FUI inversion method and system, electronic equipment and storage medium

Abstract

The invention discloses an FUI inversion method and system, electronic equipment and a storage medium, and the method comprises the steps: obtaining first satellite data, and carrying out the preprocessing of the first satellite data, and obtaining second satellite data; selecting reflectivity data of wave bands B, G and R from the second satellite data, and performing linear transformation on the reflectivity data of the wave bands B, G and R by using a principal component analysis method to obtain the reflectivity data of the wave bands B, G and R after linear transformation; inputting the reflectivity data of the wave bands B, G and R after linear transformation into an XGBoost model to obtain a predicted FUI value; and carrying out screening processing on the predicted FUI value to obtain a final FUI inversion result. According to the method, the problem of missing of remote sensing reflectivity data in a near-shore area is effectively solved, and the precision of FUI inversion is improved.

Description

Technical Field

[0001] This application relates to the technical field of FUI inversion, and particularly relates to a method, system, electronic device and storage medium for FUI inversion. Background Art

[0002] The inversion of the water color index (FUI) of water bodies has always been a research hotspot. As a key indicator for evaluating the nutritional status and transparency of water bodies, FUI is of great significance for environmental protection, water resource management, and water ecological research. However, the existing FUI inversion methods have many deficiencies, which limit their accuracy and reliability in practical applications. The existing technologies mainly use the data of remote sensing reflectance (R rs (λ)) at specific wavelengths, such as R(645nm), G(555nm), and B band (469nm), to calculate chromaticity coordinates and angles, and perform angle correction according to the remote sensing emissivity of different satellites. Finally, FUI is obtained by looking up a table. This method has limitations. First, the R rs (λ) data is missing in the nearshore area. This is mainly because the water body environment in the nearshore area is complex. Affected by factors such as river input, tidal action, and human activities, the quality of remote sensing data decreases, and even effective reflectance data cannot be obtained. And due to the complex atmospheric environment (especially aerosols), R rs needs to be corrected for the atmosphere, and the complex atmospheric environment makes it difficult to remove the contributions of atmospheric molecules and aerosols, resulting in the failure of atmospheric correction and data loss. Second, the inversion accuracy is low, especially in turbid sea areas. The water body composition in turbid sea areas is complex, and the scattering and absorption effects are strong, resulting in a large range of changes in remote sensing reflectance data. The inversion value is often significantly lower than the true value, which reduces the accuracy of FUI inversion. The above problems need to be solved. Summary of the Invention

[0003] The main purpose of this application is to overcome the shortcomings and deficiencies of the prior art, and provide a method, system, electronic device and storage medium for FUI inversion, which effectively solves the problem of missing remote sensing reflectance data in the nearshore area and improves the accuracy of FUI inversion.

[0004] To achieve the above purpose, this application adopts the following technical solutions:

[0005] In the first aspect, this application provides a method for FUI inversion, including the following steps:

[0006] Obtain first satellite data, and preprocess the first satellite data to obtain second satellite data;

[0007] Select the reflectance data of the B, G, and R bands from the second satellite data, and perform a linear transformation on the reflectance data of the B, G, and R bands using the principal component analysis method to obtain the reflectance data of the B, G, and R bands after the linear transformation;

[0008] Input the reflectance data of the B, G, and R bands after the linear transformation into the XGBoost model to obtain the predicted FUI value;

[0009] Perform a screening process on the predicted FUI value to obtain the final FUI inversion result.

[0010] As a preferred technical solution, the preprocessing includes:

[0011] Perform geometric rough correction and Rayleigh correction on the first satellite data.

[0012] As a preferred technical solution, the preprocessing further includes:

[0013] Perform masking processing on the data of cloud, land, and flare regions in the first satellite data.

[0014] As a preferred technical solution, inputting the reflectance data of the B, G, and R bands after the linear transformation into the XGBoost model to obtain the predicted FUI value includes:

[0015] Use the reflectance data of the B, G, and R bands after the linear transformation as the input features of the XGBoost model;

[0016] The XGBoost model iteratively constructs multiple trees, adding a new tree in each iteration and learning a new function to fit the residuals of the previous prediction;

[0017] When the training process is completed, an XGBoost model containing k trees is obtained. Locate the corresponding leaf nodes in each tree according to the input features, where each leaf node corresponds to a score;

[0018] Finally, add the scores corresponding to each of the k trees to obtain the predicted FUI value.

[0019] As a preferred technical solution, it further includes determining the optimal hyperparameters in the XGBoost model;

[0020] Among them, the optimal hyperparameters in the XGBoost model are obtained by the grid search combined with the cross-validation method.

