An on-orbit calibration method and device for backscattering coefficient σ° of a scatterometer based on optical remote sensing features

Abstract

The application provides a method and device for in-orbit calibration of scatterometer backscattering coefficients based on optical remote sensing features, and relates to the technical field of ocean remote sensing calibration. The method obtains ocean optical remote sensing images and scatterometer backscattering coefficient observation data that satisfy preset matching conditions, and extracts optical features representing sea surface roughness changes from the ocean optical remote sensing images, so that the optical remote sensing information can be used as an external constraint source independent of the scatterometer microwave observation system. Further, by performing geometric normalization processing on the optical features and aggregating the geometric normalization processed optical features to a wind vector element scale, a statistical corresponding relationship between the element scale optical features and the scatterometer backscattering coefficient observation data can be established at a unified scale, and the availability and reference stability of the paired samples are improved.

Description

Technical Field

[0001] This application relates to the field of marine remote sensing calibration technology, and more specifically, to a scatterometer backscattering coefficient based on optical remote sensing characteristics. In-orbit calibration method and apparatus. Background Technology

[0002] Microwave scatterometers typically measure the backscattering coefficient of the sea surface and then use geophysical model functions to invert sea surface wind speed and direction. To ensure the stability and consistency of backscattering coefficient observations during long-term operation, on-orbit calibration of the scatterometer's backscattering coefficient is necessary. Existing on-orbit calibration methods usually rely on external wind field data, geophysical model functions, stable land targets, or active calibrators, which suffer from problems such as error coupling, limited reference areas, and high construction and operation costs, making it difficult to continuously support the calibration of scatterometer backscattering coefficients across global ocean areas. Summary of the Invention

[0003] This application provides at least one embodiment of a scatterometer backscattering coefficient based on optical remote sensing characteristics. In-orbit calibration method and apparatus.

[0004] In a first aspect, embodiments of this application provide a scatterometer backscattering coefficient based on optical remote sensing characteristics. In-orbit calibration methods include:

[0005] Acquire marine optical remote sensing images and scatterometer backscattering coefficients that meet preset matching conditions. The observation data is used to establish a paired sample set; wherein, the preset matching conditions include time matching conditions and spatial matching conditions, and the backscattering coefficient observation data of the scatterometer includes the backscattering coefficient observation value, incident angle, relative azimuth angle, polarization mode, and wind vector element position information; The marine optical remote sensing image is filtered for effective pixels to obtain an effective sea surface pixel set, and optical features characterizing sea surface roughness changes are extracted based on the effective sea surface pixel set; wherein, the optical features include at least one of white cap features, flare features, and sea surface reflection texture features; Based on the observation geometric information corresponding to the marine optical remote sensing image, the optical features are geometrically normalized, and the geometrically normalized optical features are aggregated to the scale corresponding to the wind vector element to obtain the wind vector element scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer. Based on the optical eigenvalues ​​at the wind vector element scale and the backscattering coefficient observation data of the scatterometer, a statistical correspondence between the optical eigenvalues ​​and the observed backscattering coefficients is established, and a reference value for the backscattering coefficient is determined based on the statistical correspondence. Based on the backscattering coefficient reference value and the observed backscattering coefficient value, an external statistical reference quantity is constructed to characterize the system bias structure of the observed backscattering coefficient value; Based on the external statistical reference, bias parameters are determined for on-orbit calibration of the backscattering coefficient observations; wherein, the bias parameters include at least one of additive bias, multiplicative coefficients, and geometrically related system terms associated with the incident angle, the relative azimuth angle, and the polarization mode; The observed backscattering coefficients are corrected based on the bias parameters to obtain the calibrated backscattering coefficients.

[0006] Secondly, embodiments of this application also provide a backscattering coefficient of a scatterometer based on optical remote sensing characteristics. The on-orbit calibration device includes: The acquisition module is used to acquire marine optical remote sensing images and scatterometer backscattering coefficients that meet preset matching conditions. The observation data is used to establish a paired sample set; wherein, the preset matching conditions include time matching conditions and spatial matching conditions, and the backscattering coefficient observation data of the scatterometer includes the backscattering coefficient observation value, incident angle, relative azimuth angle, polarization mode, and wind vector element position information; The filtering module is used to filter the effective pixels of the marine optical remote sensing image to obtain an effective sea surface pixel set, and extract optical features characterizing sea surface roughness changes based on the effective sea surface pixel set; wherein, the optical features include at least one of white hat features, flare features, and sea surface reflection texture features; The aggregation module is used to perform geometric normalization processing on the optical features based on the observation geometric information corresponding to the marine optical remote sensing image, and aggregate the geometrically normalized optical features to the scale corresponding to the wind vector element to obtain the wind vector element scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer. The mapping module is used to establish a statistical correspondence between the optical features and the observed backscattering coefficient based on the wind vector element scale optical feature value and the backscattering coefficient observation data of the scatterometer, and to determine the backscattering coefficient reference value based on the statistical correspondence. The construction module is used to construct an external statistical reference quantity to characterize the system bias structure of the backscattering coefficient observation value based on the backscattering coefficient reference value and the backscattering coefficient observation value; The determination module is used to determine, based on the external statistical reference, bias parameters for on-orbit calibration of the backscattering coefficient observations; wherein the bias parameters include at least one of additive bias, multiplicative coefficient, and geometrically related system terms associated with the incident angle, the relative azimuth angle, and the polarization mode; The correction module is used to correct the observed backscattering coefficient based on the bias parameter to obtain the calibrated backscattering coefficient.

