Hourly aerosol optical thickness fusion algorithm, system, electronic equipment and storage medium for geostationary satellite
Through the static satellite hour-by-hour aerosol optical thickness fusion algorithm, the problem of lack of local characteristics and insufficient temporal and spatial correlation of aerosol optical thickness framing in the prior art is solved, and efficient aerosol product construction is achieved, supporting aerosol forecasting and Climate research.
Patent Information
- Application Number
- CN202410058227.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-15
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-01-15
AI Technical Summary
In the production of the sterling satellite FY-4 aerosol tertiary products, the lack of local characteristic constraints in the optical thickness of the aerosol, the insufficient spatial and temporal correlation, the difference in the temporal distribution of different seasons is not considered, and the error of the grading error weight lookup table is relatively high. Big problems make it difficult to obtain accurate hourly aerosol products.
A fusion algorithm for optical thickness of aerosols per hour is proposed. By pre-processing of the remote sensing data of stationary satellites, probability distribution division, error weight function establishment and optimal estimation theory, we construct the truncation error of space-time correlation of aerosols Weight lookup table, refines the data weights in the time and space windows, and realizes the hourly assimilation of the optical thickness of aerosols.
It improves the accuracy and calculation efficiency of aerosol products, can provide optimized observation and analysis fields for atmospheric numerical simulation and data assimilation, and supports aerosol forecasting and climate research.
Smart Images

Figure CN118296293B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of satellite remote sensing technology, and in particular relates to a geostationary satellite hourly aerosol optical thickness fusion algorithm, system, electronic equipment and storage medium. Background Art
[0002] Aerosols are suspended matter widely distributed in the air. They not only have a weather feedback mechanism, but are also important climate forcing factors. The lack of explicit consideration of aerosols is an important reason for the uncertainty in weather forecasts and climate predictions in different regions around the world. Geostationary satellite remote sensing is currently the most important means to obtain aerosol characteristics and their changing characteristics over a large area with high temporal resolution. For example, the visible / infrared radiometer AHI (Advanced Himawari Imager) carried by Japan's new meteorological satellite Himawari-8, the ocean remote sensing sensor GOCI (Geostationary Ocean Color Imager) carried by South Korea's geostationary satellite COMS, and the advanced geostationary orbit radiation imager AGRI (Advanced Geosynchronous Radiation Imager) carried by China's geostationary satellite FY-4 can all obtain the spatiotemporal distribution characteristics of high-frequency aerosols in East Asia. At present, mainstream aerosol satellite products are mainly divided into: Level-1, Level-2 and Level-3 standard data products. Level-1 is the data reconstructed and time-matched from the unprocessed instrument data. Level-2 is the data with the same spatial resolution and coverage of the geographical area developed on the basis of Level-1. Level-3 is the variables expressed in a unified time and space grid, which can be used for centralized scientific research.
[0003] At present, my country's new generation of geostationary satellite FY-4 only provides second-level products with the same time resolution as the first-level products, that is, 15 minutes. There is no third-level aerosol product based on my country's geostationary satellite FY-4. However, the input and output files of atmospheric numerical simulation and data assimilation are usually once every hour. Considering that the size of the time resolution seriously affects the calculation efficiency, it is difficult for current computing resources to use aerosol data once every 15 minutes. Atmospheric numerical simulation and data assimilation urgently need third-level products with hourly time resolution to provide observation fields for numerical prediction models, and then obtain optimized aerosol prediction fields and analysis fields (it should be noted that the hourly products here are not the average values of aerosol observations within the hour. Since the simulation results of the model are instantaneous values, the third-level aerosol products also represent instantaneous values every hour.).
[0004] The current existing technology has the following problems when generating three-level products:
[0005] (1) The aerosol optical depth classification lacks the constraints of local aerosol characteristics;
[0006] (2) Only satellite remote sensing data with a shorter time range are selected to confirm the spatiotemporal correlation of aerosols;
[0007] (3) The difference in the temporal and spatial distribution of aerosols in different seasons is not considered;
[0008] (4) There is a large error in the bin error weight lookup table. Therefore, it is difficult to strictly obtain accurate geostationary satellite FY-4 aerosol level 3 products based on existing technology.
[0009] Therefore, it is urgent to develop a geostationary satellite hourly aerosol optical thickness fusion algorithm, system, electronic equipment and storage medium to solve the above technical problems. Summary of the invention
[0010] In order to solve the above technical problems, the present invention proposes a geostationary satellite hourly aerosol optical thickness fusion algorithm, system, electronic equipment and storage medium to solve the above technical problems.
[0011] The first aspect of the present invention discloses a geostationary satellite hourly aerosol optical thickness fusion algorithm, the method comprising:
[0012] Step S1, preprocessing the acquired geostationary satellite remote sensing raw aerosol data to obtain geostationary satellite aerosol optical thickness data;
[0013] Step S2, calculating the probability distribution of aerosol optical thickness using geostationary satellite aerosol optical thickness data, and dividing the aerosol optical thickness into bins according to the calculated probability distribution;
[0014] Step S3, based on the geostationary satellite aerosol optical thickness data and the binning of the aerosol optical thickness, and using an error weight function, establish a binning error weight lookup table of aerosol spatiotemporal correlation;
[0015] Step S4: using all grid point observation data in the time and space window and the corresponding weights in the bin error weight lookup table of the aerosol spatiotemporal correlation, and estimating new center grid point values of all grid points based on the optimal estimation theory.
[0016] According to the method of the first aspect of the present invention, step S1 specifically comprises:
[0017] Step S11, obtaining geostationary satellite remote sensing original aerosol data, wherein the geostationary satellite remote sensing original aerosol data includes parameters such as aerosol optical thickness, aerosol wavelength index, fine particle ratio, suspended particle mass concentration and credibility;
[0018] Step S12: performing quality control processing on the geostationary satellite remote sensing raw aerosol data according to the credibility parameters to remove data interfered by clouds and / or surface topography, thereby obtaining geostationary satellite aerosol optical thickness data.
[0019] According to the method of the first aspect of the present invention, step S2 specifically comprises:
[0020] Step S21, extracting long-time series aerosol optical depth data from geostationary satellite aerosol optical depth data;
[0021] Step S22, calculating the probability distribution of aerosol optical depth using the long-term series aerosol optical depth data;
[0022] Step S23: dividing the aerosol optical thickness into grades according to the calculated probability distribution, wherein each grade has the same proportion.
