds of iterative training on the preset generalized regression neural network model using the sample data each provided with the label, and in each round of iterative training, respectively use the ten-fold verification method to obtain the optimal standard deviation of the trained generalized regression neural network model in the current iterative training round, and train the generalized regression neural network model in the next iterative training round based on the optimal standard deviation until the model converges or a preset number of iterative training rounds are completed, and the generalized regression neural network model is determined as the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, wherein the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

[0085] In order to further improve the effectiveness and robustness of the total atmospheric water vapor reconstruction model training of the MERSI imager on the FY-3A satellite, in a method for reconstructing the total atmospheric water vapor of the MERSI imager on the FY-3A satellite provided in an embodiment of the present application, the input layer of the generalized regression neural network model includes: neurons corresponding to the temperature value, the pressure value, the specific humidity value, the wind speed value, the downlink shortwave radiation value, the downlink longwave radiation value, the precipitation, the normalized vegetation index data, the digital elevation model data, and the longitude, latitude, altitude and time in the longitude and latitude information, respectively, and each of the neurons in the input layer is connected to each of the neurons in the pattern layer of the generalized regression neural network model, wherein the number of neurons in the pattern layer is the same as that in the input layer;

[0086] The output layer of the generalized regression neural network model includes a neuron for outputting the reconstruction result of the total atmospheric water vapor PWV data.

[0087] Based on the above-mentioned embodiment of the training method for the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, the present application also provides an embodiment of a method for reconstructing and calibrating the total atmospheric water vapor of the MERSI imager on the FY-3A satellite, see Figure 4 The reconstruction and calibration method of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite specifically includes the following contents:

[0088] Step 300: Acquire multi-source observation data of the target railway corresponding to a target unit time and satisfying a preset time resolution and a preset first spatial resolution.

[0089] Step 400: Input the multi-source observation data satisfying the preset time resolution and the preset first spatial resolution corresponding to the target unit time into the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, so that the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite outputs the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at the target unit time, wherein the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite is pre-trained based on the total atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite.

[0090] It can be understood that the training method for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite in step 500 can be specifically implemented by adopting the embodiment of the training method for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite.

[0091] Step 500: Calibrate the reconstruction result of the total atmospheric water vapor PWV data based on the pre-acquired total atmospheric water vapor PWV value of the global navigation satellite system corresponding to the target railway at the target unit time, so as to obtain the calibrated total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to the target railway at the target unit time.

[0092] From the above description, it can be seen that the reconstruction and calibration method of the total atmospheric water vapor volume PWV data of the MERSI imager on the FY-3A satellite provided in the embodiment of the present application can effectively improve the reliability and effectiveness of the reconstruction and calibration results of the total atmospheric water vapor volume PWV data of the MERSI imager on the FY-3A satellite.

[0093] To further illustrate the above embodiment, the present application also provides a method for reconstructing and calibrating the PWV of the MERSI imager (MERSI / FY-3A) on the FY-3A satellite based on multi-source observation data and a generalized regression neural network (GRNN) model, see Figure 5 , the method comprises the following steps performed in sequence:

[0094] (1) Step 1: Obtain meteorological data, normalized difference vegetation index (NDVI), digital elevation model (DEM) data, latitude, longitude, and total atmospheric water vapor from the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV) for the area surrounding a railway, and preprocess the data to obtain multi-source observation variable information with a preset temporal resolution such as day temporal resolution and a preset spatial resolution such as 0.01°.

[0095] In the step 1, the digital elevation model (DEM) data may be data provided by the Shuttle Radar Topography Mission (SRTM), and the original resolution of the product is 30 m.

[0096] The MOD13Q1 Normalized Difference Vegetation Index (NDVI) data provided by MODIS can also be used, with a temporal and spatial resolution of 250 meters and 16 days. The Fourier-Transform method–Harmonic Analysis of Time Series (HANTS) can then be used to reconstruct the NDVI time series on a daily time scale, and the gaps in the NDVI can be filled.