[0007] Thirdly, an optional implementation of this application also provides a computer device, a processor, and a memory, wherein the memory stores machine-readable instructions executable by the processor, and the processor is used to execute the machine-readable instructions stored in the memory. When the machine-readable instructions are executed by the processor, the steps of the first aspect above, or any possible implementation of the first aspect, are performed.

[0008] Fourthly, an optional implementation of this application also provides a computer-readable storage medium storing a computer program that, when run, performs the steps of the first aspect or any possible implementation of the first aspect.

[0009] Regarding the backscattering coefficient of the scatterometer based on the aforementioned optical remote sensing characteristics For a description of the effects of the on-orbit calibration device, computer equipment, and computer-readable storage medium, please refer to the above description of the backscattering coefficient of the scatterometer based on optical remote sensing characteristics. The explanation of the on-orbit calibration method will not be repeated here.

[0010] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of this application.

[0011] The backscattering coefficient of the scatterometer based on optical remote sensing characteristics provided in this application embodiment The on-orbit calibration method and apparatus acquires marine optical remote sensing images and scatterometer backscattering coefficient observation data that meet preset matching conditions. It then extracts optical features characterizing sea surface roughness changes from the marine optical remote sensing images, enabling optical remote sensing information to serve as an external constraint source independent of the scatterometer microwave observation system. Furthermore, by geometrically normalizing the optical features and aggregating them to the wind vector element scale, a statistical correspondence can be established between pixel-scale optical features and scatterometer backscattering coefficient observation data at a unified scale, improving the usability and reference stability of the paired samples.

[0012] Based on this, this application determines the reference value of the backscattering coefficient based on the optical eigenvalues ​​at the wind vector element scale, and constructs an external statistical reference quantity to characterize the systematic bias structure of the backscattering coefficient observations. This allows the bias parameters of the scatterometer's backscattering coefficients to be estimated based on optical remote sensing characteristics, thereby reducing dependence on external wind field data and geophysical model functions, and reducing the coupling between wind field errors, model errors, and scatterometer observation errors. Furthermore, since the bias parameters include at least one of additive bias, multiplicative coefficients, and geometrically related systematic terms associated with the incident angle, relative azimuth angle, and polarization mode, it is possible to characterize and correct the systematic biases in the scatterometer's backscattering coefficients that are related to the observation geometry, improving the consistency and stability of the calibrated backscattering coefficients.

[0013] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. These drawings are incorporated in and constitute a part of this specification. They illustrate embodiments conforming to this application and, together with the specification, serve to explain the technical solutions of this application. It should be understood that the following drawings only show some embodiments of this application and should not be considered as limiting the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0015] Figure 1 This application illustrates a scattering coefficient based on optical remote sensing characteristics provided by some embodiments of a scatterer. Flowchart of the on-orbit calibration method; Figure 2 This application illustrates a scattering coefficient based on optical remote sensing characteristics provided by some embodiments of a scatterer. Schematic diagram of the on-orbit calibration device; Figure 3 A schematic diagram of a computer device provided in some embodiments of this application is shown. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown herein can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0017] Research has revealed that existing on-orbit calibration methods typically rely on external wind field data, geophysical model functions, stable land targets, or active calibrators. These methods suffer from problems such as error coupling, limited reference areas, and high construction and operation costs, making it difficult to sustainably support scatterometer backscattering coefficient calibration across global ocean areas. Furthermore, the systematic bias of the scatterometer backscattering coefficient is usually related to observational conditions such as incident angle, relative azimuth, and polarization mode. Existing methods have limited ability to characterize and correct for these geometrically correlated biases. Optical features in marine optical remote sensing images, such as whitecaps, flares, and sea surface reflection textures, can reflect changes in sea surface roughness. However, current technologies mostly use these features for wind speed or roughness inversion, and have not yet developed an on-orbit calibration method that utilizes optical features to construct a backscattering coefficient reference and further determine the scatterometer backscattering coefficient bias parameters. Therefore, how to estimate and correct the systematic bias of the scatterometer backscattering coefficient using optical remote sensing information is an urgent technical problem to be solved.

[0018] Based on the above research, this application provides a backscattering coefficient of a scatterometer based on optical remote sensing characteristics. The on-orbit calibration method and apparatus acquires marine optical remote sensing images and scatterometer backscattering coefficient observation data that meet preset matching conditions. It then extracts optical features characterizing sea surface roughness changes from the marine optical remote sensing images, enabling optical remote sensing information to serve as an external constraint source independent of the scatterometer microwave observation system. Furthermore, by geometrically normalizing the optical features and aggregating them to the wind vector element scale, a statistical correspondence can be established between pixel-scale optical features and scatterometer backscattering coefficient observation data at a unified scale, improving the usability and reference stability of the paired samples.

