Dual correction method and system for observation deviation of satellite-borne microwave detector and medium

Abstract

The embodiment of the invention discloses a double correction method and system for observation deviation of a satellite-borne microwave detector and a medium, and the method comprises the steps: introducing surface parameters on the basis of the first correction of the observation deviation of the satellite-borne microwave detector, and accurately describing the correction residual error of a near-surface detection channel and a window region channel of the satellite-borne microwave detector; the double correction of the observation deviation of the satellite-borne microwave detector is realized, so that the observation brightness temperature of the corrected satellite-borne microwave detector is closer to the simulated brightness temperature of the satellite-borne microwave detector. According to the method, the adverse effect of complex observation deviation characteristics of the near-surface detection channel and the window area detection channel of the satellite-borne microwave detector on the observation deviation correction effect of the satellite-borne microwave detector is overcome, and the observation deviation correction effect of the near-surface detection channel and the window area detection channel of the satellite-borne microwave detector is effectively improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of microwave remote sensing, and in particular, to a dual correction method, system and medium for the observation deviation of a spaceborne microwave sounder. Background Art

[0002] A spaceborne microwave sounder is an important instrument for obtaining global atmospheric temperature profiles, humidity profiles, cloud parameters, and surface parameters, and plays an important role in many fields such as atmospheric environmental monitoring, numerical weather forecasting, and climate change research. The observed brightness temperature of the spaceborne microwave sounder can be converted into atmospheric parameter information using an inversion system, and can also be directly assimilated in a numerical weather prediction assimilation system. The physical variational algorithm is the core algorithm of the inversion system and the assimilation system, and the requirement of this algorithm for the observed brightness temperature of the spaceborne microwave sounder input into it is that the observation deviation of the spaceborne microwave sounder needs to satisfy unbiased and Gaussian characteristics as much as possible. When the observation deviation of the spaceborne microwave sounder is smaller, the accuracy of the atmospheric parameter inversion from the observed brightness temperature of the spaceborne microwave sounder will be higher, or the assimilation application effect of the observed brightness temperature of the spaceborne microwave sounder will be better. Therefore, before the inversion or assimilation application of the observed brightness temperature of the spaceborne microwave sounder, it is necessary to correct the observation deviation of the spaceborne microwave sounder. The observation deviation of the spaceborne microwave sounder is caused by multiple error sources, such as inaccurate calibration of microwave sounder data, adverse observation environments, spectral data errors of radiation transfer models, inaccuracies of radiation transfer models, and atmospheric parameter errors involved in radiation transfer calculations, etc. The complexity of the sources of the observation deviation of the spaceborne microwave sounder makes it very difficult to determine the error sources and model corrections from a physical perspective. When the spaceborne microwave sounder detects the atmosphere, its high-altitude detection channels only detect the distribution characteristics of atmospheric parameters at high altitudes, and the nonlinearity of the observation deviation characteristics of the high-altitude detection channels is relatively simple; its near-surface detection channels or window channels not only detect the atmospheric parameter information at low altitudes / near the surface, but also detect the surface parameter information, and the microwave radiation of atmospheric parameters and surface parameters is usually mixed with the surface reflected radiation, which further leads to a more complex nonlinear relationship in the observation deviation characteristics of the near-surface detection channels and window channels. When the existing technology uses a neural network to correct the observation deviation of the spaceborne microwave sounder, the correction effect of the observation deviation of the high-altitude detection channels of the spaceborne microwave sounder is significant, but the correction effect of the observation deviation of the near-surface detection channels and window detection channels is poor.

[0003] The above problems need to be solved urgently. Summary of the Invention

[0004] To solve the related technical problems, the present invention provides a dual correction method, system and medium for the observation deviation of a spaceborne microwave sounder to solve the problems mentioned in the above background art section.

[0005] To achieve the above object, the embodiments of the present invention adopt the following technical solutions:

[0006] In a first aspect, an embodiment of the present invention provides a dual correction method for the observation deviation of a spaceborne microwave sounder, including:

[0007] Establish a first correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0008] Based on the first correction data set, obtain the first corrected brightness temperature of the spaceborne microwave sounder;

[0009] Establish a second correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0010] Based on the second correction data set, obtain the second corrected brightness temperature of the spaceborne microwave sounder;

[0011] Compare the first corrected brightness temperature with the second corrected brightness temperature, and complete the dual correction of the observation deviation of the spaceborne microwave sounder according to the comparison result.

[0012] As an optional implementation manner, the establishment of the first correction data set for correcting the observation deviation of the spaceborne microwave sounder includes:

[0013] Calculate the observation deviation of the spaceborne microwave sounder, and establish a first correction data set for correcting the observation deviation of the spaceborne microwave sounder.

[0014] As an optional implementation manner, the calculation of the observation deviation of the spaceborne microwave sounder includes:

[0015] Establish a matching data set of atmospheric parameters and the observed brightness temperature of the spaceborne microwave sounder in time and space, and calculate the observation deviation of the spaceborne microwave sounder.

[0016] As an optional implementation manner, the obtaining of the first corrected brightness temperature of the spaceborne microwave sounder based on the first correction data set includes:

[0017] Based on the first correction data set, establish a first correction model for the observation deviation of the spaceborne microwave sounder, and obtain the first corrected brightness temperature of the spaceborne microwave sounder through the first correction model.