[0023] According to the method of the first aspect of the present invention, step S3 specifically comprises:
[0024] Step S31, using an error weight function and based on the geostationary satellite aerosol optical thickness data, calculate the time root mean square error between the grid points and the central grid point at different times within the previous hour, and use the time root mean square error as a measure of time correlation;
[0025] Step S32, establishing a bin error weight lookup table of aerosol time correlation at different times within the first hour based on the bins of aerosol optical thickness and the time root mean square error;
[0026] Step S33, using the error weight function and based on the geostationary satellite aerosol optical thickness data, calculating the spatial root mean square error of the grid points at different distances within a preset distance range and the central grid point, and using the spatial root mean square error as a measure of spatial correlation;
[0027] Step S34: establishing a bin error weight lookup table of aerosol spatial correlation at different distances within a preset distance range based on the bins of aerosol optical thickness and the spatial root mean square error.
[0028] According to the method of the first aspect of the present invention, in step S3, the error weight function specifically includes:
[0029] The formula for calculating the time root mean square error at non-zero time is:
[0030]
[0031] Where N represents the total number of geostationary satellite remote sensing observations within the time window Δt, AOT oeiginalrepresents the aerosol optical thickness data of geostationary satellite remote sensing at the original temporal and spatial resolution, x0 represents the longitude of the center point, y0 represents the latitude of the center point, t0 represents the time of the center point, t i Represents the grid time within the Δt time window from the center point;
[0032] The formula for calculating the root mean square error in non-0km space is:
[0033]
[0034] Where N represents the total number of satellite remote sensing observations within the spatial window ΔL, AOT original represents the aerosol optical thickness data from geostationary satellite remote sensing at the original temporal and spatial resolution, x i and i Indicates the distance to the center point; ΔL indicates the longitude and latitude of the grid point within the distance window;
[0035] The root mean square error at time 0 or 0 km is obtained by polynomial fitting, and the polynomial is expressed as:
[0036] P(z)=a3z 3 +a2z 2 +a1z+a0 (3)
[0037] Where P(z) is the fitting polynomial, a3, a2, a1 and a0 are the polynomial coefficients, z represents the time window, Δt is the time or space window, and ΔL represents the internal distance.
[0038] According to the method of the first aspect of the present invention, step S4 specifically comprises:
[0039] Step S41, using the central grid point observation data in the time and space window and the corresponding weights in the bin error weight lookup table of the aerosol spatiotemporal correlation to calculate the estimated value of the aerosol optical thickness at the central grid point:
[0040] Step S42, based on the calculated estimated value of the aerosol optical thickness at the central grid point, the geostationary satellite aerosol optical thickness data is eliminated to obtain new geostationary satellite aerosol optical thickness data;
[0041] Step S43, based on the new geostationary satellite aerosol optical thickness data and the binning of aerosol optical thickness, and using an error weight function, establish a binning error weight lookup table of aerosol spatiotemporal correlation;
[0042] Step S44: using all grid point observation data in the time and space window and the corresponding weights in the new bin error weight lookup table of aerosol spatiotemporal correlation, and estimating new center grid point values of all grid points based on optimal estimation theory.
[0043] According to the method of the first aspect of the present invention, the estimating of the new central grid point value AOT merged The calculation formula is:
[0044]
[0045]
[0046] Among them, AOT pure is the new geostationary satellite aerosol optical depth, σ pure is the temporal and spatial correlation of the new geostationary satellite aerosol optical depth data, σ merged To estimate the temporal and spatial correlation of the new central grid.
[0047] The second aspect of the present invention discloses a geostationary satellite hourly aerosol optical thickness fusion system, the system comprising:
[0048] The first processing module is configured to pre-process the acquired geostationary satellite remote sensing raw aerosol data to obtain geostationary satellite aerosol optical thickness data;
[0049] The second processing module is configured to calculate the probability distribution of the aerosol optical thickness using the geostationary satellite aerosol optical thickness data, and divide the aerosol optical thickness into bins according to the calculated probability distribution;
[0050] The third processing module is configured to establish a bin error weight lookup table of aerosol spatiotemporal correlation based on the geostationary satellite aerosol optical thickness data and the bins of aerosol optical thickness using an error weight function;
[0051] The fourth processing module is configured to estimate new center grid point values of all grid points based on the optimal estimation theory, using all grid point observation data within the time and space window and the corresponding weights in the binned error weight lookup table of the aerosol spatiotemporal correlation.
[0052] According to the system of the second aspect of the present invention, the first processing module specifically includes:
[0053] The acquisition unit is specifically configured to acquire the original aerosol data sensed by the geostationary satellite, wherein the parameters of the original aerosol data sensed by the geostationary satellite include aerosol optical thickness, aerosol wavelength index, fine particle ratio, suspended particle mass concentration and credibility;
[0054] The preprocessing unit is specifically configured to perform quality control processing on the geostationary satellite remote sensing raw aerosol data according to the credibility parameters to remove the data interfered by clouds and / or surface topography, and then obtain the geostationary satellite aerosol optical thickness data.
[0055] According to the system of the second aspect of the present invention, the second processing module specifically includes:
[0056] The extraction unit is specifically configured to extract long-time series aerosol optical thickness data from geostationary satellite aerosol optical thickness data;
[0057] The calculation unit is specifically configured to calculate the probability distribution of aerosol optical thickness using long-time series aerosol optical thickness data;
[0058] The binning unit is specifically configured to divide the aerosol optical thickness into bins according to the calculated probability distribution, wherein each bin has the same proportion.
[0059] According to the system of the second aspect of the present invention, the third processing module specifically includes:
[0060] The first error calculation unit is specifically configured to calculate the time root mean square error between the grid points and the central grid point at different times within the previous hour based on the geostationary satellite aerosol optical thickness data by using the error weight function, and use the time root mean square error as a measure of time correlation;
[0061] The first establishing unit is specifically configured to establish a bin error weight lookup table of aerosol time correlation at different times within 1 hour before the binning of aerosol optical thickness and the time root mean square error;
[0062] The second error calculation unit is specifically configured to calculate the spatial root mean square error of the grid points at different distances and the central grid point within a preset distance range by using the error weight function and based on the geostationary satellite aerosol optical thickness data, and use the spatial root mean square error as a measure of spatial correlation;
[0063] The second establishing unit is specifically configured to establish a bin error weight lookup table of aerosol spatial correlation at different distances within a preset distance range based on the bins of aerosol optical thickness and the spatial root mean square error.