[0019] Based on this, this application determines the reference value of the backscattering coefficient based on the optical eigenvalues ​​at the wind vector element scale, and constructs an external statistical reference quantity to characterize the systematic bias structure of the backscattering coefficient observations. This allows the bias parameters of the scatterometer's backscattering coefficients to be estimated based on optical remote sensing characteristics, thereby reducing dependence on external wind field data and geophysical model functions, and reducing the coupling between wind field errors, model errors, and scatterometer observation errors. Furthermore, since the bias parameters include at least one of additive bias, multiplicative coefficients, and geometrically related systematic terms associated with the incident angle, relative azimuth angle, and polarization mode, it is possible to characterize and correct the systematic biases in the scatterometer's backscattering coefficients that are related to the observation geometry, improving the consistency and stability of the calibrated backscattering coefficients.

[0020] The shortcomings of the above solutions are the result of the inventor's practical experience and careful research. Therefore, the discovery process of the above problems and the solutions proposed in this application below should be considered as the inventor's contributions to this application.

[0021] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0022] To facilitate understanding of this embodiment, we will first discuss the backscattering coefficient of a scatterometer based on optical remote sensing characteristics disclosed in this application. The on-orbit calibration method is described in detail. The backscattering coefficient of the scatterometer based on optical remote sensing characteristics provided in the embodiments of this application is described in detail. The on-orbit calibration method is typically implemented using computer equipment with a certain computing power. In some possible implementations, the backscattering coefficient of the scatterometer based on optical remote sensing characteristics... On-orbit calibration methods can be implemented by the processor calling computer-readable instructions stored in memory.

[0023] The following example uses a terminal device as the execution subject to illustrate the backscattering coefficient of a scatterometer based on optical remote sensing features provided in the embodiments of this application. The in-orbit calibration method is explained.

[0024] See Figure 1 The figure shown illustrates a backscattering coefficient of a scatterometer based on optical remote sensing features, provided in an embodiment of this application. A flowchart of the on-orbit calibration method is provided, the method including steps S101~S107, wherein: S101. Obtain marine optical remote sensing images and backscattering coefficient observation data of scatterometers that meet preset matching conditions, and establish a paired sample set; wherein, the preset matching conditions include time matching conditions and spatial matching conditions, and the backscattering coefficient observation data of scatterometers includes backscattering coefficient observation values, incident angle, relative azimuth angle, polarization mode, and position information of wind vector elements.

[0025] The marine optical remote sensing imagery can be in the visible light band and / or near-infrared band. For example, marine optical remote sensing imagery may include apparent reflectance imagery or sea surface reflectance imagery after atmospheric correction. Marine optical remote sensing imagery may also include corresponding observation geometry information, which may include at least one of solar zenith angle, observation zenith angle, and relative optical azimuth angle. The data format of the marine optical remote sensing imagery may include geotagged image file format and / or common web data format. The reflectance variable in the marine optical remote sensing imagery can be used as input data for subsequent extraction of white-hat features, flare features, and sea surface reflectance texture features.

[0026] A scatterometer is an active spaceborne / airborne microwave radar that inverts surface parameters (such as sea surface wind speed and direction, soil moisture, and vegetation status) by emitting microwaves and receiving backscattered echoes (i.e. echoes opposite to the incident direction) from the Earth's surface.

[0027] Backscattering coefficient of the scatterometer (usually referring to the normalized backscattering coefficient, symbol) This is the power scattered per unit area from an incident microwave pulse back towards the radar, measured by a microwave scattering meter. It is used to quantify the scattering intensity of a ground feature at a specific incident angle, polarization, and spectral band. The unit is a dimensionless ratio or decibel (dB).

[0028] Backscattering coefficient of scatterer The observation data can be either primary or secondary data from the scatterometer. In addition to the observed backscattering coefficient values, incident angle, relative azimuth, polarization mode, and wind vector element location information, the scatterometer backscattering coefficient observation data may also include observation time, quality control information, number of observations, and / or observation error parameters. The wind vector element location information may include the center latitude and longitude of the wind vector element, or the spatial range corresponding to the wind vector element. The observed backscattering coefficient values ​​can be expressed in linear units or decibel units; unit conversion can be performed during subsequent processing according to model requirements.

[0029] In some embodiments, establishing a paired sample set can be achieved as follows: for each wind vector element, determine the scatterometer observation time and center position corresponding to the wind vector element; based on time matching conditions, determine the target optical remote sensing image from the ocean optical remote sensing image where the time difference between the optical observation time and the scatterometer observation time is less than or equal to a preset time threshold; based on spatial matching conditions, determine the set of optical pixels corresponding to the wind vector element from the target optical remote sensing image according to the center position of the wind vector element; and establish a paired sample set based on the wind vector element and the corresponding set of optical pixels.