[0021] As a preferred technical solution, the screening process on the predicted FUI value to obtain the final FUI inversion result includes:

[0022] Exclude FUI values less than 1 or greater than 21, exclude FUIs with the reflectance of the corresponding pixel points greater than 1 or less than or equal to 0, and exclude FUIs with the reflectance of the corresponding pixel points at a wavelength of 1240 nm greater than 0.08, to obtain the final FUI inversion results for each pixel point.

[0023] In a second aspect, the present application provides an FUI inversion system, which is applied to the described FUI inversion method, and includes a preprocessing module, a linear transformation module, a prediction module, and a screening module;

[0024] The preprocessing module is used to obtain first satellite data and preprocess the first satellite data to obtain second satellite data;

[0025] The linear transformation module is used to select the reflectance data of the B, G, and R bands from the second satellite data, and perform a linear transformation on the reflectance data of the B, G, and R bands by using the principal component analysis method to obtain the linearly transformed reflectance data of the B, G, and R bands;

[0026] The prediction module is used to input the linearly transformed reflectance data of the B, G, and R bands into the XGBoost model to obtain the predicted FUI value;

[0027] The screening module is used to perform screening processing on the predicted FUI value to obtain the final FUI inversion result.

[0028] As a preferred technical solution, the preprocessing module is specifically used to perform geometric rough correction and Rayleigh correction on the first satellite data.

[0029] In a third aspect, the present application provides an electronic device, and the electronic device includes:

[0030] At least one processor; and a memory communicatively connected to the at least one processor;

[0031] Wherein, the memory stores computer program instructions executable by the at least one processor, and the computer program instructions are executed by the at least one processor so that the at least one processor can execute the described FUI inversion method.

[0032] In a fourth aspect, the present application provides a computer-readable storage medium, storing a program, and when the program is executed by a processor, the described FUI inversion method is implemented.

[0033] In summary, compared with the prior art, the effective effects brought by the technical solution provided by the present application at least include:

[0034] The present application proposes a method for FUI inversion, which obtains first satellite data and preprocesses the first satellite data to obtain second satellite data; selects the reflectance data of the B, G, and R bands from the second satellite data, and performs a linear transformation on the reflectance data of the B, G, and R bands by using the principal component analysis method (PCA) to obtain the reflectance data of the B, G, and R bands after the linear transformation; inputs the reflectance data of the B, G, and R bands after the linear transformation into the XGBoost model to obtain the predicted FUI value; performs a screening process on the predicted FUI value to obtain the final FUI inversion result. Through the Rayleigh correction in the preprocessing, the data loss caused by aerosol correction is avoided; at the same time, by performing principal component analysis on the reflectance data of the B, G, and R bands and combining the XGBoost model to predict the FUI value, the accuracy of FUI inversion is significantly improved, and the deviation between the inversion value and the true value is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0036] Figure 1 It is a flowchart of a method for FUI inversion provided by an embodiment of the present application;

[0037] Figure 2 It is a map of the FUI spatial distribution result obtained by the FUI inversion method of the present application according to an embodiment of the present application;

[0038] Figure 3 It is a map of the monthly average FUI distribution result obtained by the FUI inversion method of the present application according to an embodiment of the present application;

[0039] Figure 4 It is a block diagram of a FUI inversion system provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] In order to enable those skilled in the art to better understand the solutions of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present application.

[0041] References to "embodiments" in this application mean that the specific features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of this application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0042] Embodiment:

[0043] Please refer to Figure 1 , in an embodiment of this application, a method for FUI inversion is provided, including the following steps:

[0044] S1. Obtain first satellite data and preprocess the first satellite data to obtain second satellite data.

[0045] Furthermore, obtaining the first satellite data of the MODIS-AQUA satellite from the Earth Observation Data Platform means obtaining the Level 1A product data of the satellite in the daytime with a spatial resolution of 1KM. Among them, the Level 1A product refers to the data in which the original radiation directly captured by the satellite sensor is converted into the physical unit of radiance.

[0046] In this embodiment, the preprocessing includes performing geometric rough correction and Rayleigh correction on the first satellite data, and masking the data in the cloud, land, and flare regions in the first satellite data.

[0047] Specifically, the geometric rough correction utilizes the modis_GEO (module for processing MODIS data geolocation information) and modis_L1B (module for geometric rough correction) functions in the SeaDAS software to correct image distortions caused by satellite motion and Earth rotation, etc., to ensure that the geographical features of the first satellite data are consistent with the actual Earth's surface. Secondly, Rayleigh correction is performed on the data after geometric rough correction by adjusting parameter settings to disable unnecessary aerosol subtraction and atmospheric correction functions to obtain the second satellite data after Rayleigh correction.