[0064] According to the system of the second aspect of the present invention, in the third processing module, the error weight function specifically includes:
[0065] The formula for calculating the time root mean square error at non-zero time is:
[0066]
[0067] Where N represents the total number of geostationary satellite remote sensing observations within the time window Δt, AOT original represents the aerosol optical thickness data of geostationary satellite remote sensing at the original temporal and spatial resolution, x0 represents the longitude of the center point, y0 represents the latitude of the center point, t0 represents the time of the center point, ti Represents the grid time within the Δt time window from the center point;
[0068] The formula for calculating the root mean square error in non-0km space is:
[0069]
[0070] Where N represents the total number of satellite remote sensing observations within the spatial window ΔL, AOT original represents the aerosol optical thickness data from geostationary satellite remote sensing at the original temporal and spatial resolution, x i and i Indicates the distance to the center point; ΔL indicates the longitude and latitude of the grid point within the distance window;
[0071] The root mean square error at time 0 or 0 km is obtained by polynomial fitting, and the polynomial is expressed as:
[0072] P(z)=a3z 3 +a2z 2 +a1z+a0 (3)
[0073] Wherein, P(z) is the fitting polynomial, a3, a2, a1 and a0 are the polynomial coefficients, z represents the time window, Δt represents the inner time or space window, and ΔL represents the inner distance.
[0074] According to the system of the second aspect of the present invention, the fourth processing module specifically includes:
[0075] The first estimation unit is specifically configured to calculate an estimated value of the aerosol optical thickness at the central grid point using the central grid point observation data within the time and space window and the corresponding weights in the binned error weight lookup table of the aerosol spatiotemporal correlation:
[0076] The elimination unit is specifically configured to eliminate the geostationary satellite aerosol optical thickness data based on the calculated estimated value of the central grid point aerosol optical thickness to obtain new geostationary satellite aerosol optical thickness data;
[0077] The weight table establishment unit is specifically configured to establish a binned error weight lookup table of aerosol spatiotemporal correlation based on new geostationary satellite aerosol optical thickness data and binned aerosol optical thickness using an error weight function;
[0078] The second estimation unit is specifically configured to estimate the new center grid point values of all grid points based on the optimal estimation theory, using all grid point observation data within the time and space window and the corresponding weights in the new binned error weight lookup table of the aerosol spatiotemporal correlation.
[0079] According to the system of the second aspect of the present invention, in the fourth processing module, the estimating of the new central grid point value AOT merged The calculation formula is:
[0080]
[0081]
[0082] Among them, AOT pure is the new geostationary satellite aerosol optical depth, σ pure is the temporal and spatial correlation of the new geostationary satellite aerosol optical depth data, σ merged To estimate the temporal and spatial correlation of the new central grid.
[0083] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps in any one of the geostationary satellite hourly aerosol optical thickness fusion algorithms in the first aspect of the present disclosure are implemented.
[0084] The fourth aspect of the present invention discloses a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any one of the geostationary satellite hourly aerosol optical thickness fusion algorithms in the first aspect of the present disclosure are implemented.
[0085] In summary, the scheme proposed in the present invention is based on accurate spatiotemporal correlation of aerosols, refines the binned error weight function of observations at different times and distances to the center point, and uses optimal estimation technology to realize the construction of hourly assimilable products of satellite remote sensing aerosol optical thickness. It can also provide important data support for revealing the long-term variation law and spatial differentiation characteristics of atmospheric haze pollution in East Asia, the estimation of health risks of atmospheric pollution and the radiation effect of aerosols. BRIEF DESCRIPTION OF THE DRAWINGS
[0086] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0087] Figure 1 The process of a geostationary satellite hourly aerosol optical thickness fusion algorithm according to an embodiment of the present invention is as follows Figure 1 ;
[0088] Figure 2A schematic diagram of probability distribution of aerosol optical thickness according to an embodiment of the present invention;
[0089] Figure 3 It is a structural diagram of a geostationary satellite hourly aerosol optical thickness fusion system according to an embodiment of the present invention;
[0090] Figure 4 The figure is a structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0091] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0092] The first aspect of the present invention discloses a geostationary satellite hourly aerosol optical thickness fusion algorithm. Figure 1 FIG. 4 is a flow chart of an hourly aerosol optical thickness fusion algorithm for a geostationary satellite according to an embodiment of the present invention. Figure 1 As shown, the method includes:
[0093] Step S1, preprocessing the acquired geostationary satellite remote sensing raw aerosol data to obtain geostationary satellite aerosol optical thickness data;
[0094] Step S2, calculating the probability distribution of aerosol optical thickness using geostationary satellite aerosol optical thickness data, and dividing the aerosol optical thickness into bins according to the calculated probability distribution;
[0095] Step S3, based on the geostationary satellite aerosol optical thickness data and the binning of the aerosol optical thickness, and using an error weight function, establish a binning error weight lookup table of aerosol spatiotemporal correlation;
[0096] Step S4: using all grid point observation data in the time and space window and the corresponding weights in the bin error weight lookup table of the aerosol spatiotemporal correlation, and estimating new center grid point values of all grid points based on the optimal estimation theory.
[0097] In step S1, the acquired geostationary satellite remote sensing raw aerosol data is preprocessed to obtain geostationary satellite aerosol optical thickness data;
[0098] In some embodiments, step S1 of the present invention specifically includes:
[0099] Step S11, obtaining geostationary satellite remote sensing original aerosol data, wherein the geostationary satellite remote sensing original aerosol data includes parameters such as aerosol optical thickness, aerosol wavelength index, fine particle ratio, suspended particle mass concentration and credibility;
[0100] Step S12: performing quality control processing on the geostationary satellite remote sensing raw aerosol data according to the credibility parameters to remove data interfered by clouds and / or surface topography, thereby obtaining geostationary satellite aerosol optical thickness data.