[0018] As an optional implementation manner, the establishment of the second correction data set for correcting the observation deviation of the spaceborne microwave sounder includes:

[0019] Calculate the correction residual of the spaceborne microwave sounder, and based on the correction residual of the spaceborne microwave sounder, establish a second correction data set for correcting the observation deviation of the spaceborne microwave sounder.

[0020] As an alternative implementation, after establishing the second correction data set for correcting the observation deviation of the spaceborne microwave sounder, the following steps are further included:

[0021] Based on the second correction data set, establish a second correction model for the observation deviation of the spaceborne microwave sounder.

[0022] As an alternative implementation, the obtaining of the second corrected brightness temperature of the spaceborne microwave sounder based on the second correction data set specifically includes:

[0023] Based on the second correction data set, establish a second correction model for the observation deviation of the spaceborne microwave sounder, and obtain the second corrected brightness temperature of the spaceborne microwave sounder according to the second correction model.

[0024] As an alternative implementation, the completing of the dual correction of the observation deviation of the spaceborne microwave sounder according to the comparison result includes:

[0025] According to the comparison result, obtain the corrected brightness temperature with the highest accuracy, and complete the dual correction of the observation deviation of the spaceborne microwave sounder.

[0026] In a second aspect, an embodiment of the present invention provides a dual correction system for the observation deviation of a spaceborne microwave sounder. This system adopts the dual correction method for the observation deviation of a spaceborne microwave sounder proposed in the first aspect above, and includes:

[0027] A first correction data set establishment module, configured to establish a first correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0028] A first corrected brightness temperature acquisition module, configured to obtain the first corrected brightness temperature of the spaceborne microwave sounder based on the first correction data set;

[0029] A second correction data set establishment module, configured to establish a second correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0030] A second corrected brightness temperature acquisition module, configured to obtain the second corrected brightness temperature of the spaceborne microwave sounder based on the second correction data set;

[0031] A dual correction module, configured to compare the first corrected brightness temperature with the second corrected brightness temperature, and complete the dual correction of the observation deviation of the spaceborne microwave sounder according to the comparison result.

[0032] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium. Computer-executable instructions are stored in the computer-readable storage medium. When the computer-executable instructions are executed by a processor, they are used to implement the dual correction method for the observation deviation of a spaceborne microwave sounder provided in the first aspect embodiment above.

[0033] The technical solution proposed in the embodiment of the present invention overcomes the adverse effects of the complex observation deviation characteristics of the near-surface detection channel and the window area detection channel of the spaceborne microwave sounder on the observation deviation correction effect of the spaceborne microwave sounder. By introducing surface parameters on the basis of the first correction of the observation deviation of the spaceborne microwave sounder, the correction residuals of the near-surface detection channel and the window area channel of the spaceborne microwave sounder are accurately described, realizing the double correction of the observation deviation of the spaceborne microwave sounder, and making the observed brightness temperature of the corrected spaceborne microwave sounder closer to the simulated brightness temperature of the spaceborne microwave sounder. The technical solution proposed in the embodiment of the present invention effectively improves the observation deviation correction effect of the near-surface detection channel and the window area detection channel of the spaceborne microwave sounder, is simple and easy to operate, and is suitable for popularization and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate and understand the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the background technology and the description of the embodiments of the present invention. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the content of the embodiments of the present invention and these drawings without creative efforts.

[0035] Figure 1 It is a schematic flowchart of the double correction method for the observation deviation of the spaceborne microwave sounder provided in Embodiment 1 of the present invention;

[0036] Figure 2 It is a display diagram of the root mean square error between the first correction brightness temperature and the observed brightness temperature of the MWTS-II on FY-3E satellite provided in Embodiment 1 of the present invention and the simulated brightness temperature of the MWTS-II respectively;

[0037] Figure 3 It is a display diagram of the comparison of the root mean square error between the second correction brightness temperature and the first correction brightness temperature of the MWTS-II on FY-3E satellite provided in Embodiment 1 of the present invention and the simulated brightness temperature of the MWTS-II respectively;

[0038] Figure 4 It is a schematic diagram of the principle of the double correction system for the observation deviation of the spaceborne microwave sounder provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] In order to make the technical problems solved by the present invention, the technical solutions adopted and the achieved technical effects clearer, the following will further describe the technical solutions in the embodiments of the present invention in detail with reference to the drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0040] Example 1

[0041] Please refer to Figure 1 as described Figure 1 which is a schematic flow diagram of the dual correction method for the observation deviation of the spaceborne microwave sounder provided in Example 1 of the present invention.

[0042] As shown in the figure, the dual correction method 100 for the observation deviation of the spaceborne microwave sounder in this embodiment includes:

[0043] S101. Establish a first correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0044] S102. Obtain the first corrected brightness temperature of the spaceborne microwave sounder based on the first correction data set;

[0045] S103. Establish a second correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0046] S104. Obtain the second corrected brightness temperature of the spaceborne microwave sounder based on the second correction data set;

[0047] S105. Compare the first corrected brightness temperature with the second corrected brightness temperature, and complete the dual correction of the observation deviation of the spaceborne microwave sounder according to the comparison result.

[0048] Exemplarily, the establishment of the first correction data set for correcting the observation deviation of the spaceborne microwave sounder includes:

[0049] Calculate the observation deviation of the spaceborne microwave sounder, and establish a first correction data set for correcting the observation deviation of the spaceborne microwave sounder.