[0048] In the embodiment of this application, it also includes constructing a dataset to train a model. Specifically, the second satellite data is matched with the measured FUI data to obtain valid second satellite data. Among them, the valid second satellite data is used to train the subsequent XGBoost model.

[0049] Matching the second satellite data with the measured FUI data to obtain valid second satellite data means matching the satellite data on a daily basis with a 3X3 pixel as the matching window with the measured FUI data, specifically including:

[0050] (1) Calculate the distances between the longitude and latitude of the measured FUI data points and the longitude and latitude of each pixel point in the first satellite data.

[0051] (2) Screen out the pixel points whose distances between the longitude and latitude of the measured FUI data points and the longitude and latitude of each pixel point in the first satellite data are less than a preset distance threshold.

[0052] (3) Sort the screened pixel points in ascending order.

[0053] (4) Take the pixel point ranked first as the center of the matched window.

[0054] (5) Select the pixel point ranked first and a preset number of pixel points around the pixel point ranked first as the matching window, that is, select the pixel point ranked first and 8 pixel points around the pixel point ranked first as the matching window (a total of 9 pixel points as the matching window); in addition, it should be noted that if the distances of all pixel points from the measured data point are greater than the preset distance threshold, it means that there is no matching item for this measured FUI data point.

[0055] (6) Eliminate the pixel point data with reflectance data greater than 1 or less than 0 in the matching window, take the average value of the remaining pixels, and obtain the effective second satellite data.

[0056] S2. Select the reflectance data of the B, G, and R bands from the second satellite data, and perform a linear transformation on the reflectance data of the B, G, and R bands by using the principal component analysis method (PCA) to obtain the linearly transformed reflectance data of the B, G, and R bands.

[0057] Furthermore, in the embodiments of the present application, considering the applicability of multi-source satellites (most mainstream water color satellite sensors have data in the R, G, B bands, namely 645nm, 555nm, 469nm (or nearby bands), and the data of other bands vary greatly among different satellites), and the RGB band data has high validity, and the missing rate of other band data is greater than that of the RGB band; therefore, the principal component analysis method (PCA) is used to perform a linear transformation on the reflectance data of the B, G, and R bands (i.e., R rs (469), R rs (555), R rs (645) and divide the linearly transformed data set into a training set and a test set at a ratio of 8:2.

[0058] S3. Input the linearly transformed reflectance data of the B, G, and R bands into the XGBoost model to obtain the predicted FUI value.

[0059] Furthermore, this application uses the XGBoost model as the benchmark model, obtains the optimal hyperparameters of the model through grid search combined with cross-validation, sets the hyperparameters of the XGBoost model accordingly, and replaces the traditional loss function MSE (mean squared error loss function) with Huber Loss to train the XGBoost model. Among them, the optimal hyperparameter settings are as follows: the learning rate is 0.01, the maximum depth of the tree is 7, the number of trees is 300, the sample sampling ratio is 0.2, the objective function is HuberLoss, the minimum sum of child node weights is 1, the feature sampling ratio is 1, and the minimum loss reduction required for splitting is 0.

[0060] Furthermore, the reflectance data of the B, G, and R bands after the linear transformation are used as the input features of the XGBoost model. During the training process of the XGBoost model, the second-order Taylor expansion is used to approximate the loss function, and the optimal tree structure and the values of the leaf nodes are solved by minimizing the approximated loss function. The core is: iteratively construct multiple trees, add a new tree in each iteration, and learn a new function to fit the residuals of the previous prediction; when the training process is completed, an XGBoost model containing k trees is obtained, and the corresponding leaf nodes are located in each tree according to the input features, and each leaf node corresponds to a score; finally, the scores corresponding to each of the k trees are added together to obtain the predicted FUI value, and the predicted FUI value is rounded to the nearest integer.

[0061] In the embodiment of this application, for the convenience of use, the model is packaged as a Transform function, and the user only needs to input the grid-type Rayleigh reflectance data R rs (469), R rs (555), R rs (645) and R rs (1240), and the corresponding grid-type FUI results can be obtained.

[0062] S4. Screen the predicted FUI values to obtain the final FUI inversion result.

[0063] Finally, eliminate the FUI values less than 1 or greater than 21, eliminate the FUI with the reflectance of the corresponding pixel point greater than 1 or less than or equal to 0, and eliminate the FUI with the reflectance of the corresponding pixel point greater than 0.08 at the wavelength of 1240 nm to obtain the final FUI inversion result of each pixel point.