[0030] Specifically, the preset time threshold can be determined based on the rate of change in sea surface state, the satellite transit time interval, and the calibration accuracy requirements. For example, the preset time threshold can be any value within the range of 30 to 90 minutes. If multiple target optical remote sensing images satisfy the time matching condition exist for the same wind vector element, the target optical remote sensing image with the smallest time difference from the scatterometer observation time can be selected from among the multiple target optical remote sensing images. Alternatively, based on the time differences corresponding to the multiple target optical remote sensing images, the time weights corresponding to the multiple target optical remote sensing images can be determined, and the optical feature statistics corresponding to the multiple target optical remote sensing images can be weighted and fused based on the time weights. It should be noted that the object of this weighted fusion can be optical feature statistics, without requiring the generation of new optical remote sensing image products.

[0031] In some embodiments, spatial matching conditions can be used to constrain the spatial overlap between marine optical remote sensing images and wind vector elements. Specifically, marine optical remote sensing images and scatterometer backscattering coefficient observation data can be unified into the same geographic coordinate system or the same projected coordinate system, and a spatial index can be established between the center location of the wind vector element and the set of optical pixels. The spatial index can be used to determine the set of optical pixels corresponding to each wind vector element, thereby providing a basis for subsequently aggregating optical features to the scale corresponding to the wind vector element.

[0032] S102. The effective pixels of the marine optical remote sensing image are screened to obtain an effective sea surface pixel set, and optical features characterizing sea surface roughness changes are extracted based on the effective sea surface pixel set; wherein, the optical features include at least one of white hat features, flare features and sea surface reflection texture features.

[0033] In some embodiments, effective pixel screening of marine optical remote sensing images to obtain an effective set of sea surface pixels can be achieved through the following steps: generating a pixel-level effective identifier based on the product quality mark corresponding to the marine optical remote sensing image and / or a preset screening rule; and removing at least one of cloud pixels, cloud image pixels, land pixels, saturated pixels, striped anomalous pixels, and observed geometric anomalous pixels from the marine optical remote sensing image based on the pixel-level effective identifier to obtain an effective set of sea surface pixels.

[0034] The pixel-level valid identifier can be a binary mask used to indicate whether a corresponding pixel is a valid sea surface pixel. The pixel-level valid identifier can be determined by the quality indicators inherent in the marine optical remote sensing image, such as cloud masks, land masks, or quality classification indicators; or it can be determined by preset filtering rules, such as a combination of at least one of cloud probability thresholds, brightness anomaly thresholds, land-sea masks, and observation geometry thresholds. The set of valid sea surface pixels is the set of pixels retained after invalid pixels are removed, and it does not necessarily have to form a new image product.

[0035] In some embodiments, quality control can also be performed on the scatterometer backscattering coefficient observation data. For example, invalid scatterometer backscattering coefficient observation data can be removed based on quality control information in the scatterometer backscattering coefficient observation data; and / or, scatterometer backscattering coefficient observation data contaminated by precipitation can be removed based on precipitation contamination indicators; and / or, external precipitation products or rainfall rate products can be introduced to remove samples with significant precipitation influence. By performing quality control on both the optical and scatterometer sides, the influence of factors such as clouds, cloud shadows, precipitation, nearshore mixed pixels, and anomalous observation geometry on the calibration results can be reduced.

[0036] In some embodiments, the effective optical coverage corresponding to each wind vector element can be determined, and the corresponding wind vector element sample can be removed if the effective optical coverage is lower than a coverage threshold. The effective optical coverage can be the ratio between the number of effective sea surface pixels within the aggregation window corresponding to the wind vector element and the total number of sea surface pixels within that aggregation window. The coverage threshold can be determined based on actual data quality requirements, for example, it can be 0.5, or other empirical thresholds. Screening by effective optical coverage can avoid instability in the statistical results of optical features caused by cloud boundaries, nearshore mixed pixels, or local missing measurements.

[0037] In some embodiments, extracting optical features characterizing sea surface roughness variations based on an effective sea surface pixel set may include extracting white-hat features. Specifically, a white-hat discrimination value corresponding to each pixel may be determined based on the reflectance of each pixel in the effective sea surface pixel set in one or more visible light bands and / or near-infrared bands; pixels whose white-hat discrimination values ​​satisfy the white-hat discrimination criteria are identified as white-hat pixels; a white-hat score is determined based on the number of white-hat pixels and the number of effective sea surface pixels, and the white-hat score is used as at least one of the optical features. The white-hat discrimination value may include at least one of single-band reflectance, band ratio, and multi-band linear combination value. The white-hat discrimination criteria may include a white-hat discrimination value greater than or equal to a white-hat threshold. The white-hat threshold may be a fixed empirical threshold or an adaptive threshold determined based on a local window.

[0038] In other embodiments, the white-hat feature may further include white-hat brightness statistics and / or white-hat connected component morphology statistics. White-hat brightness statistics may include at least one of the white-hat pixel brightness mean and white-hat pixel brightness quantiles. White-hat connected component morphology statistics may include at least one of the white-hat connected component area and the number of white-hat connected components. The above-mentioned white-hat features can reflect changes in sea surface roughness under high wind or breaking wave conditions.