[0097] Meteorological data may include temperature (TEM), pressure (PRES), specific humidity (SHUM), wind speed (WIND), downward shortwave radiation (SRAD), downward longwave radiation (LRAD) and precipitation (PREC), all of which can be obtained from the China Meteorological Forcing Dataset (CMFD).

[0098] In step 1, according to the scope of the area around the railway, the 5-min PWV data of land water vapor from the MERSI imager (MERSI / FY-3A) on the FY-3A satellite can be obtained from the Fengyun satellite remote sensing data service network. The spatial resolution of the data is 1 km. The total atmospheric water vapor PWV can also be calculated based on the pre-acquired tropospheric static delay value, tropospheric zenith delay value and atmospheric weighted average temperature value corresponding to each historical time period of the target railway. The calculation method of the total atmospheric water vapor PWV data can be:

[0099] The Zenith Hydrostatic Delay (ZHD) can generally be calculated using empirical models, such as the Saastamoninen model. The Zenith Wet Delay (ZWD) can be obtained by subtracting the ZHD from the Zenith Hydrostatic Delay (ZTD). The total atmospheric water vapor PWV can be further obtained using the conversion formula:

[0100]

[0101] Where, T m is the weighted average temperature of the atmosphere, which can be calculated using an empirical model based on temperature, k'2, k3 and R v Are physical constants, whose values ​​are 16.25K / hPa, 3.776×105K2 / hPa and 461.495J / K / kg respectively.

[0102] (2) Step 2: All data are sampled to a larger spatial resolution, such as 0.25°, and then grid point temporal and spatial matching is performed. Sample data are selected to train the generalized regression neural network (GRNN) model.

[0103] Specifically, in an example of step 2, if the number of grid point samples involved in modeling is 1325, then 202104 groups of sample data can be finally selected to train the generalized regression neural network (GRNN) model.

[0104] (3) Step 3: The standard deviation σ of the generalized regression neural network (GRNN) model is determined by the ten-fold verification method. The generalized regression neural network (GRNN) model trained with the optimal σ is used as the final reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV).

[0105] The method for determining the standard deviation σ in step three is as follows: setting the σ value range to 0.01 to 0.3 with an interval of 0.01 to repeatedly train the generalized regression neural network (GRNN) model, and using the mean absolute error RMS of 10-fold CV (i.e., cross validation fold cv=10) as the standard to determine the optimal σ.

[0106] (4) Step 4: Using daily and 0.01° resolution meteorological data, normalized difference vegetation index (NDVI), digital elevation model (DEM) data and other variables as input data, the optimal generalized regression neural network (GRNN) model was used to reconstruct the missing values ​​of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV).

[0107] The specific method of step four is:

[0108] The input layer consists of 12 neurons: latitude, longitude, altitude, time, normalized difference vegetation index (NDVI) and 7 meteorological variables together constitute the observation vector X i =[Latitude Longitude DEM DOY{Met.}NDVI] T ; The output layer is a neuron: Y i =[FY-3A / MERSIPWV], 1≤i≤n. The number of nodes in the model layer is the same as that of the observation vector. The purpose of constructing this generalized regression neural network (GRNN) model is to establish a functional mapping relationship f between input data and output data. After obtaining f, given X = [Latitude Longitude DEM DOY{Met.}NDVI] T , the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite under the mapping relationship (FY-3A / MERSIPWV) can be obtained.

[0109] (5) Step 5: Use the PWV value of the global navigation satellite system (GNSS) to calibrate the reconstructed PWV value of the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV).

[0110] The calibration method of the total atmospheric water vapor product (FY-3A / MERSIPWV) of the MERSI imager on the FY-3A satellite in step 5 is:

[0111]

[0112] Where, 1≤i≤365, refers to the total atmospheric water vapor of the MERSI imager on the FY-3A satellite after calibration on the i-th day (FY-3A / MERSIPWV), is the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite on the i-th day (FY-3A / MERSIPWV), ε i is the system deviation on the i-th day, a and b are the calibration coefficients, which are the fixed error and proportional error, respectively, and are obtained by estimation.