[0064] In another embodiment of this application, some areas of the Pearl River Estuary and the South China Sea are selected for the inversion of the water color FUI value. The selected date is 2003, and the inversion method of this application is used to obtain the FUI spatial distribution result as Figure 2 shown; among them, from Figure 2It can be seen that the inversion method of the present application can also have a high data coverage in the nearshore area (between 22°N and 23°N, between 113°E and 114°E), and the water color index in the nearshore area of the Pearl River Estuary (between 22°N and 23°N, between 113°E and 114°E) is on the high side, and its distribution trend conforms to the actual situation of the sea body.

[0065] In addition, the present application also uses the daily data of the satellite in 2023 for experiments. After inversion, the daily FUI is obtained, and then the daily FUI is averaged by month to obtain the average FUI of the offshore area for each month. The results are as Figure 3 . From Figure 3 it can be seen that the FUI in winter is generally higher than that in summer. Focusing on the East China Sea area, it is found that its turbid area will contract from February to August and expand from August to the next February. Especially in October, its turbid area will extend outward in a tongue shape from the Yangtze Estuary, reflecting the powerful influence of the Yangtze River runoff. Focusing on the Yellow Sea area, it is found that there are high-turbidity waters in the southern coast of Shidao, which is particularly obvious in February and weakens in summer. And the water color situation in the Yellow Sea is relatively complex in winter. It may be affected by strong winter winds, causing strong vertical and horizontal mixing of seawater. When the wind speed is relatively high, the energy on the seawater surface is transferred to the deep layer, breaking the thermocline and density stratification structure that may exist in summer, stirring up suspended substances such as sediment on the seabed, and making the seawater turbid. Focusing on the Bohai Sea area, it is found that the water color of the sea body along the coast from Tianjin to Shandong is generally yellowish, especially the turbid area expands significantly in February and converges in August.

[0066] In summary, the FUI inversion method provided by the present application, through Rayleigh correction in the preprocessing, avoids data loss caused by aerosol correction; at the same time, by performing principal component analysis on the reflectance data of the B, G, and R bands and combining with the XGBoost model to predict the FUI value, it can learn deeper non-linear mapping relationships. Especially in complex water body environments such as turbid sea areas, it significantly improves the accuracy of FUI inversion and reduces the deviation between the inversion value and the true value.

[0067] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously.

[0068] Based on the same idea as a FUI inversion method in the above embodiments, the present application also provides a FUI inversion system, which can be used to execute the above-mentioned FUI inversion method. For the convenience of description, in the structural schematic diagram of an embodiment of the FUI inversion system, only the parts related to the embodiments of the present application are shown. Those skilled in the art can understand that the illustrated structure does not constitute a limitation on the device, and it may include more or fewer components than those shown, or combine certain components, or have different component arrangements.

[0069] Please refer to Figure 4 , in another embodiment of the present application, a FUI inversion system is provided, which includes a preprocessing module 101, a linear transformation module 102, a prediction module 103, and a screening module 104;

[0070] The preprocessing module 101 is used to obtain the first satellite data and preprocess the first satellite data to obtain the second satellite data;

[0071] The linear transformation module 102 is used to select the reflectance data of the B, G, and R bands from the second satellite data, and perform a linear transformation on the reflectance data of the B, G, and R bands by using the principal component analysis method to obtain the linearly transformed reflectance data of the B, G, and R bands;

[0072] The prediction module 103 is used to input the linearly transformed reflectance data of the B, G, and R bands into the XGBoost model to obtain the predicted FUI value;

[0073] The screening module 104 is used to perform screening processing on the predicted FUI value to obtain the final FUI inversion result.

[0074] It should be noted that the FUI inversion system of the present application corresponds one-to-one with the FUI inversion method of the present application. The technical features and their beneficial effects described in the embodiments of the above-mentioned FUI inversion method are all applicable to the embodiments of the FUI inversion system. For the specific content, reference can be made to the description in the method embodiments of the present application, which will not be repeated here. This is hereby declared.

[0075] In addition, in the implementation manner of the FUI inversion system in the above embodiments, the logical division of each program module is only for illustration. In actual applications, according to needs, for example, considering the configuration requirements of the corresponding hardware or the convenience of software implementation, the above functions can be assigned to different program modules to complete, that is, the internal structure of the FUI inversion system is divided into different program modules to complete all or part of the functions described above.