[0039] In some embodiments, extracting optical features characterizing sea surface roughness variations based on an effective sea surface pixel set may further include extracting flare features. Specifically, pixels satisfying flare observation geometry conditions can be determined from the effective sea surface pixel set; based on the reflectivity of the pixels satisfying the flare observation geometry conditions, at least one of flare intensity statistics, flare gradient statistics, and flare region area fraction is determined, and the determined features are used as at least one of the optical features. The flare observation geometry conditions can be determined based on the solar zenith angle, the observed zenith angle, and the relative optical azimuth angle. The flare intensity statistics may include the local brightness mean, brightness quantile, or brightness peak; the flare gradient statistics can be used to characterize the spatial variation of the flare region; and the flare region area fraction can be used to characterize the proportion of the flare-enhanced region in the effective sea surface pixel set. The aforementioned flare features can reflect sea surface roughness variations under low to medium wind conditions or specular reflection sensitive conditions.

[0040] In some embodiments, extracting optical features characterizing sea surface roughness variations based on an effective set of sea surface pixels may further include extracting sea surface reflection texture features. Specifically, within one or more spatial scale windows, at least one of local variance, spectral energy, and directional texture statistics can be determined based on the reflectance of each pixel in the effective set of sea surface pixels, and the determined feature can be used as at least one of the optical features. Sea surface reflection texture features can characterize the texture variations in the spatial distribution of sea surface reflectance, thereby reflecting the spatial differences caused by changes in sea surface roughness.

[0041] S103. Based on the observation geometric information corresponding to the marine optical remote sensing image, the optical features are geometrically normalized, and the geometrically normalized optical features are aggregated to the scale corresponding to the wind vector element to obtain the wind vector element scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer.

[0042] The observation geometry information corresponding to the marine optical remote sensing image can include the solar zenith angle, the observation zenith angle, and the relative optical azimuth angle. Since optical reflectivity, flare intensity, and texture statistics may be affected by the solar-sensor observation geometry, the optical features can be geometrically normalized before being used to determine the backscattering coefficient reference value.

[0043] In some embodiments, geometric normalization can be implemented in at least one of the following ways.

[0044] The first approach is binning normalization. Specifically, based on the solar zenith angle, the observed zenith angle, and the optical relative azimuth angle corresponding to the marine optical remote sensing image, samples corresponding to optical features can be geometrically binned, and the optical features can be statistically analyzed or modeled within each geometric bin. By performing statistical analysis or modeling within the same geometric bin, the influence of different optical observation geometries on optical features can be reduced.

[0045] The second approach is function normalization. Specifically, optical features can be normalized based on a geometric normalization function determined by the solar zenith angle, the observed zenith angle, and the optical relative azimuth angle corresponding to the marine optical remote sensing image. For example, for flare-related features, a geometric normalization function can be used to correct the flare intensity statistics, or the geometric terms in the flare-related features can be eliminated through regression.

[0046] The third approach is joint modeling. Specifically, the solar zenith angle, the observed zenith angle, and the optical relative azimuth angle corresponding to the ocean optical remote sensing image can be used as input features of the model, and together with the optical features, they can be used to establish subsequent statistical correspondences. Thus, when establishing the statistical correspondence between optical features and observed backscattering coefficients, the influence of optical observation geometry on the optical features can be considered simultaneously.

[0047] After geometric normalization, the geometrically normalized optical features can be aggregated to the scale corresponding to the wind vector element. Specifically, an aggregation window can be determined based on the center position of each wind vector element; effective sea surface pixels are determined within the aggregation window; and based on the geometrically normalized optical features corresponding to the effective sea surface pixels within the aggregation window, at least one of the following statistics—mean, quantiles, and weighted mean—is determined as the optical feature value at the wind vector element scale. The aggregation window can be a circular window, a rectangular window, or a spatial window of other shapes centered on the wind vector element's center position. The radius or size of the aggregation window can be on the same order of magnitude as the spatial resolution of the wind vector element.

[0048] In some embodiments, the weighted mean can be determined based on distance weights and / or time weights. Specifically, the distance weight can be determined based on the distance between the center locations of the effective sea surface pixels and the wind vector pixels; the time weight can be determined based on the time difference between the optical observation time of the marine optical remote sensing image to which the effective sea surface pixel belongs and the scatterometer observation time corresponding to the wind vector pixel; and the weighted mean is obtained by weighting the geometrically normalized optical features corresponding to the effective sea surface pixels based on the distance weights and time weights. Thus, optical pixels that are closer to the center location of the wind vector pixel and whose observation time is closer to the scatterometer observation time can have higher weights, thereby improving the representativeness of the optical feature values ​​at the wind vector pixel scale.

[0049] S104. Based on the optical characteristic value of the wind vector element scale and the backscattering coefficient observation data of the scatterometer, establish a statistical correspondence between the optical characteristic and the backscattering coefficient observation value, and determine the backscattering coefficient reference value based on the statistical correspondence.