[0113] That is to say, the application example of this application provides a method for reconstructing and calibrating the PWV of the MERSI imager (MERSI / FY-3A) on the FY-3A satellite based on multi-source observation data and a generalized regression neural network (GRNN) model, which involves the field of railway meteorological remote sensing. The water vapor product coverage of the medium-resolution spectral imager (MERSI) of China's second-generation polar-orbiting satellite FengYun-3A (FY-3A) is incomplete, data is missing, and the accuracy is low in the surrounding area of ​​a railway. The method comprises the following steps: step 1: obtaining meteorological data, normalized difference vegetation index (NDVI), digital elevation model (DEM) data, latitude, longitude and the total atmospheric water vapor of the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV) in the surrounding area of ​​a railway, and preprocessing the data to obtain variable information with a time resolution of days and a spatial resolution of 0.01°; step 2: sampling all data to 0.25°, then performing spatiotemporal matching of grid points, and selecting sample data for training a generalized regression neural network (GRNN) model; step 3: using a ten-fold verification method to determine the generalized regression The key parameter σ of the neural network (GRNN) model is used. The GRNN model trained with the optimal σ is used as the final reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV); Step 4: Using daily and 0.01° resolution meteorological data, normalized difference vegetation index (NDVI), digital elevation model (DEM) data and other variables as input data, the optimal generalized regression neural network (GRNN) model is used to reconstruct the missing values ​​of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV); Step 5: The total atmospheric water vapor of the MERSI imager on the FY-3A satellite (FY-3A / MERSIPWV) reconstructed using GNSS PWV calibration. The performance of the total atmospheric water vapor (FY-3A / MERSIPWV) reconstruction model of the MERSI imager on the FY-3A satellite in this application is stable and reliable, and can provide reliable PWV data for the area around a railway, which is of great value for the hydrological and ecological research in the region.

[0114] Therefore, this application example introduces multi-source earth observation data to construct a spatiotemporally continuous reconstruction and calibration method for the total atmospheric water vapor (FY-3A / MERSIPWV) of the MERSI imager on the FY-3A satellite, so as to fill in the missing values ​​of the total atmospheric water vapor (FY-3A / MERSIPWV) product of the MERSI imager on the FY-3A satellite and calibrate it.

[0115] The technical problem to be solved by the application example of this application is a method for reconstructing and calibrating the PWV of the MERSI imager (MERSI / FY-3A) on the FY-3A satellite based on multi-source observation data and a generalized regression neural network (GRNN) model, so as to fill in the missing values ​​of the total atmospheric water vapor product (FY-3A / MERSIPWV) of the MERSI imager on the FY-3A satellite and improve its accuracy.

[0116] Compared with the prior art, the application example of this application has the following beneficial effects:

[0117] 1. This application uses a generalized regression neural network (GRNN) to establish a reconstruction model of the total atmospheric water vapor (FY-3A / MERSIPWV) of the MERSI imager on the FY-3A satellite. The generalized regression neural network (GRNN) is based on the maximum likelihood algorithm and is a special form of the radial basis function neural network. The model has the advantages of relatively simple structure, fast training speed, low computational cost and global convergence. Compared with the commonly used back-propagation neural network, it can overcome the shortcomings of slow convergence and easy to fall into local minima.

[0118] 2. This application uses a generalized regression neural network (GRNN) to establish a reconstruction model for the total atmospheric water vapor (FY-3A / MERSIPWV) of the MERSI imager on the FY-3A satellite. The model has strong robustness and the reconstruction results are highly reliable. This method is of reference value for other regions.

[0119] The present application also provides a training device for the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite for executing all or part of the training method for the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, see Figure 6 The atmospheric water vapor total amount reconstruction model training device of the MERSI imager on the FY-3A satellite specifically includes the following contents:

[0120] The sample setting module 10 is used to use the multi-source observation data corresponding to the target railway at each historical unit time as the sample data corresponding to each historical unit time, and use the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time as the label of the sample data corresponding to each historical unit time.