[0050] Exemplarily, the calculation of the observation deviation of the spaceborne microwave sounder includes:

[0051] Establish a data set of atmospheric parameters and a matching data set of the observed brightness temperature of the spaceborne microwave sounder in time and space, and calculate the observation deviation of the spaceborne microwave sounder.

[0052] Specifically, in this embodiment, atmospheric parameter datasets are established by selecting atmospheric parameters such as temperature profiles, humidity profiles, cloud water profiles, surface temperature, surface humidity, surface pressure, and 10m wind speed from historical climate datasets. The atmospheric parameter datasets are matched with the observed brightness temperatures of spaceborne microwave sounders in terms of time and space to establish a matching dataset. The atmospheric parameters and the observation angles of the observed brightness temperatures of spaceborne microwave sounders in the matching dataset are input into a radiative transfer model to calculate the simulated brightness temperatures of spaceborne microwave sounders. Based on the simulated brightness temperatures and the observed brightness temperatures of spaceborne microwave sounders, the observation biases of spaceborne microwave sounders can be calculated. The observation bias of the spaceborne microwave sounder is:

[0053] B i =O i -S i

[0054] where B represents the observation bias of the spaceborne microwave sounder, B i represents the observation bias of the i-th channel of the spaceborne microwave sounder, and i = 1, 2, 3…, n represents the n channels of the spaceborne microwave sounder; O represents the observed brightness temperature of the spaceborne microwave sounder, O i represents the observed brightness temperature of the i-th channel of the spaceborne microwave sounder; S represents the simulated brightness temperature of the spaceborne microwave sounder, S i represents the simulated brightness temperature of the i-th channel of the spaceborne microwave sounder.

[0055] By calculating the observation biases of spaceborne microwave sounders, the matching dataset includes the atmospheric parameter dataset, the observed brightness temperatures of spaceborne microwave sounders, the simulated brightness temperatures of spaceborne microwave sounders, and the observation biases of spaceborne microwave sounders. This matching dataset is used as the first calibration dataset for the first calibration of the observation biases of spaceborne microwave sounders.

[0056] Exemplarily, obtaining the first calibrated brightness temperature of the spaceborne microwave sounder based on the first calibration dataset includes:

[0057] Based on the first calibration dataset, a first calibration model for the observation biases of the spaceborne microwave sounder is established, and the first calibrated brightness temperature of the spaceborne microwave sounder is obtained through the first calibration model.

[0058] Specifically, in this embodiment, first, 80% of the matching data in the first calibration dataset is randomly selected as the training dataset of the first calibration data, and the remaining 20% of the matching data in the first calibration dataset is used as the validation dataset of the first calibration dataset. Secondly, a three-layer neural network structure (an input layer, a hidden layer, and an output layer) is established. Taking the observed brightness temperature of the spaceborne microwave sounder in the training dataset of the first calibration dataset as the input and the corresponding observed deviation of the spaceborne microwave sounder as the output, the three-layer neural network structure is trained. By adjusting the number of neurons in the hidden layer, different predicted values of the observed deviation of the spaceborne microwave sounder can be obtained. Then, the root mean square error between the predicted values of different observed deviations of the spaceborne microwave sounder and the corresponding observed deviations of the spaceborne microwave sounder is calculated, and the corresponding relationship between the root mean square error and the number of neurons in the hidden layer is established. The neural network structure corresponding to the number of neurons in the hidden layer with the minimum root mean square error is taken as the optimal neural network, that is, the first calibration model of the observed deviation of the spaceborne microwave sounder. Finally, the observed brightness temperature of the spaceborne microwave sounder in the validation dataset of the first calibration dataset is input into the first calibration model of the observed deviation of the spaceborne microwave sounder to obtain the predicted value B' of the observed deviation of the spaceborne microwave sounder, and then the first calibrated brightness temperature C' of the spaceborne microwave sounder is obtained. The first calibrated brightness temperature C' of the i-th channel of the spaceborne microwave sounder i is expressed as:

[0059] C' i = O i - B' i

[0060] where B' i represents the predicted value of the observed deviation of the i-th channel of the spaceborne microwave sounder.

[0061] Exemplarily, the establishment of the second calibration dataset for correcting the observed deviation of the spaceborne microwave sounder includes:

[0062] Calculating the calibration residual of the spaceborne microwave sounder, and based on the calibration residual of the spaceborne microwave sounder, establishing a second calibration dataset for correcting the observed deviation of the spaceborne microwave sounder.

[0063] Exemplarily, after the establishment of the second calibration dataset for correcting the observed deviation of the spaceborne microwave sounder, it further includes:

[0064] Based on the second calibration dataset, establishing a second calibration model of the observed deviation of the spaceborne microwave sounder.

[0065] Specifically, in this embodiment, first, the first calibrated brightness temperature of the spaceborne microwave sounder is added to the validation dataset of the first calibration dataset. The calibration residual R of the spaceborne microwave sounder is calculated in the validation dataset of the first calibration dataset, and the calibration residual of the spaceborne microwave sounder is added to the validation dataset of the first calibration dataset. The calibration residual R of the i-th channel of the spaceborne microwave sounder i is expressed as:

[0066] R i = C i '- S i

[0067] Secondly, 80% of the matching datasets in the validation dataset of the first calibration dataset are randomly selected as the training dataset of the second calibration dataset, and the remaining 20% of the matching data in the validation dataset of the first calibration dataset are used as the validation dataset of the second calibration dataset. Then, a three-layer neural network structure (an input layer, a hidden layer, and an output layer) is established. The first calibrated brightness temperature, surface temperature, surface humidity, surface pressure, and 10m wind speed of the spaceborne microwave sounder in the training dataset of the second calibration dataset are used as inputs, and the corresponding calibrated residual of the spaceborne microwave sounder is used as the output. The three-layer neural network structure is trained, and different predicted values of the calibrated residual of the spaceborne microwave sounder can be obtained by adjusting the number of neurons in the hidden layer. Finally, the root mean square error between the predicted values of different calibrated residuals of the spaceborne microwave sounder and the corresponding calibrated residuals of the spaceborne microwave sounder is calculated, and the corresponding relationship between the root mean square error and the number of neurons in the hidden layer is established. The neural network structure corresponding to the number of neurons in the hidden layer with the minimum root mean square error is used as the optimal neural network, that is, the second calibration model for the observation deviation of the spaceborne microwave sounder.

[0068] Exemplarily, obtaining the second calibrated brightness temperature of the spaceborne microwave sounder based on the second calibration dataset specifically includes:

[0069] Based on the second calibration dataset, a second calibration model for the observation deviation of the spaceborne microwave sounder is established, and the second calibrated brightness temperature of the spaceborne microwave sounder is obtained according to the second calibration model.

[0070] Specifically, in this embodiment, first, the first calibrated brightness temperature, surface temperature, surface humidity, surface pressure, and 10m wind speed of the spaceborne microwave sounder in the validation dataset of the second calibration dataset are input into the second calibration model for the observation deviation of the spaceborne microwave sounder to obtain the predicted value R' of the calibration residual of the spaceborne microwave sounder i . Then, the second calibrated brightness temperature C″ of the spaceborne microwave sounder is calculated. The second calibrated brightness temperature C of the i-th channel of the spaceborne microwave sounder i is expressed as:

[0071] C i ” = C i '-R i '

[0072] wherein, R' i represents the predicted value of the calibration residual of the i-th channel of the spaceborne microwave sounder.

[0073] Then, in the validation dataset of the second calibration dataset, for each sounding channel of the spaceborne microwave sounder, calculate the root mean square error between the first calibrated brightness temperature of the spaceborne microwave sounder and the simulated brightness temperature of the spaceborne microwave sounder as the root mean square error after the first calibration, and at the same time calculate the root mean square error between the second calibrated brightness temperature of the spaceborne microwave sounder and the simulated brightness temperature of the spaceborne microwave sounder as the root mean square error after the second calibration. Finally, for each sounding channel of the spaceborne microwave sounder, compare the root mean square error after the first calibration with the root mean square error after the second calibration, discard the calibrated brightness temperature corresponding to the larger root mean square error, and take the calibrated brightness temperature corresponding to the smaller root mean square error as the calibrated brightness temperature of this sounding channel. Furthermore, according to the comparison result of the root mean square error after the first calibration and the root mean square error after the second calibration of the spaceborne microwave sounder, obtain the double-calibrated brightness temperature of the spaceborne microwave sounder, and realize the double calibration of the observation deviation of the spaceborne microwave sounder. It should be noted that the larger root mean square error and the smaller root mean square error can be flexibly set according to the actual situation, and are not limited to a certain quantitative value.

[0074] To facilitate the understanding of the double calibration method for the observation deviation of the spaceborne microwave sounder proposed in Embodiment 1, the following gives a detailed example. In this embodiment, the selected spaceborne microwave sounder is the Microwave Temperature Sounder II (MWTS-II) carried on the Fengyun-3 E satellite, and the ERA5 reanalysis dataset of the European Centre for Medium-Range Weather Forecasts is selected as the historical climate dataset to establish the atmospheric parameter dataset required by the present invention. This embodiment conducts double calibration on the observation deviation of MWTS-II.

[0075] The geographical and temporal ranges of the MWTS-II observed brightness temperature and ERA5 reanalysis datasets used in the selection are the same. The geographical range is (15°N - 35°N, 160°E - 210°E), and the temporal range is from January 2023 to December 2023. Atmospheric parameter datasets are established by selecting atmospheric parameters such as temperature profiles, humidity profiles, cloud water profiles, surface temperature, surface humidity, surface pressure, and 10m wind speed from the ERA5 reanalysis dataset. The atmospheric parameter datasets are matched with the MWTS-II observed brightness temperature in both time and space to establish a matching dataset. Among them, the matching rule is that the time error between the atmospheric parameters and the MWTS-II observed brightness temperature is less than 20 minutes, and the longitude and latitude error is less than 0.1°. 635,287 pairs of matching data can be obtained.

[0076] The observation angles of the atmospheric parameters and the MWTS-II observed brightness temperature in the matching dataset are input into the radiative transfer model to calculate the MWTS-II simulated brightness temperature. Based on the MWTS-II simulated brightness temperature and the MWTS-II observed brightness temperature, the MWTS-II observation deviation can be calculated. The MWTS-II observation deviation is:

[0077] B i =O i -S i

[0078] where B represents the MWTS-II observation deviation, B i represents the observation deviation of the i-th channel of MWTS-II, and i = 1, 2, 3…, 15 represents the 15 channels of MWTS-II; O represents the MWTS-II observed brightness temperature, O i represents the observed brightness temperature of the i-th channel of MWTS-II; S represents the MWTS-II simulated brightness temperature, S i represents the simulated brightness temperature of the i-th channel of MWTS-II.