[0050] The optical eigenvalues ​​at the wind vector element scale can include at least one of the following statistical values ​​at the wind vector element scale: white hat fraction, white hat brightness statistics, white hat connected domain morphology statistics, flare intensity statistics, flare gradient statistics, flare region area fraction, local variance, spectral energy, and directional texture statistics. The backscattering coefficient reference value is a numerical value determined based on the optical eigenvalues ​​at the wind vector element scale, used to characterize the reference level of the backscattering coefficient of the corresponding wind vector element under the corresponding observation conditions.

[0051] In some embodiments, samples in the paired sample set can be stratified or binned according to polarization and incident angle. Within each stratum or bin, a statistical correspondence is established based on the mapping relationship between wind vector element-scale optical eigenvalues ​​and observed backscattering coefficients. Based on the statistical correspondence, a reference value for the backscattering coefficient corresponding to the wind vector element-scale optical eigenvalue is determined. Furthermore, the relative azimuth angle in the backscattering coefficient observation data from the scatterometer can also be used in stratification or binning, or as an input parameter for establishing the statistical correspondence. Thus, the influence of the scatterometer observation geometry on the observed backscattering coefficients can be considered when establishing the statistical correspondence.

[0052] In some embodiments, statistical correspondences can be established using at least one of linear regression, piecewise regression, monotonic constrained regression, nonlinear regression, binning statistics, and robust regression. For example, within each bin determined by polarization mode and incident angle, the wind vector element-scale optical eigenvalues ​​can be used as independent variables, and the observed backscattering coefficients as dependent variables, to fit a mapping relationship for predicting backscattering coefficient reference values. As another example, under high wind or breaking wave conditions, a monotonic constrained regression model can be established based on the monotonic relationship between the white hat score and the observed backscattering coefficients. Furthermore, in cases where outliers may exist in the sample, robust regression can be used to establish statistical correspondences to reduce the impact of outliers on the determination of backscattering coefficient reference values.

[0053] In some embodiments, the reference value for the backscattering coefficient can be expressed as: The wind vector element-scale optical eigenvalues ​​can be expressed as: The statistical correspondence can be expressed as: .

[0054] in, Indicates the angle of incidence. Indicates relative azimuth. This indicates the polarization mode. This expression is only used to illustrate the possible forms of statistical correspondences and does not limit the statistical correspondences to adopting a specific functional form.

[0055] S105. Based on the backscattering coefficient reference value and the backscattering coefficient observation value, construct an external statistical reference quantity to characterize the system bias structure of the backscattering coefficient observation value.

[0056] In some embodiments, the external statistical reference may include a backscattering coefficient reference value and / or a bias term. The bias term may be determined based on the difference between the observed backscattering coefficient value and the backscattering coefficient reference value. For example, the bias term may be expressed as... ,in, This represents the observed value of the backscattering coefficient. This represents the reference value for the backscattering coefficient. The deviation term can be used to characterize the degree of deviation of the observed backscattering coefficient value from the reference value.

[0057] In some embodiments, the external statistical reference quantity may further include statistical residual information, bin mean difference information, or equivalent bias index determined based on multiple samples. For example, within a bin determined by polarization mode, incident angle, and / or relative azimuth angle, the mean difference, standard deviation, or quantile difference between the observed backscattering coefficients and the reference backscattering coefficients can be statistically analyzed and used as an external statistical reference quantity to characterize the structure of the system bias. In this way, the external statistical reference quantity can reflect not only the overall bias but also the structural bias of the observed backscattering coefficients as the incident angle, relative azimuth angle, and polarization mode change.

[0058] In some embodiments, since the backscattering coefficient reference value is determined based on optical features in marine optical remote sensing images, the external statistical reference can be an external constraint independent of the scatterometer microwave observation system. Using an external statistical reference can reduce error coupling problems that arise when calibration relies solely on external wind field data or geophysical model functions.

[0059] S106. Based on the external statistical reference, determine the bias parameters for on-orbit calibration of the backscattering coefficient observations; wherein the bias parameters include at least one of additive bias, multiplicative coefficient, and geometric correlation system terms related to the incident angle, the relative azimuth angle, and the polarization mode.

[0060] In some embodiments, an objective function for estimating the bias parameter can be constructed based on an external statistical reference. The objective function may include the sum of residual loss values ​​corresponding to multiple samples. The residual loss value can be determined based on the residual between a candidate corrected backscattering coefficient and a reference backscattering coefficient. The candidate corrected backscattering coefficient is the backscattering coefficient obtained by correcting the observed backscattering coefficient values ​​based on the bias parameter to be determined. The bias parameter can be determined by minimizing the objective function.

[0061] Specifically, the bias parameter can be used to correct the observed backscattering coefficients according to the following model: .

[0062] in, This represents the calibrated backscattering coefficient. This represents the observed value of the backscattering coefficient. Indicates the multiplicative coefficient. Indicates additive bias. Represents geometrically related system terms. Indicates the angle of incidence. Indicates relative azimuth. The polarization mode is indicated. Geometrically relevant system terms can be represented using binning constants, low-order polynomials, piecewise functions, or lookup tables.