[0101] Specifically, the specific process of step S1 of the hourly aerosol optical thickness fusion algorithm of the geostationary satellite in this embodiment is as follows: first, the original domestic geostationary satellite remote sensing data (such as the remote sensing data of the geostationary satellite Fengyun-4) is obtained, and the parameters contained in the data set include aerosol optical thickness (AOT), aerosol wavelength index (ASI), fine particle fraction (FMR), suspended particle mass concentration (CMR) and credibility (DQF). The aerosol optical thickness here has multiple bands, such as: 470nm, 550nm and 650nm; wherein aerosol optical depths is abbreviated as AOT, angstrom exponent is abbreviated as ASI, fineparticle fraction is abbreviated as FMR, suspended particle mass concentration is abbreviated as CMR, and quality flag is abbreviated as DQF. In this embodiment, the aerosol optical thickness and credibility information of the 550nm band contained in the data set can be extracted. Afterwards, according to the quality control indicator of credibility, the control factors are screened, the aerosol optical thickness data is quality controlled, and the data of poor quality interfered by factors such as clouds and / or surface topography are eliminated. The credibility parameters of the FY-4 geostationary satellite aerosol products are quality flag (DQF), which are "no_value", "bad_pixel", "conditionally_usable_pixel" and "good_pixel", and the corresponding bit codes are 0, 1, 2 and 3, respectively, representing that the pixel has no valid value, poor quality, usable in some cases and good quality. Therefore, when DQF is "no_value", "bad_pixel" and "conditionally_usable_pixel", the credibility of the observation in the pixel is low and it is eliminated.
[0102] In step S2, the probability distribution of aerosol optical thickness is calculated using the geostationary satellite aerosol optical thickness data, and the aerosol optical thickness is divided into bins according to the calculated probability distribution.
[0103] In some embodiments, step S2 of the present invention specifically includes:
[0104] Step S21, extracting long-time series aerosol optical depth data from geostationary satellite aerosol optical depth data;
[0105] Step S22, calculating the probability distribution of aerosol optical depth using the long-term series aerosol optical depth data;
[0106] Step S23: dividing the aerosol optical thickness into grades according to the calculated probability distribution, wherein each grade has the same proportion.
[0107] Specifically, the specific process of step S2 of the hourly aerosol optical thickness fusion algorithm of the geostationary satellite in this embodiment is: first, select a long-term series of geostationary satellite remote sensing observation data, and extract the aerosol optical thickness data in the geostationary satellite data; then use the aerosol optical thickness data in the geostationary satellite data to establish a probability distribution; finally, divide the aerosol optical thickness into grades according to the probability distribution, and the proportion of each grade is basically the same (for example, about 10%). The following is an example of using a one-year time series data to perform aerosol optical thickness grading:
[0108] (1) The remote sensing observation data of Fengyun-4 geostationary satellite from September 2022 to August 2023 were selected, and the aerosol optical thickness data in the 550nm band were extracted from the geostationary satellite data.
[0109] (2) Calculate the corresponding probabilities of different aerosol optical thickness values of the aerosol optical thickness data within the time period, select the data with aerosol optical thickness values ranging from 0 to 5, divide them into 100 levels, and start accumulating from the first level. When the accumulation reaches 10%, the aerosol optical thickness ranges of these levels are used as the first level in the aerosol classification rule used in subsequent fusion. And so on, a total of 10 levels are calculated.
[0110] (3) The data in this time period are analyzed and divided into 10 levels, namely: (0, 0.025), [0.025, 0.05), [0.05, 0.075), [0.075, 0.1), [0.1, 0.15), [0.15, 0.225), [0.225, 0.35), [0.35, 0.6), [0.6, 1.2), [1.2, ∞). Figure 2 shown.
[0111] In step S3, based on the geostationary satellite aerosol optical thickness data and the binning of aerosol optical thickness, an error weight function is used to establish a binning error weight lookup table of aerosol spatiotemporal correlation;
[0112] In some embodiments, step S3 of the present invention specifically includes:
[0113] Step S31, using an error weight function and based on the geostationary satellite aerosol optical thickness data, calculate the time root mean square error between the grid points and the central grid point at different times within the previous hour, and use the time root mean square error as a measure of time correlation;
[0114] Step S32, establishing a bin error weight lookup table of aerosol time correlation at different times within the first hour based on the bins of aerosol optical thickness and the time root mean square error;
[0115] Step S33, using an error weight function and based on the geostationary satellite aerosol optical thickness data, calculating the spatial root mean square error of the grid points at different distances within a preset distance range and the central grid point, and using the spatial root mean square error as a measure of spatial correlation;
[0116] Step S34: establishing a bin error weight lookup table of aerosol spatial correlation at different distances within a preset distance range based on the binning of aerosol optical thickness and the spatial root mean square error.
[0117] In some embodiments, the error weight function in step S3 of the present invention specifically includes:
[0118] The formula for calculating the time root mean square error at non-zero time is:
[0119]
[0120] Where N represents the total number of geostationary satellite remote sensing observations within the time window Δt, AOT original represents the aerosol optical thickness data of geostationary satellite remote sensing at the original temporal and spatial resolution, x0 represents the longitude of the center point, y0 represents the latitude of the center point, t0 represents the time of the center point, t i Represents the grid time within the Δt time window from the center point;
[0121] The formula for calculating the root mean square error in non-0km space is:
[0122]
[0123] Where N represents the total number of satellite remote sensing observations within the spatial window ΔL, AOT original represents the aerosol optical thickness data from geostationary satellite remote sensing at the original temporal and spatial resolution, x i and i Indicates the distance to the center point; ΔL indicates the longitude and latitude of the grid point within the distance window;
[0124] The root mean square error at time 0 or 0 km is obtained by polynomial fitting, and the polynomial is expressed as:
[0125] P(z)=a3z 3 +a2z 2 +a1z+a0 (3)
[0126] Wherein, P(z) is the fitting polynomial, a3, a2, a1 and a0 are the polynomial coefficients, z represents the time window, Δt represents the inner time or space window, and ΔL represents the inner distance.
[0127] Specifically, the specific content of step S3 in the geostationary satellite hourly aerosol optical thickness fusion algorithm of this embodiment is as follows:
[0128] Firstly, a bin error weight lookup table of the original aerosol optical thickness time correlation is established.
[0129] First, calculate the root mean square error between the grid point and the central grid point observations in the first hour, and use the root mean square error σ original This statistical parameter is used as a measure of time correlation. The specific calculation is shown in the following formula:
[0130]
[0131] Where N represents the total number of satellite remote sensing observations in the time window (Δt), AOT original Represents satellite remote sensing aerosol optical thickness data at the original temporal and spatial resolution. x, y and t represent the longitude, latitude and time dimensions respectively. The time window is the previous hour, divided into the previous 15, 30 and 45 minutes.