[0079] By calculating the MWTS-II observation deviation, the matching dataset includes the atmospheric parameter dataset, the MWTS-II observed brightness temperature, the MWTS-II simulated brightness temperature, and the MWTS-II observation deviation. This matching dataset is used as the first calibration dataset for the first calibration of the MWTS-II observation deviation.

[0080] For the first correction of the MWTS-II observation deviation, first, randomly select 80% of the matching data in the first correction dataset as the training dataset of the first correction data, and the remaining 20% of the matching data in the first correction dataset as the validation dataset of the first correction dataset; secondly, establish a three-layer neural network structure (one input layer, one hidden layer, and one output layer), take the MWTS-II observed brightness temperature in the training dataset of the first correction dataset as the input, and the matched MWTS-II observation deviation as the output, train the three-layer neural network structure, and different predicted values of the MWTS-II observation deviation can be obtained by adjusting the number of neurons in the hidden layer; then, calculate the root mean square error between the predicted values of different MWTS-II observation deviations and the corresponding MWTS-II observation deviations, establish the corresponding relationship between the root mean square error and the number of neurons in the hidden layer, and take the neural network structure corresponding to the number of neurons in the hidden layer with the minimum root mean square error as the optimal neural network, that is, the first correction model of the MWTS-II observation deviation; finally, input the MWTS-II observed brightness temperature in the validation dataset of the first correction dataset into the first correction model of the MWTS-II observation deviation to obtain the predicted value B' of the MWTS-II observation deviation, and then obtain the first correction brightness temperature C' of the MWTS-II. The first correction brightness temperature C' of the i-th channel of the MWTS-II i It is expressed as:

[0081] C' i = O i - B' i

[0082] Among them, B' i represents the predicted value of the observation deviation of the i-th channel of the MWTS-II.

[0083] Add the first correction brightness temperature of the MWTS-II to the validation dataset of the first correction dataset, calculate the MWTS-II correction residual R in the validation dataset of the first correction dataset, and add the MWTS-II correction residual to the validation dataset of the first correction dataset. The correction residual R of the i-th channel of the MWTS-II i It is expressed as:

[0084] R i = C i '- S i

[0085] Randomly select 80% of the matching datasets in the validation dataset of the first calibration dataset as the training dataset of the second calibration dataset, and the remaining 20% of the matching data in the validation dataset of the first calibration dataset as the validation dataset of the second calibration dataset; establish a three-layer neural network structure (one input layer, one hidden layer, and one output layer), with the first calibrated brightness temperature, surface temperature, surface humidity, surface pressure, and 10m wind speed of MWTS-II in the training dataset of the second calibration dataset as the input, and the matching MWTS-II calibration residuals as the output. Train the three-layer neural network structure, and different predicted values of MWTS-II calibration residuals can be obtained by adjusting the number of neurons in the hidden layer; finally, calculate the root mean square error between the predicted values of different MWTS-II calibration residuals and the corresponding MWTS-II calibration residuals, establish the corresponding relationship between the root mean square error and the number of neurons in the hidden layer, and take the neural network structure corresponding to the number of neurons in the hidden layer with the minimum root mean square error as the optimal neural network, that is, the second calibration model of MWTS-II observation deviation.

[0086] Input the first calibrated brightness temperature, surface temperature, surface humidity, surface pressure, and 10m wind speed of MWTS-II in the validation dataset of the second calibration dataset into the second calibration model of MWTS-II observation deviation to obtain the predicted value R of the MWTS-II calibration residual i '. Then, calculate the second calibrated brightness temperature C″ of MWTS-II, and the second calibrated brightness temperature C of the i-th channel of MWTS-II i ” is expressed as:

[0087] C i ” = C i '- R i '

[0088] Among them, R iIt represents the predicted value of the calibration residual of the \(i\)-th channel of MWTS-II. In the validation dataset of the second calibration dataset, for each detection channel of MWTS-II, the root mean square error between the first calibrated brightness temperature of MWTS-II and the simulated brightness temperature of MWTS-II is calculated as the root mean square error after the first calibration. At the same time, the root mean square error between the second calibrated brightness temperature of MWTS-II and the simulated brightness temperature of MWTS-II is calculated as the root mean square error after the second calibration. Finally, for each detection channel of MWTS-II, the root mean square error after the first calibration is compared with the root mean square error after the second calibration, and the calibrated brightness temperature corresponding to the larger root mean square error is discarded. The calibrated brightness temperature corresponding to the smaller root mean square error is taken as the calibrated brightness temperature of this detection channel. Furthermore, based on the comparison result of the root mean square error after the first calibration and the root mean square error after the second calibration of MWTS-II, the double-calibrated brightness temperature of MWTS-II is obtained, realizing the double calibration of the observation deviation of MWTS-II.

[0089] Calculate the root mean square errors between the first calibrated brightness temperature of MWTS-II and the observed brightness temperature (i.e., the uncalibrated brightness temperature) in the validation dataset of the first calibration dataset and the simulated brightness temperature of MWTS-II as Figure 2 shown. Calculate the root mean square error between the second calibrated brightness temperature of MWTS in the validation dataset of the second calibration dataset and the simulated brightness temperature of MWTS-II, and the comparison with the root mean square error between the first calibrated brightness temperature of MWTS-II and the simulated brightness temperature of MWTS-II in the validation dataset of the first calibration dataset is as Figure 3 shown.