[0121] The model training module 20 is used to train a preset generalized regression neural network model using the sample data each provided with the label, so as to train the generalized regression neural network model into a total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, which is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

[0122] The embodiment of the device for training a model for reconstructing the total amount of atmospheric water vapor for the MERSI imager on the FY-3A satellite provided in the present application can be specifically used to execute the processing flow of the embodiment of the training method for reconstructing the total amount of atmospheric water vapor for the MERSI imager on the FY-3A satellite in the above-mentioned embodiment. Its functions are not described in detail here, and reference can be made to the detailed description of the embodiment of the training method for reconstructing the total amount of atmospheric water vapor for the MERSI imager on the FY-3A satellite.

[0123] The part of the atmospheric water vapor reconstruction model training device for the MERSI imager on the FY-3A satellite that performs the atmospheric water vapor reconstruction model training for the MERSI imager on the FY-3A satellite can be completed in the server or the client device. It can be selected specifically according to the processing capability of the client device and the limitations of the user's usage scenario. This application is not limited to this. If all operations are completed in the client device, the client device may also include a processor for specific processing of the atmospheric water vapor reconstruction model training for the MERSI imager on the FY-3A satellite.

[0124] The client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side, and other implementation scenarios may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, or a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0125] The server and the client device may communicate with each other using any suitable network protocol, including network protocols that have not yet been developed on the date of filing this application. The network protocols may include, for example, TCP / IP, UDP / IP, HTTP, HTTPS, etc. Of course, the network protocols may also include, for example, RPC (Remote Procedure Call Protocol) and REST (Representational State Transfer) protocols used on top of the above protocols.

[0126] From the above description, it can be seen that the training device for reconstructing the total atmospheric water vapor volume PWV data of the MERSI imager on the FY-3A satellite provided in the embodiment of the present application can design a training method specifically suitable for the total atmospheric water vapor volume reconstruction model of the MERSI imager on the FY-3A satellite, which can effectively improve the training efficiency, reduce the computing cost and improve the robustness of the model, thereby realizing the intelligent and automatic reconstruction of the total atmospheric water vapor volume PWV data of the MERSI imager on the FY-3A satellite, and can effectively improve the reliability of the reconstruction results of the total atmospheric water vapor volume PWV data of the MERSI imager on the FY-3A satellite.

[0127] The present application also provides a device for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, which is used to perform all or part of the method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, see Figure 7 The reconstruction and calibration device of the total atmospheric water vapor amount of the MERSI imager on the FY-3A satellite specifically includes the following contents:

[0128] The data acquisition module 30 is used to acquire multi-source observation data of the target railway corresponding to a target unit time and satisfying a preset time resolution and a preset first spatial resolution.

[0129] The model reconstruction module 40 is used to input the multi-source observation data that meets the preset time resolution and the preset first spatial resolution corresponding to the target unit time into the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, so that the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite outputs the reconstruction result of the total atmospheric water vapor PWV data of the target railway corresponding to the target unit time, wherein the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite is pre-trained based on the total atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite.

[0130] The calibration module 50 is used to calibrate the reconstruction result of the total atmospheric water vapor PWV data based on the pre-acquired total atmospheric water vapor PWV value of the global navigation satellite system corresponding to the target railway at the target unit time, so as to obtain the calibrated total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to the target railway at the target unit time.

[0131] The embodiment of the device for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite provided in the present application can be specifically used to execute the processing flow of the embodiment of the method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite in the above-mentioned embodiment. Its functions are not repeated here, and reference can be made to the detailed description of the embodiment of the method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite.

[0132] The embodiment of the present application also provides an electronic device, which may include a processor, a memory, a receiver and a transmitter, wherein the processor is used to execute the atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite mentioned in the above embodiment and / or the atmospheric water vapor reconstruction and calibration method of the MERSI imager on the FY-3A satellite, wherein the processor and the memory may be connected via a bus or other means, with the bus connection being taken as an example. The receiver may be connected to the processor and the memory via a wired or wireless manner.