[0132] Since the above calculation can only obtain the aerosol optical thickness at 15, 30 and 45 minutes, the root mean square error at time 0 needs to be obtained through polynomial fitting. The general form of polynomial fitting is:
[0133] P(x)=a3x 3 +a2x 2 +a1x+a0 (3)
[0134] Where P(x) is the fitting polynomial, and a3, a2, a1, and a0 are the polynomial coefficients. The sum of squares of the fitting errors S is minimized using the least squares method:
[0135]
[0136] Where m represents the number of all available grid points. The root mean square error between the grid points and the central grid point observations for the first 15, 30, and 45 minutes is substituted and the polynomial coefficients that minimize S are found by solving the equation.
[0137] Here is a specific example to explain in detail:
[0138] (1) The period selected for calculation is from September 2022 to August 2023, and the area selected for calculation is 0-50N and 80-150E. The root mean square error between the grid and center grid observations of the geostationary satellite Fengyun-4 aerosol optical thickness in the first 15, 30 and 45 minutes is calculated.
[0139] (2) The root mean square error at time 0 is calculated by polynomial fitting.
[0140] (3) The bin error weight lookup table of the time correlation of the original aerosol optical thickness obtained based on (1) and (2) is shown in Table 1.
[0141] Table 1: Bin error weight lookup table for raw aerosol optical thickness time correlation
[0142]
[0143]
[0144] Secondly, a bin error weight lookup table of the original aerosol optical thickness spatial correlation is established.
[0145] First, calculate the root mean square error between the observations of the grid points within the 12 km distance range and the central grid point, and use the root mean square error σ original This statistical parameter is used as a measure of spatial correlation. The specific calculation is shown in the following formula:
[0146]
[0147] Where N represents the total number of satellite remote sensing observations in the space (ΔL) window, AOT original Represents satellite remote sensing aerosol optical thickness data at the original spatial and temporal resolution. x, y and t represent the longitude, latitude and time dimensions respectively. The spatial window is 12 km, which can be divided into 4, 8 and 12 km.
[0148] Since the above calculations can only obtain the aerosol optical depths of 4, 8, and 12 km, the root mean square error of 0 km needs to be obtained through polynomial fitting. The general form of polynomial fitting is:
[0149] P(z)=a3z 3 +a2xz 2 +a1z+a0
[0150] Where P(z) is the fitting polynomial, a3, a2, a1 and a0 are the polynomial coefficients, z represents the time window, Δt represents the inner time or space window, and ΔL represents the inner distance.
[0151] Minimize the sum of squares of the fitting errors S using the least least squares method:
[0152]
[0153] Substitute the root mean square error between the 4, 8 and 12 km grid points and the central grid point observations, and find the polynomial coefficients that minimize S by solving the equation.
[0154] Here is a specific example to explain in detail:
[0155] (1) The period selected for calculation is from September 2022 to August 2023, and the area selected for calculation is 0-50N and 80-150E. The root mean square error between the observations of the aerosol optical thickness of the geostationary satellite Fengyun-4 at the nearby grid points of 4, 8 and 12 km and the central grid point is calculated.
[0156] (2) The root mean square error of 0 km is calculated by polynomial fitting.
[0157] (3) The bin error weight lookup table of the original aerosol optical depth spatial correlation obtained based on (1) and (2) is shown in Table 2.
[0158] Table 2: Bin error weight lookup table for the original aerosol optical depth spatial correlation
[0159]
[0160] In step S4, all grid point observation data in the time and space window and the corresponding weights in the binned error weight lookup table of the aerosol spatiotemporal correlation are used, and based on the optimal estimation theory, new central grid point values of all grid points are estimated.
[0161] In some embodiments, step S4 of the present invention specifically includes:
[0162] Step S41, using the central grid point observation data in the time and space window and the corresponding weights in the bin error weight lookup table of the aerosol spatiotemporal correlation to calculate the estimated value of the aerosol optical thickness at the central grid point:
[0163] Step S42, based on the calculated estimated value of the aerosol optical thickness at the central grid point, the geostationary satellite aerosol optical thickness data is eliminated to obtain new geostationary satellite aerosol optical thickness data;
[0164] Step S43, based on the new geostationary satellite aerosol optical thickness data and the binning of aerosol optical thickness, and using an error weight function, establish a binning error weight lookup table of aerosol spatiotemporal correlation;
[0165] Step S44: using all grid point observation data in the time and space window and the corresponding weights in the new bin error weight lookup table of aerosol spatiotemporal correlation, and estimating new center grid point values of all grid points based on optimal estimation theory.
[0166] In some embodiments, the new center grid point value AOT is estimated in step S4 of the present invention. merged The calculation formula is:
[0167]
[0168]
[0169] Among them, AOT pure is the new geostationary satellite aerosol optical depth, σ pure is the temporal and spatial correlation of the new geostationary satellite aerosol optical depth data, σ merged To estimate the temporal and spatial correlation of the new central grid.
[0170] Specifically, step S4 in the geostationary satellite hourly aerosol optical thickness fusion algorithm of this embodiment is specifically: constructing the geostationary satellite aerosol optical thickness hourly assimilable product based on the optimal estimation theory and the bin error weight function. The specific process is as follows:
[0171] First, the estimated value of the aerosol optical depth at the central grid is calculated using the observations and their weights within the time and space window:
[0172]
[0173]
[0174] Where N represents the total number of satellite remote sensing observations in the time and space window, σ original (x i ,y i ,t i ) is AOT original (x i ,y i ,t i ) The error variance of the inversion value of the central point x, y and t represent the longitude dimension, latitude dimension and time dimension respectively. And σ est (x0,y0,t0) is AOT est (x o ,y0,t0) is the error variance of the central point estimate. Since the error weight is related to both time and space, the final error weight is the square root of the sum of the squares of the weight in the time window and the weight in the space window.
[0175] The original aerosol optical thickness inversion value is further eliminated. When the original aerosol optical thickness inversion value meets the AOT original (x o ,y0,t0)>AOT est (x o ,y0,t0)+2.58σ pure (x o ,y0,t0), then the AOT is removed original (x o ,y0,t0). If there are grid points that may be identified as clouds within a radius of 12km around the central grid point, the AOT is removed. original (x o ,y0,t0). When the total number of valid observations in the time and space window is less than 20% of the total number of grid points, the AOT is eliminated. original (x o ,y0,t0). The aerosol optical thickness value retained after removal is AOT pure , the weight σ corresponding to the retained aerosol optical depth value pure for
[0176]
[0177] Then, the new aerosol optical thickness value AOT after the above steps is eliminated pure , establish a new bin error weight lookup table for the time correlation of aerosol optical thickness. The specific process includes:
[0178] Calculate the root mean square error between the grid point and the central grid point observations in the first hour, and convert the root mean square error σ pure This statistical parameter is used as a measure of time correlation. The specific calculation is shown in the following formula:
[0179]
[0180] Where N represents the total number of satellite remote sensing observations in the time window (Δt), AOT pure Represents satellite remote sensing aerosol optical thickness data at the original temporal and spatial resolution. x, y and t represent the longitude, latitude and time dimensions respectively. The time window is the first hour, divided into the first 15, 30 and 45 minutes.