[0090] From Figure 2 it can be seen that compared with the uncalibrated brightness temperature of MWTS-II, the first observation deviation calibration can effectively correct the observation deviation in all 13 channels of MWTS-II. However, the calibration residuals in the near-surface detection channels 1, 2, 3, and 4 of MWTS-II are still very large, being 4.5K, 2.7K, 1.4K, and 0.7K respectively. When the second calibration of the MWTS-II observation deviation is performed, from Figure 3 it can be seen that the calibration residuals in channels 1, 2, 3, 4, and 5 of MWTS-II are further reduced by 1.1K, 0.95K, 0.5K, 0.24K, and 0.16K respectively. And the calibration residuals in the detection channels 6 - 13 of MWTS-II remain unchanged. It can be seen that the proposed double calibration method for the observation deviation of a spaceborne microwave sounder can effectively improve the calibration effect of the observation deviation in the near-surface detection channels 1 - 5 of MWTS-II.

[0091] In Figure 3For each detection channel of MWTS-II, the root mean square error (RMSE) after the first correction is compared with the RMSE after the second correction. The corrected brightness temperature corresponding to the larger RMSE is discarded, and the corrected brightness temperature corresponding to the smaller RMSE is taken as the corrected brightness temperature of this detection channel. Furthermore, the second corrected brightness temperature of MWTS-II is the double-corrected brightness temperature of MWTS-II, achieving the double correction of the observation deviation of MWTS-II.

[0092] The double correction method 100 for the observation deviation of the spaceborne microwave sounder proposed in this embodiment overcomes the adverse effects of the complex observation deviation characteristics of the near-surface detection channel and the window region detection channel of the spaceborne microwave sounder on the correction effect of the observation deviation of the spaceborne microwave sounder. By introducing surface parameters on the basis of the first correction of the observation deviation of the spaceborne microwave sounder, it accurately describes the correction residuals of the near-surface detection channel and the window region channel of the spaceborne microwave sounder, realizes the double correction of the observation deviation of the spaceborne microwave sounder, and makes the observed brightness temperature of the spaceborne microwave sounder after correction closer to the simulated brightness temperature of the spaceborne microwave sounder. The technical solution proposed in the embodiment of the present invention effectively improves the correction effect of the observation deviation of the near-surface detection channel and the window region detection channel of the spaceborne microwave sounder, and is simple and easy to operate.

[0093] Embodiment 2

[0094] Please refer to Figure 4 as described Figure 4 This is a schematic diagram of the principle of the double correction system for the observation deviation of the spaceborne microwave sounder provided in Embodiment 2 of the present invention. As shown in the figure, in this embodiment, the double correction system 400 for the observation deviation of the spaceborne microwave sounder adopts the double correction method 100 for the observation deviation of the spaceborne microwave sounder proposed in the above Embodiment 1, and includes:

[0095] The first correction data set establishment module 401 is used to establish the first correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0096] The first corrected brightness temperature acquisition module 402 is used to obtain the first corrected brightness temperature of the spaceborne microwave sounder based on the first correction data set;

[0097] The second correction data set establishment module 403 is used to establish the second correction data set for correcting the observation deviation of the spaceborne microwave sounder;

[0098] The second corrected brightness temperature acquisition module 404 is used to obtain the second corrected brightness temperature of the spaceborne microwave sounder based on the second correction data set;

[0099] The double correction module 405 is used to compare the first corrected brightness temperature with the second corrected brightness temperature, and complete the double correction of the observation deviation of the spaceborne microwave sounder according to the comparison result.

[0100] Specifically, the first calibration dataset building module 401 is specifically configured to:

[0101] Select atmospheric parameters such as temperature profile, humidity profile, cloud water profile, surface temperature, surface humidity, surface pressure, and 10m wind speed from the historical climate dataset to establish an atmospheric parameter dataset, match the atmospheric parameter dataset with the observed brightness temperature of the spaceborne microwave sounder in terms of time and space to establish a matching dataset. Input the atmospheric parameters and the observation angles of the observed brightness temperature of the spaceborne microwave sounder in the matching dataset into the radiative transfer model to calculate the simulated brightness temperature of the spaceborne microwave sounder. Based on the simulated brightness temperature and the observed brightness temperature of the spaceborne microwave sounder, the observation deviation of the spaceborne microwave sounder can be calculated. The observation deviation of the spaceborne microwave sounder is:

[0102] B i = O i - S i

[0103] Where B represents the observation deviation of the spaceborne microwave sounder, and B i represents the observation deviation of the i-th channel of the spaceborne microwave sounder, where i = 1, 2, 3…, n represents the n channels of the spaceborne microwave sounder; O represents the observed brightness temperature of the spaceborne microwave sounder, and O i represents the observed brightness temperature of the i-th channel of the spaceborne microwave sounder; S represents the simulated brightness temperature of the spaceborne microwave sounder, and S i represents the simulated brightness temperature of the i-th channel of the spaceborne microwave sounder.

[0104] By calculating the observation deviation of the spaceborne microwave sounder, the matching dataset includes the atmospheric parameter dataset, the observed brightness temperature of the spaceborne microwave sounder, the simulated brightness temperature of the spaceborne microwave sounder, and the observation deviation of the spaceborne microwave sounder. This matching dataset is used as the first calibration dataset for the first calibration of the observation deviation of the spaceborne microwave sounder.