[0063] In some embodiments, the objective function can be represented as the sum of residual loss values ​​for multiple samples. The residuals can be the difference between the candidate corrected backscattering coefficient and the reference backscattering coefficient. The loss function can include a quadratic loss function or a robust loss function. A robust loss function can include the Huber loss function or other loss functions that can reduce the impact of outliers. By employing a robust loss function, the impact of cloud remnants, precipitation pollution, nearshore mixed pixels, or other outliers on the bias parameter estimation results can be reduced.

[0064] In other embodiments, the bias parameter can also be determined by calculating the difference between bin means. Specifically, within each bin determined by the polarization mode and the incident angle, a first mean of the observed backscattering coefficients can be determined, and a second mean of the reference backscattering coefficients can be determined; based on the difference between the first and second means, the additive bias within the corresponding bin can be determined. Furthermore, the relative azimuth angle can also be used in binning to determine the geometrically dependent system terms related to the incident angle, relative azimuth angle, and polarization mode.

[0065] In other embodiments, the bias parameters can also be determined by robust regression. Specifically, the bias parameters can be obtained by using the difference between the observed backscattering coefficient and the reference backscattering coefficient as the dependent variable, and robust regression with at least one of the incident angle, relative azimuth angle, polarization mode, and wind vector element-scale optical eigenvalues ​​as independent variables. In this way, both the overall bias and the geometrically related system terms can be estimated simultaneously.

[0066] S107. The observed backscattering coefficient is corrected based on the bias parameter to obtain the calibrated backscattering coefficient.

[0067] In some embodiments, the bias parameters determined in step S106 can be applied to the backscattering coefficient observations to obtain calibrated backscattering coefficients. Specifically, the backscattering coefficient observations can be corrected based on at least one of multiplicative coefficients, additive bias, and geometrically dependent system terms to reduce the overall bias and / or geometrically dependent system bias in the backscattering coefficient observations.

[0068] In some embodiments, the calibrated backscattering coefficients can be used to generate subsequent wind field products and / or backscattering coefficient products. By using calibrated backscattering coefficients, the consistency and stability of subsequent wind field inversion or marine environmental monitoring results can be improved.

[0069] In some embodiments, consistency assessment can also be performed based on the bias parameters and / or the calibrated backscattering coefficients. For example, the mean, standard deviation, and residual structure changes of the backscattering coefficients before and after correction can be statistically analyzed; alternatively, the bias parameters can be statistically analyzed according to the time and / or spatial dimensions to obtain the time series and / or spatial distribution of the bias parameters, and the long-term consistency of the backscattering coefficients of the scatterometer can be assessed based on the time series and / or spatial distribution of the bias parameters. Furthermore, if the time series of the bias parameters meets a preset drift condition, and / or if there is an anomalous region in the spatial distribution of the bias parameters that meets a preset anomaly condition, the anomalous trajectory segment and / or anomalous region can be identified.

[0070] Through the above steps, this application can construct an external statistical reference quantity to characterize the systematic bias structure of backscattering coefficient observations based on optical features reflecting sea surface roughness variations in marine optical remote sensing images, and further determine the bias parameters used for on-orbit calibration. Since the optical features are geometrically normalized and aggregated to the wind vector element scale, a statistical correspondence can be established with the backscattering coefficient observation data of the scatterometer at a unified scale. Furthermore, since the bias parameters can include geometrically correlated systematic terms related to the incident angle, relative azimuth angle, and polarization mode, the geometrically correlated systematic bias in the backscattering coefficient observations can be characterized and corrected, thereby improving the consistency and stability of the calibrated backscattering coefficients.

[0071] The backscattering coefficient of the scatterometer based on optical remote sensing characteristics provided in this application The on-orbit calibration method acquires marine optical remote sensing images and scatterometer backscattering coefficient observation data that meet preset matching conditions. It then extracts optical features characterizing sea surface roughness changes from the marine optical remote sensing images, enabling optical remote sensing information to serve as an external constraint source independent of the scatterometer microwave observation system. Furthermore, by geometrically normalizing the optical features and aggregating them to the wind vector element scale, a statistical correspondence can be established between pixel-scale optical features and scatterometer backscattering coefficient observation data at a unified scale, improving the usability and reference stability of the paired samples.

[0072] Based on this, this application determines the reference value of the backscattering coefficient based on the optical eigenvalues ​​at the wind vector element scale, and constructs an external statistical reference quantity to characterize the systematic bias structure of the backscattering coefficient observations. This allows the bias parameters of the scatterometer's backscattering coefficients to be estimated based on optical remote sensing characteristics, thereby reducing dependence on external wind field data and geophysical model functions, and reducing the coupling between wind field errors, model errors, and scatterometer observation errors. Furthermore, since the bias parameters include at least one of additive bias, multiplicative coefficients, and geometrically related systematic terms associated with the incident angle, relative azimuth angle, and polarization mode, it is possible to characterize and correct the systematic biases in the scatterometer's backscattering coefficients that are related to the observation geometry, improving the consistency and stability of the calibrated backscattering coefficients.