[0181] Since the above calculation can only obtain the aerosol optical thickness at 15, 30 and 45 minutes, the root mean square error at time 0 needs to be obtained through polynomial fitting. The general form of polynomial fitting is:
[0182] P(z)=a3z 3 +a2xz 2 +a1z+a0
[0183] Where P(z) is the fitting polynomial, a3, a2, a1 and a0 are the polynomial coefficients, z represents the time window, Δt represents the inner time or space window, and ΔL represents the inner distance.
[0184] Minimize the sum of squares of the fitting errors S using the least least squares method:
[0185]
[0186] Substituting the RMS error between the grid and central grid observations for the first 15, 30, and 45 minutes, the polynomial coefficients that minimize S are found by solving the equation.
[0187] The following is a detailed description using a specific example:
[0188] (1) The period selected for calculation is from September 2022 to August 2023, the area selected for calculation is 0-50N and 80-150E, and the new geostationary satellite FY-4 aerosol optical thickness AOT is calculated. pure Root mean square error between grid and center grid observations for the first 15, 30, and 45 minutes.
[0189] (2) The root mean square error at time 0 is calculated by polynomial fitting.
[0190] (3) The new bin error weight lookup table of aerosol optical thickness time correlation obtained based on (1) and (2) is shown in Table 3.
[0191] Table 3: New aerosol optical thickness time correlation bin error weight lookup table
[0192]
[0193] Then, based on the new aerosol optical thickness value AOT after the above steps, pure , establish a new bin error weight lookup table for the spatial correlation of aerosol optical thickness. The specific process includes:
[0194] Calculate the root mean square error between the observations of the grid points within the 12 km distance and the central grid point, and use the root mean square error σ original This statistical parameter is used as a measure of spatial correlation. The specific calculation is shown in the following formula:
[0195]
[0196] Where N represents the total number of satellite remote sensing observations in the space (ΔL) window, AOT pureRepresents the new satellite remote sensing aerosol optical thickness data, x, y and t represent the longitude dimension, latitude dimension and time dimension respectively. The spatial window is 12km, which can be divided into 4, 8 and 12km.
[0197] Since the above calculations can only obtain the aerosol optical depths of 4, 8, and 12 km, the root mean square error of 0 km needs to be obtained through polynomial fitting. The general form of polynomial fitting is:
[0198] P(z)=a3z 3 +a2xz 2 +a1z+a0
[0199] Where P(z) is the fitting polynomial, a3, a2, a1 and a0 are the polynomial coefficients, z represents the time window, Δt represents the inner time or space window, and ΔL represents the inner distance.
[0200] Minimize the sum of squares of the fitting errors S using the least least squares method:
[0201]
[0202] Substitute the root mean square error between the 4, 8 and 12 km grid points and the central grid point observations, and find the polynomial coefficients that minimize S by solving the equation.
[0203] The following is a detailed description using a specific example:
[0204] (1) The period selected for calculation is from September 2022 to August 2023, and the area selected for calculation is 0-50N and 80-150E. The root mean square error between the observations of the aerosol optical thickness of the new geostationary satellite Fengyun-4 at the nearby grid points of 4, 8 and 12 km and the central grid point is calculated.
[0205] (2) The root mean square error of 0 km is calculated by polynomial fitting.
[0206] (3) The bin error weight lookup table of the spatial correlation of the original aerosol optical depth obtained based on (1) and (2) is shown in Table 4.
[0207] Table 4: New aerosol optical depth spatial correlation bin error weight lookup table
[0208]
[0209]
[0210] Finally, using all grid point observations within the time and space window and the weights corresponding to the bin error lookup table of the new aerosol optical thickness temporal and spatial correlation, the new center grid point value is estimated based on the optimal estimation theory. Estimation of new center grid point value AOTmerged The method used is shown in the following formula:
[0211]
[0212]
[0213] AOT pure is the new geostationary satellite aerosol optical depth, σ pure is the temporal and spatial correlation of the new geostationary satellite aerosol optical depth data, σ merged To estimate the temporal and spatial correlation of the new central grid point. x, y and t represent the longitude, latitude and time dimensions respectively. The spatial window is 12km, which can be divided into 4, 8 and 12km. The temporal window is 1 hour, which can be divided into 15, 30 and 45 minutes.
[0214] The following is a detailed description using a specific example:
[0215] Based on the binned error weight lookup table of the spatiotemporal correlation of the new aerosol optical thickness, the optimal estimated value of the FY-4 aerosol optical thickness at a certain moment is obtained.
[0216] Compared with the prior art, the aerosol optical thickness level in the hourly aerosol optical thickness fusion algorithm of the geostationary satellite in the embodiment of the present invention is reduced (from 13 to 10); and the level division is based on the probability distribution of the aerosol optical thickness in the area, which improves the calculation efficiency while ensuring the algorithm accuracy; the amount of data used in calculating the lookup table is larger and more representative;
[0217] The consideration of the temporal and spatial correlation of aerosols has been refined; the impact of changes in the temporal and spatial distribution characteristics of aerosols before and after quality control has been considered.
[0218] In summary, the geostationary satellite hourly aerosol optical depth fusion algorithm of the embodiment of the present invention is based on the accurate spatiotemporal correlation of aerosols, refines the binned error weight function of observations at different times and distances to the center point, and uses the optimal estimation technology to realize the construction of hourly assimilable products of satellite remote sensing aerosol optical depth. It can also provide important data support for revealing the long-term variation law and spatial differentiation characteristics of atmospheric haze pollution in East Asia, the estimation of health risks of atmospheric pollution and the radiation effect of aerosols.