[0133] The processor may be a central processing unit (CPU). The processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0134] As a non-transient computer-readable storage medium, the memory can be used to store non-transient software programs, non-transient computer executable programs and modules, such as the program instructions / modules corresponding to the method for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite and / or the method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite in the embodiments of the present application. The processor executes various functional applications and data processing of the processor by running the non-transient software programs, instructions and modules stored in the memory, that is, implementing the method for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite and / or the method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite in the above method embodiments.

[0135] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created by the processor, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0136] The one or more modules are stored in the memory, and when executed by the processor, perform the total atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite and / or the reconstruction and calibration method of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite in the embodiment.

[0137] In some embodiments of the present application, the user equipment may include a processor, a memory, and a transceiver unit, which may include a receiver and a transmitter. The processor, memory, receiver, and transmitter may be connected through a bus system. The memory is used to store computer instructions, and the processor is used to execute the computer instructions stored in the memory to control the transceiver unit to send and receive signals.

[0138] As an implementation method, the functions of the receiver and the transmitter in the present application can be considered to be implemented through a transceiver circuit or a dedicated chip for transceiver, and the processor can be considered to be implemented through a dedicated processing chip, a processing circuit or a general chip.

[0139] As another implementation method, it is possible to use a general-purpose computer to implement the server provided in the embodiment of the present application, that is, to store the program code for implementing the functions of the processor, receiver, and transmitter in a memory, and the general-purpose processor implements the functions of the processor, receiver, and transmitter by executing the code in the memory.

[0140] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite and / or the reconstruction and calibration method of the atmospheric water vapor total amount of the MERSI imager on the FY-3A satellite are implemented. The computer-readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the technical field.

[0141] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements a method for training a model for reconstructing the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, and / or implements a method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite.

[0142] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is specifically performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.

[0143] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.

[0144] In the present application, features described and / or illustrated for one embodiment may be used in the same manner or in a similar manner in one or more other embodiments, and / or combined with features of other embodiments or replace features of other embodiments.

[0145] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the embodiments of the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for reconstructing a model of total atmospheric water vapor from the MERSI imager on the FY-3A satellite, characterized in that: include: The multi-source observation data corresponding to each historical unit time of the target railway are respectively used as the sample data corresponding to each historical unit time, and the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time are respectively used as the labels of the sample data corresponding to each historical unit time; The preset generalized regression neural network model is trained using the sample data each provided with the label, so as to train the generalized regression neural network model into a total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, which is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

2. The method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite according to claim 1, characterized in that: The multi-source observation data includes: environmental meteorological data, normalized vegetation index data, digital elevation model data, and longitude and latitude information; The environmental meteorological data include: temperature value, pressure value, specific humidity value, wind speed value, downward shortwave radiation value, downward longwave radiation value and precipitation.

3. The method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite according to claim 2, characterized in that: Before taking the multi-source observation data corresponding to each historical unit time of the target railway as the sample data corresponding to each historical unit time, the method further includes: Collecting the environmental meteorological data, the normalized vegetation index data, the digital elevation model data and the longitude and latitude information corresponding to each historical unit time of the target railway; And, obtaining the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time of the target railway; The environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data corresponding to each historical unit time are preprocessed respectively to obtain the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data corresponding to each historical unit time and meeting the preset time resolution and the preset first spatial resolution.

4. The method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite according to claim 3, characterized in that: Before taking the multi-source observation data corresponding to each historical unit time of the target railway as the sample data corresponding to each historical unit time, the method further includes: Based on the NDVI time series harmonic analysis method, the normalized difference vegetation index data satisfying the preset time resolution and the preset first spatial resolution corresponding to each historical unit time is reconstructed in time series.