[0105] Specifically, the first corrected brightness temperature acquisition module 402 is specifically configured to: randomly select 80% of the matching data in the first corrected dataset as the training dataset of the first corrected data, and the remaining 20% of the matching data in the first corrected dataset as the validation dataset of the first corrected dataset. Establish a three-layer neural network structure (an input layer, a hidden layer, and an output layer), use the observed brightness temperature of the spaceborne microwave sounder in the training dataset of the first corrected dataset as the input, and the corresponding observed deviation of the spaceborne microwave sounder as the output. Train the three-layer neural network structure, and different predicted values of the observed deviation of the spaceborne microwave sounder can be obtained by adjusting the number of neurons in the hidden layer. Calculate the root mean square error between the predicted values of different observed deviations of the spaceborne microwave sounder and the corresponding observed deviations of the spaceborne microwave sounder, establish the corresponding relationship between the root mean square error and the number of neurons in the hidden layer, and use the neural network structure corresponding to the number of neurons in the hidden layer with the minimum root mean square error as the optimal neural network, that is, the first correction model of the observed deviation of the spaceborne microwave sounder. Finally, input the observed brightness temperature of the spaceborne microwave sounder in the validation dataset of the first corrected dataset into the first correction model of the observed deviation of the spaceborne microwave sounder to obtain the predicted value B' of the observed deviation of the spaceborne microwave sounder, and then obtain the first corrected brightness temperature C' of the spaceborne microwave sounder. The first corrected brightness temperature C' of the i-th channel of the spaceborne microwave sounder i is expressed as:

[0106] C' i = O i - B' i

[0107] where B' i represents the predicted value of the observed deviation of the i-th channel of the spaceborne microwave sounder.

[0108] In this embodiment, the second corrected dataset establishment module 403, the second corrected brightness temperature acquisition module 404, and the double correction module 405 are specifically configured to: add the first corrected brightness temperature of the spaceborne microwave sounder to the validation dataset of the first corrected dataset, calculate the correction residual R of the spaceborne microwave sounder in the validation dataset of the first corrected dataset, and add the correction residual of the spaceborne microwave sounder to the validation dataset of the first corrected dataset. The correction residual R of the i-th channel of the spaceborne microwave sounder i is expressed as:

[0109] R i = C i '- S i

[0110] Randomly select 80% of the matching data set in the validation data set of the first calibration data set as the training data set of the second calibration data set, and the remaining 20% of the matching data in the validation data set of the first calibration data set as the validation data set of the second calibration data set. Establish a three-layer neural network structure (an input layer, a hidden layer, and an output layer), with the first calibration brightness temperature, surface temperature, surface humidity, surface pressure, and 10m wind speed of the spaceborne microwave sounder in the training data set of the second calibration data set as the input, and the matching spaceborne microwave sounder calibration residual as the output. Train the three-layer neural network structure, and different predicted values of the spaceborne microwave sounder calibration residual can be obtained by adjusting the number of neurons in the hidden layer. Calculate the root mean square error between the predicted values of different spaceborne microwave sounder calibration residuals and the corresponding spaceborne microwave sounder calibration residuals, establish the corresponding relationship between the root mean square error and the number of neurons in the hidden layer, and take the neural network structure corresponding to the number of neurons in the hidden layer with the minimum root mean square error as the optimal neural network, that is, the second calibration model of the spaceborne microwave sounder observation deviation. Input the first calibration brightness temperature, surface temperature, surface humidity, surface pressure, and 10m wind speed of the spaceborne microwave sounder in the validation data set of the second calibration data set into the second calibration model of the spaceborne microwave sounder observation deviation to obtain the spaceborne

[0111] predicted value R of the microwave sounder calibration residual i . Then, calculate the second calibration brightness temperature C″ of the spaceborne microwave sounder, and the second calibration brightness temperature C of the i-th channel of the spaceborne microwave sounder i is expressed as:

[0112] C i ” = C i ' - R i '

[0113] where R iIt represents the predicted value of the calibration residual of the i-th channel of the spaceborne microwave sounder. In the validation dataset of the second calibration dataset, for each detection channel of the spaceborne microwave sounder, the root mean square error between the first-calibrated brightness temperature and the simulated brightness temperature of the spaceborne microwave sounder is calculated as the root mean square error after the first calibration. At the same time, the root mean square error between the second-calibrated brightness temperature and the simulated brightness temperature of the spaceborne microwave sounder is calculated as the root mean square error after the second calibration. For each detection channel of the spaceborne microwave sounder, the root mean square error after the first calibration is compared with the root mean square error after the second calibration. The calibrated brightness temperature corresponding to the larger root mean square error is discarded, and the calibrated brightness temperature corresponding to the smaller root mean square error is used as the calibrated brightness temperature of the detection channel. Furthermore, based on the comparison result of the root mean square error after the first calibration and the root mean square error after the second calibration of the spaceborne microwave sounder, the double-calibrated brightness temperature of the spaceborne microwave sounder is obtained, realizing the double calibration of the observation deviation of the spaceborne microwave sounder. It should be noted that the larger root mean square error and the smaller root mean square error can be flexibly set according to the actual situation and are not limited to a certain quantitative value.