[0219] The second aspect of the present invention discloses a geostationary satellite hourly aerosol optical thickness fusion system. Figure 3 is a structural diagram of a geostationary satellite hourly aerosol optical thickness fusion system according to an embodiment of the present invention; Figure 3 As shown, the system 100 includes:
[0220] The first processing module 10 is configured to pre-process the acquired geostationary satellite remote sensing raw aerosol data to obtain geostationary satellite aerosol optical thickness data;
[0221] The second processing module 20 is configured to calculate the probability distribution of the aerosol optical thickness using the geostationary satellite aerosol optical thickness data, and divide the aerosol optical thickness into bins according to the calculated probability distribution;
[0222] The third processing module 30 is configured to establish a bin error weight lookup table of aerosol spatiotemporal correlation based on the geostationary satellite aerosol optical thickness data and the bins of aerosol optical thickness and using an error weight function;
[0223] The fourth processing module 40 is configured to estimate new center grid point values of all grid points based on the optimal estimation theory, using all grid point observation data within the time and space window and the corresponding weights in the binned error weight lookup table of the aerosol spatiotemporal correlation.
[0224] According to the system of the second aspect of the present invention, the first processing module 10 specifically includes:
[0225] The acquisition unit is specifically configured to acquire the original aerosol data sensed by the geostationary satellite, wherein the parameters of the original aerosol data sensed by the geostationary satellite include aerosol optical thickness, aerosol wavelength index, fine particle ratio, suspended particle mass concentration and credibility;
[0226] The preprocessing unit is specifically configured to perform quality control processing on the geostationary satellite remote sensing raw aerosol data according to the credibility parameters to remove the data interfered by clouds and / or surface topography, and then obtain the geostationary satellite aerosol optical thickness data.
[0227] According to the system of the second aspect of the present invention, the second processing module 20 specifically includes:
[0228] The extraction unit is specifically configured to extract long-time series aerosol optical thickness data from geostationary satellite aerosol optical thickness data;
[0229] The calculation unit is specifically configured to calculate the probability distribution of aerosol optical thickness using long-time series aerosol optical thickness data;
[0230] The binning unit is specifically configured to divide the aerosol optical thickness into bins according to the calculated probability distribution, wherein each bin has the same proportion.
[0231] According to the system of the second aspect of the present invention, the third processing module 30 specifically includes:
[0232] The first error calculation unit is specifically configured to calculate the time root mean square error between the grid points and the central grid point at different times within the previous hour based on the geostationary satellite aerosol optical thickness data by using the error weight function, and use the time root mean square error as a measure of time correlation;
[0233] The first establishing unit is specifically configured to establish a bin error weight lookup table of aerosol time correlation at different times within 1 hour before the binning of aerosol optical thickness and the time root mean square error;
[0234] The second error calculation unit is specifically configured to calculate the spatial root mean square error of the grid points at different distances and the central grid point within a preset distance range by using the error weight function and based on the geostationary satellite aerosol optical thickness data, and use the spatial root mean square error as a measure of spatial correlation;
[0235] The second establishing unit is specifically configured to establish a bin error weight lookup table of aerosol spatial correlation at different distances within a preset distance range based on the bins of aerosol optical thickness and the spatial root mean square error.
[0236] According to the system of the second aspect of the present invention, in the third processing module 30, the error weight function specifically includes:
[0237] The formula for calculating the time root mean square error at non-zero time is:
[0238]
[0239] Where N represents the total number of geostationary satellite remote sensing observations within the time window Δt, AOT original represents the aerosol optical thickness data of geostationary satellite remote sensing at the original temporal and spatial resolution, x0 represents the longitude of the center point, y0 represents the latitude of the center point, t0 represents the time of the center point, t i Represents the grid time within the Δt time window from the center point;
[0240] The formula for calculating the root mean square error in non-0km space is:
[0241]
[0242] Where N represents the total number of satellite remote sensing observations within the spatial window ΔL, AOT original represents the aerosol optical thickness data from geostationary satellite remote sensing at the original temporal and spatial resolution, x i and i Indicates the distance to the center point; ΔL indicates the longitude and latitude of the grid point within the distance window;
[0243] The root mean square error at time 0 or 0 km is obtained by polynomial fitting, and the polynomial is expressed as:
[0244] P(z)=a3z 3 +a2z 2 +a1z+a0 (3)
[0245] Wherein, P(z) is the fitting polynomial, a3, a2, a1 and a0 are the polynomial coefficients, z represents the time window, Δt represents the inner time or space window, and ΔL represents the inner distance.
[0246] According to the system of the second aspect of the present invention, the fourth processing module 40 specifically includes:
[0247] The first estimation unit is specifically configured to calculate an estimated value of the aerosol optical thickness at the central grid point using the central grid point observation data within the time and space window and the corresponding weights in the binned error weight lookup table of the aerosol spatiotemporal correlation:
[0248] The elimination unit is specifically configured to eliminate the geostationary satellite aerosol optical thickness data based on the calculated estimated value of the central grid point aerosol optical thickness to obtain new geostationary satellite aerosol optical thickness data;
[0249] The weight table establishment unit is specifically configured to establish a binned error weight lookup table of aerosol spatiotemporal correlation based on new geostationary satellite aerosol optical thickness data and binned aerosol optical thickness using an error weight function;
[0250] The second estimation unit is specifically configured to estimate the new center grid point values of all grid points based on the optimal estimation theory, using all grid point observation data within the time and space window and the corresponding weights in the new binned error weight lookup table of the aerosol spatiotemporal correlation.
[0251] According to the system of the second aspect of the present invention, in the fourth processing module 40, the new central grid point value AOT is estimated. merged The calculation formula is:
[0252]
[0253]
[0254] Among them, AOT pure is the new geostationary satellite aerosol optical depth, σ pure is the temporal and spatial correlation of the new geostationary satellite aerosol optical depth data, σ merged To estimate the temporal and spatial correlation of the new central grid.
[0255] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps in the hourly aerosol optical thickness fusion algorithm of a geostationary satellite disclosed in any one of the first aspects of the embodiments of the present invention are implemented.
[0256] Figure 4 is a structural diagram of an electronic device according to an embodiment of the present invention, such as Figure 4 As shown, the electronic device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the electronic device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, near field communication (NFC) or other technologies. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device can be a touch layer covered on the display screen, or a button, a trackball or a touch pad set on the housing of the electronic device, or an external keyboard, touch pad or mouse, etc.
[0257] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a structural diagram of the part related to the technical solution of the present disclosure, and does not constitute a limitation on the electronic device to which the technical solution of the present application is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.
[0258] The fourth aspect of the present invention discloses a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the hourly aerosol optical thickness fusion algorithm for a geostationary satellite disclosed in any one of the first aspects of the embodiments of the present invention are implemented.