[0076] In another embodiment, an electronic device for implementing the FUI inversion method is provided, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor; when the processor executes the computer program, a FUI inversion method according to any embodiment of the present application is implemented.

[0077] Exemplarily, in this embodiment, the computer program may be divided into one or more modules, and the one or more modules are stored in the memory and executed by the processor to complete the present application. The one or more module elements may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the device.

[0078] The device may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The device may include, but is not limited to, a processor and a memory.

[0079] The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the device, and connects various parts of the entire device through various interfaces and lines.

[0080] The memory may be used to store the computer program and / or modules. The processor realizes various functions of the device by running or executing the computer program and / or modules stored in the memory, and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system, application programs required for at least one function, etc.; in addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0081] Correspondingly, the present application also provides a computer-readable storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute a FUI inversion method described in any one of the above embodiments.

[0082] Those of ordinary skill in the art can understand that all or part of the processes in the above embodiment methods can be completed by instructing relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above various methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided by the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0083] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0084] The above embodiments are preferred embodiments of the present application, but the embodiments of the present application are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principle of the present application shall be equivalent replacement methods and are all included in the protection scope of the present application.

Claims

1. A FUI inversion method, characterized in that: The steps include: Acquire first satellite data, and preprocess the first satellite data to obtain second satellite data; Selecting reflectance data of the B, G, and R bands from the second satellite data, and performing linear transformation on the reflectance data of the B, G, and R bands using a principal component analysis method to obtain reflectance data of the B, G, and R bands after linear transformation; Inputting the reflectivity data of the B, G, and R bands after the linear transformation into the XGBoost model to obtain the predicted FUI value; The predicted FUI value is screened to obtain a final FUI inversion result.

2. A FUI inversion method according to claim 1, characterized in that: The pre-processing comprises: Performing rough geometric correction and Rayleigh correction on the first satellite data.

3. A FUI inversion method according to claim 2, characterized in that: The pre-processing further comprises: Mask processing is performed on the data of the cloud layer, land and flare area in the first satellite data.

4. The FUI inversion method according to claim 1, characterized in that: The linearly transformed reflectivity data of the B, G, and R bands are input into the XGBoost model to obtain the predicted FUI value, including: The reflectivity data of the B, G, and R bands after the linear transformation are used as input features of the XGBoost model; The XGBoost model iteratively builds multiple trees, adds a new tree in each iteration, and learns a new function to fit the residual of the previous prediction; When the training process is completed, an XGBoost model containing k trees is obtained, and the corresponding leaf node is located in each tree according to the input features, wherein each leaf node corresponds to a score; Finally, the scores corresponding to each of the k trees are added together to obtain the predicted FUI value.

5. A FUI inversion method according to claim 4, characterized in that: Also included is determining optimal hyperparameters in the XGBoost model; The optimal hyperparameters in the XGBoost model are obtained by grid search combined with cross-validation method.

6. A FUI inversion method according to claim 1, characterized in that: The predicted FUI value is screened to obtain a final FUI inversion result, including: The FUI values ​​less than 1 or greater than 21, the FUIs with corresponding pixel reflectance greater than 1 or less than or equal to 0, and the FUIs with corresponding pixel reflectance greater than 0.08 at a wavelength of 1240 nm are eliminated to obtain the final FUI inversion results for each pixel.

7. A FUI remote sensing inversion system, characterized in that: A FUI remote sensing inversion method applied to any one of claims 1-6, comprising a preprocessing module, a linear transformation module, a prediction module and a screening module; The preprocessing module is used to obtain first satellite data and preprocess the first satellite data to obtain second satellite data; The linear transformation module is used to select the reflectivity data of the B, G, and R bands from the second satellite data, and perform linear transformation on the reflectivity data of the B, G, and R bands using a principal component analysis method to obtain the reflectivity data of the B, G, and R bands after linear transformation; The prediction module is used to input the reflectivity data of the B, G, and R bands after the linear transformation into the XGBoost model to obtain the predicted FUI value; The screening module is used to screen the predicted FUI value to obtain a final FUI inversion result.

8. A FUI inversion system according to claim 7, characterized in that: The preprocessing module is specifically used to perform rough geometric correction and Rayleigh correction on the first satellite data.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores computer program instructions executable by the at least one processor, and the computer program instructions are executed by the at least one processor so that the at least one processor can perform a FUI inversion method as described in any one of claims 1 to 6.

10. A computer-readable storage medium storing a program, characterized in that: When the program is executed by a processor, a FUI inversion method according to any one of claims 1 to 6 is implemented.

Images