[0114] The double-calibration system 400 for the observation deviation of the spaceborne microwave sounder proposed in this embodiment overcomes the adverse effects of the complex observation deviation characteristics of the near-surface detection channel and the window area detection channel of the spaceborne microwave sounder on the calibration effect of the observation deviation of the spaceborne microwave sounder. By introducing surface parameters on the basis of the first calibration of the observation deviation of the spaceborne microwave sounder, it accurately describes the calibration residuals of the near-surface detection channel and the window area channel of the spaceborne microwave sounder, realizes the double calibration of the observation deviation of the spaceborne microwave sounder, and makes the observed brightness temperature of the calibrated spaceborne microwave sounder closer to the simulated brightness temperature of the spaceborne microwave sounder. The technical solution proposed in the embodiment of the present invention effectively improves the calibration effect of the observation deviation of the near-surface detection channel and the window area detection channel of the spaceborne microwave sounder, and is simple and easy to operate.

[0115] Embodiment III

[0116] This embodiment provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the double-calibration method 100 for the observation deviation of the spaceborne microwave sounder provided in the above Embodiment I.

[0117] It should be noted that the above-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium accessible by a general-purpose or special-purpose computer.

[0118] Note that the above is only a preferred embodiment of the present invention and the technical principles applied. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein. Various obvious changes, re-adjustments and substitutions can be made by those skilled in the art without departing from the protection scope of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments. Without departing from the concept of the present invention, more other equivalent embodiments can be included, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A dual correction method for the observation deviation of a spaceborne microwave sounder, characterized in that Including: Establishing a first calibration dataset for correcting the observation deviation of the spaceborne microwave sounder; Obtaining the first calibrated brightness temperature of the spaceborne microwave sounder based on the first calibration dataset; Establishing a second calibration dataset for correcting the observation deviation of the spaceborne microwave sounder; Obtaining the second calibrated brightness temperature of the spaceborne microwave sounder based on the second calibration dataset; Comparing the first calibrated brightness temperature with the second calibrated brightness temperature, and completing the dual calibration of the observation deviation of the spaceborne microwave sounder according to the comparison result.

2. The dual correction method for the observation deviation of the spaceborne microwave sounder according to claim 1, wherein The establishment of the first calibration dataset for correcting the observation deviation of the spaceborne microwave sounder includes: Calculating the observation deviation of the spaceborne microwave sounder and establishing a first calibration dataset for correcting the observation deviation of the spaceborne microwave sounder.

3. The dual correction method for the observation deviation of the spaceborne microwave sounder according to claim 2, wherein, The calculation of the observation deviation of the spaceborne microwave sounder includes: Establishing a matching dataset of atmospheric parameter datasets and the observed brightness temperature of the spaceborne microwave sounder in time and space, and calculating the observation deviation of the spaceborne microwave sounder.

4. The dual correction method for the observation deviation of the spaceborne microwave sounder according to claim 3, characterized in that The obtaining of the first calibrated brightness temperature of the spaceborne microwave sounder based on the first calibration dataset includes: Based on the first calibration dataset, establishing a first calibration model for the observation deviation of the spaceborne microwave sounder, and obtaining the first calibrated brightness temperature of the spaceborne microwave sounder through the first calibration model.

5. The dual correction method for the observation deviation of the spaceborne microwave sounder according to claim 1, wherein, The establishment of the second calibration dataset for correcting the observation deviation of the spaceborne microwave sounder includes: Calculating the calibration residual of the spaceborne microwave sounder, and based on the calibration residual of the spaceborne microwave sounder, establishing a second calibration dataset for correcting the observation deviation of the spaceborne microwave sounder.

6. The dual correction method for the observation deviation of the spaceborne microwave sounder according to claim 5, wherein After the establishment of the second calibration dataset for correcting the observation deviation of the spaceborne microwave sounder, it further includes: Based on the second calibration dataset, establishing a second calibration model for the observation deviation of the spaceborne microwave sounder.

7. The dual correction method for the observation deviation of the spaceborne microwave sounder according to claim 6, wherein The obtaining of the second calibrated brightness temperature of the spaceborne microwave sounder based on the second calibration dataset specifically includes: Based on the second calibration dataset, establishing a second calibration model for the observation deviation of the spaceborne microwave sounder, and obtaining the second calibrated brightness temperature of the spaceborne microwave sounder according to the second calibration model.

8. The dual correction method for the observation deviation of the spaceborne microwave sounder according to claim 7, characterized in that, The completion of the dual calibration of the observation deviation of the spaceborne microwave sounder according to the comparison result includes: According to the comparison result, obtaining the calibration brightness temperature with the highest accuracy, and completing the dual calibration of the observation deviation of the spaceborne microwave sounder.

9. A dual correction system for observation deviation of a spaceborne microwave sounder, characterized in that, This system adopts the dual calibration method for the observation deviation of the spaceborne microwave sounder described in claim 1, including: A first calibration dataset establishment module for establishing a first calibration dataset for correcting the observation deviation of the spaceborne microwave sounder; A first calibrated brightness temperature obtaining module for obtaining the first calibrated brightness temperature of the spaceborne microwave sounder based on the first calibration dataset; A second calibration dataset establishment module for establishing a second calibration dataset for correcting the observation deviation of the spaceborne microwave sounder; A second calibrated brightness temperature obtaining module for obtaining the second calibrated brightness temperature of the spaceborne microwave sounder based on the second calibration dataset; A dual correction module is configured to compare the first corrected brightness temperature with the second corrected brightness temperature, and complete the dual correction of the observation deviation of the spaceborne microwave sounder according to the comparison result.

10. A computer-readable storage medium storing computer-executable instructions, characterized in that, When the computer-executed instructions are executed by a processor, they are used to implement the dual correction method for the observation deviation of the spaceborne microwave sounder provided in the embodiment of the first aspect as described above.

Images