[0259] Please note that the technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The above embodiments only express several implementation methods of the present application, and their descriptions are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present application, several variations and improvements can be made, which all belong to the scope of protection of the present application. Therefore, the scope of protection of the patent in this application shall be based on the attached claims.
[0260] The above are preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.
Claims
1. A method for fusion of hourly aerosol optical thickness of geostationary satellites, characterized in that: The method comprises: Step S1, preprocessing the acquired geostationary satellite remote sensing raw aerosol data to obtain geostationary satellite aerosol optical thickness data; Step S2, calculating the probability distribution of aerosol optical thickness using geostationary satellite aerosol optical thickness data, and dividing the aerosol optical thickness into bins according to the calculated probability distribution; Step S3, based on the geostationary satellite aerosol optical thickness data and the binning of the aerosol optical thickness, and using an error weight function, establish a binning error weight lookup table of aerosol spatiotemporal correlation; Step S4, using all grid point observation data in the time and space window and the corresponding weights in the bin error weight lookup table of the aerosol spatiotemporal correlation, and estimating new center grid point values of all grid points based on the optimal estimation theory; The step S1 specifically includes: Step S11, obtaining geostationary satellite remote sensing original aerosol data, wherein the geostationary satellite remote sensing original aerosol data includes parameters such as aerosol optical thickness, aerosol wavelength index, fine particle ratio, suspended particle mass concentration and credibility; Step S12, performing quality control processing on the geostationary satellite remote sensing raw aerosol data according to the credibility parameter to remove data interfered by clouds and / or surface topography, thereby obtaining geostationary satellite aerosol optical thickness data; The step S2 specifically includes: Step S21, extracting long-time series aerosol optical depth data from geostationary satellite aerosol optical depth data; Step S22, calculating the probability distribution of aerosol optical depth using the long-term series aerosol optical depth data; Step S23, dividing the aerosol optical thickness into grades according to the calculated probability distribution, wherein each grade has the same proportion; The step S3 specifically includes: Step S31, using an error weight function and based on the geostationary satellite aerosol optical thickness data, calculate the time root mean square error between the grid points and the central grid point at different times within the previous hour, and use the time root mean square error as a measure of time correlation; Step S32, establishing a bin error weight lookup table of aerosol time correlation at different times within the first hour based on the bins of aerosol optical thickness and the time root mean square error; Step S33, using the error weight function and based on the geostationary satellite aerosol optical thickness data, calculating the spatial root mean square error of the grid points at different distances within a preset distance range and the central grid point, and using the spatial root mean square error as a measure of spatial correlation; Step S34, establishing a bin error weight lookup table of aerosol spatial correlation at different distances within a preset distance range based on the bins of aerosol optical thickness and the spatial root mean square error; In step S3, the error weight function specifically includes: The formula for calculating the time root mean square error at non-zero time is: (1) Where N represents the time window The total number of observations of geostationary satellite remote sensing, Represents the aerosol optical depth data from geostationary satellite remote sensing at the original temporal and spatial resolution, Indicates the center time, Indicates the distance from the center point time; The formula for calculating the root mean square error in non-0km space is: (2) Where N represents the total number of satellite remote sensing observations within the spatial window ΔL, represents the aerosol optical thickness data from geostationary satellite remote sensing at the original temporal and spatial resolution, x i and i Indicates the distance to the center point; ΔL indicates the longitude and latitude of the grid point within the distance window; The root mean square error at time 0 or 0 km is obtained by polynomial fitting, and the polynomial is expressed as: (3) in, is the fitting polynomial, , , and are the polynomial coefficients, z represents the time window, represents the inner time or space window, and ΔL represents the inner distance; The step S4 specifically includes: Step S41, using the central grid point observation data in the time and space window and the corresponding weights in the bin error weight lookup table of the aerosol spatiotemporal correlation to calculate the estimated value of the aerosol optical thickness at the central grid point: Step S42, based on the calculated estimated value of the aerosol optical thickness at the central grid point, the geostationary satellite aerosol optical thickness data is eliminated to obtain new geostationary satellite aerosol optical thickness data; Step S43, based on the new geostationary satellite aerosol optical thickness data and the binning of aerosol optical thickness, and using an error weight function, establish a binning error weight lookup table of aerosol spatiotemporal correlation; Step S44: using all grid point observation data in the time and space window and the corresponding weights in the new bin error weight lookup table of aerosol spatiotemporal correlation, and estimating new center grid point values of all grid points based on optimal estimation theory.
2. The method for fusion of hourly aerosol optical thickness of geostationary satellites according to claim 1, characterized in that: Estimate new center grid point values The calculation formula is: (4) (5) in, is the new geostationary satellite aerosol optical depth, is the temporal and spatial correlation of the new geostationary satellite aerosol optical depth data, To estimate the temporal and spatial correlation of the new central grid.
3. A geostationary satellite hourly aerosol optical thickness fusion system, the system adopts the method according to any one of claims 1-2, characterized in that: The system comprises: The first processing module is configured to pre-process the acquired geostationary satellite remote sensing raw aerosol data to obtain geostationary satellite aerosol optical thickness data; The second processing module is configured to calculate the probability distribution of the aerosol optical thickness using the geostationary satellite aerosol optical thickness data, and divide the aerosol optical thickness into bins according to the calculated probability distribution; The third processing module is configured to establish a bin error weight lookup table of aerosol spatiotemporal correlation based on the geostationary satellite aerosol optical thickness data and the bins of aerosol optical thickness using an error weight function; The fourth processing module is configured to estimate new center grid point values of all grid points based on the optimal estimation theory, using all grid point observation data within the time and space window and the corresponding weights in the binned error weight lookup table of the aerosol spatiotemporal correlation.
4. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps in the method for fusing hourly aerosol optical thickness of a geostationary satellite described in any one of claims 1 to 2 are implemented.
5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the method for merging hourly aerosol optical thickness of a geostationary satellite according to any one of claims 1 to 2 are implemented.
Citation Information
Patent Citations
Gaofen-4 satellite-based inversion method and equipment for optical thickness of aerosol
CN107389617A
Data fusion method and system for monitoring aerosol optical thickness by polar orbit satellite
CN109272053A
Method for acquiring hourly high-resolution PM2.5 data
CN110909309A
Fusion method and device for multi-source satellite aerosol optical thickness data
CN112819026A
Method and system for quantifying correlation between aerosol optical thickness and landscape index
CN113849773A