5. The method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite according to claim 3 or 4, characterized in that: The multi-source observation data corresponding to each historical unit time of the target railway are used as sample data corresponding to each historical unit time, and the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time are used as labels of the sample data corresponding to each historical unit time, including: Sampling the preprocessed environmental meteorological data, the normalized difference vegetation index data, the digital elevation model data, the longitude and latitude information, and the total atmospheric water vapor PWV data to a preset second spatial resolution, wherein the preset second spatial resolution is greater than the preset first spatial resolution; Grid point spatiotemporal matching is performed on the environmental meteorological data, the normalized vegetation index data, the digital elevation model data, the longitude and latitude information and the total atmospheric water vapor PWV data that have been sampled to the second spatial resolution to select sample data and labels corresponding to each historical unit time, wherein the sample data includes the environmental meteorological data, the normalized vegetation index data, the digital elevation model data and the longitude and latitude information that have been sampled to the second spatial resolution corresponding to a historical unit time, and the label includes: the total atmospheric water vapor PWV data corresponding to a historical unit time.

6. The method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite according to claim 1, characterized in that: The method uses the sample data with the labels to train a preset generalized regression neural network model, so as to train the generalized regression neural network model into a reconstruction model of the total atmospheric water vapor of the MERSI imager on the FY-3A satellite, which is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time, including: The sample data with the labels are used to perform multiple rounds of iterative training on the preset generalized regression neural network model, and in each round of iterative training, the ten-fold verification method is used to obtain the optimal standard deviation of the trained generalized regression neural network model in the current iterative training round, and the generalized regression neural network model is trained in the next iterative training round based on the optimal standard deviation until the model converges or a preset number of iterative training rounds are completed, and the generalized regression neural network model is determined as the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, wherein the total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

7. The method for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite according to claim 2, characterized in that: The input layer of the generalized regression neural network model includes: neurons corresponding to the temperature value, the pressure value, the specific humidity value, the wind speed value, the downward shortwave radiation value, the downward longwave radiation value, the precipitation, the normalized vegetation index data, the digital elevation model data, and the longitude, latitude, altitude and time in the longitude and latitude information, respectively, and each of the neurons in the input layer is connected to each of the neurons in the pattern layer of the generalized regression neural network model, wherein the number of neurons in the pattern layer is the same as that in the input layer; The output layer of the generalized regression neural network model includes a neuron for outputting the reconstruction result of the total atmospheric water vapor PWV data.

8. A method for reconstructing and calibrating the total amount of atmospheric water vapor of the MERSI imager on the FY-3A satellite, characterized in that: include: Acquire multi-source observation data of the target railway corresponding to a target unit time and satisfying a preset time resolution and a preset first spatial resolution; Input the multi-source observation data satisfying the preset time resolution and the preset first spatial resolution corresponding to the target unit time into the atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, so that the atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite outputs the reconstruction result of the atmospheric water vapor PWV data corresponding to the target railway at the target unit time, wherein the atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite is pre-trained based on the atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite according to any one of claims 1 to 7; Based on the pre-acquired total atmospheric water vapor PWV value of the global navigation satellite system corresponding to the target railway at the target unit time, the reconstruction result of the total atmospheric water vapor PWV data is calibrated to obtain the calibrated total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to the target railway at the target unit time.

9. A training device for reconstructing the total atmospheric water vapor volume of the MERSI imager on the FY-3A satellite, characterized in that: include: A sample setting module is used to use the multi-source observation data corresponding to each historical unit time of the target railway as the sample data corresponding to each historical unit time, and use the total atmospheric water vapor PWV data of the MERSI imager on the FY-3A satellite corresponding to each historical unit time as the label of the sample data corresponding to each historical unit time; The model training module is used to train a preset generalized regression neural network model using the sample data each provided with the label, so as to train the generalized regression neural network model into a total atmospheric water vapor reconstruction model of the MERSI imager on the FY-3A satellite, which is used to output the reconstruction result of the total atmospheric water vapor PWV data corresponding to the target railway at a target unit time according to the multi-source observation data corresponding to the target railway at the target unit time.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the total atmospheric water vapor reconstruction model training method of the MERSI imager on the FY-3A satellite as described in any one of claims 1 to 7, and / or implements the total atmospheric water vapor reconstruction and calibration method of the MERSI imager on the FY-3A satellite as described in claim 8.

Images