[0073] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0074] Based on the same inventive concept, this application also provides a backscattering coefficient of a scatterometer based on optical remote sensing characteristics. The on-orbit calibration method corresponds to the scattering coefficient of the scatterometer based on optical remote sensing characteristics. The on-orbit calibration device, because the principle of the device in this application embodiment solves the problem is similar to the backscattering coefficient of the scatterometer based on optical remote sensing characteristics described above in this application embodiment. The on-orbit calibration method is similar, so the implementation of the device can be found in the implementation of the method, and the repeated parts will not be described again.

[0075] Reference Figure 2 The figure shown illustrates a backscattering coefficient of a scatterometer based on optical remote sensing features, provided in an embodiment of this application. A schematic diagram of an on-orbit calibration device, the device comprising: The acquisition module 210 is used to acquire marine optical remote sensing images and scatterometer backscattering coefficients that meet preset matching conditions. The observation data is used to establish a paired sample set; wherein, the preset matching conditions include time matching conditions and spatial matching conditions, and the backscattering coefficient observation data of the scatterometer includes the backscattering coefficient observation value, incident angle, relative azimuth angle, polarization mode, and wind vector element position information; The filtering module 220 is used to filter the effective pixels of the marine optical remote sensing image to obtain an effective sea surface pixel set, and extract optical features characterizing sea surface roughness changes based on the effective sea surface pixel set; wherein, the optical features include at least one of white hat features, flare features, and sea surface reflection texture features; The aggregation module 230 is used to perform geometric normalization processing on the optical features based on the observation geometric information corresponding to the marine optical remote sensing image, and aggregate the geometrically normalized optical features to the scale corresponding to the wind vector element to obtain the wind vector element scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer. The mapping module 240 is used to establish a statistical correspondence between the optical features and the observed backscattering coefficient based on the wind vector element scale optical feature value and the backscattering coefficient observation data of the scatterometer, and to determine the backscattering coefficient reference value based on the statistical correspondence. Construction module 250 is used to construct an external statistical reference quantity to characterize the system bias structure of the observed backscattering coefficient based on the backscattering coefficient reference value and the observed backscattering coefficient value; The determination module 260 is used to determine, based on the external statistical reference, bias parameters for on-orbit calibration of the backscattering coefficient observations; wherein the bias parameters include at least one of additive bias, multiplicative coefficient, and geometric correlation system terms related to the incident angle, the relative azimuth angle, and the polarization mode; The correction module 270 is used to correct the observed backscattering coefficient based on the bias parameter to obtain the calibrated backscattering coefficient.

[0076] In some embodiments, the acquisition module 210 is used for: For each wind vector element, determine the scatterometer observation time and center position corresponding to that wind vector element; Based on the time matching condition, a target optical remote sensing image is determined from the ocean optical remote sensing image whose time difference between the optical observation time and the scatterometer observation time is less than or equal to a preset time threshold. Based on the spatial matching conditions, the set of optical pixels corresponding to the wind vector element is determined from the target optical remote sensing image according to the center position of the wind vector element. Based on the wind vector element and the corresponding optical pixel set, the paired sample set is established.

[0077] In some embodiments, the filtering module 220 is used for: Based on the product quality marks and / or preset screening rules corresponding to the marine optical remote sensing images, generate pixel-level valid identifiers; Based on the pixel-level effective identifier, at least one of cloud pixels, cloud image pixels, land pixels, saturation pixels, stripe anomaly pixels, and observed geometric anomaly pixels in the marine optical remote sensing image is removed to obtain the effective sea surface pixel set. The device further includes a processing module for: Based on the quality control information in the backscattering coefficient observation data of the scatterer, invalid backscattering coefficient observation data of the scatterer are removed; and / or, Backscattering coefficient observation data from scatterometers were excluded based on precipitation pollution identification; and / or, Determine the effective optical coverage for each wind vector element, and remove the corresponding wind vector element sample if the effective optical coverage is lower than the coverage threshold.

[0078] In some embodiments, the filtering module 220 is used for: Based on the reflectance of each pixel in the effective sea surface pixel set in one or more visible light bands and / or near-infrared bands, the white hat discrimination value corresponding to each pixel is determined. Pixels whose white hat discrimination value meets the white hat discrimination condition are determined as white hat pixels. The white hat score is determined based on the number of white hat pixels and the number of effective sea surface pixels. Pixels that satisfy the flare observation geometry conditions are determined from the effective set of sea surface pixels. Based on the reflectance of the pixels that satisfy the flare observation geometry conditions, at least one of the following is determined: flare intensity statistics, flare gradient statistics, and flare region area fraction. Within a window of one or more spatial scales, at least one of the following is determined based on the reflectance of each pixel in the effective sea surface pixel set: local variance, spectral energy, and directional texture statistics.

[0079] In some embodiments, the aggregation module 230 is used for: Based on the solar zenith angle, the observed zenith angle, and the relative azimuth angle corresponding to the ocean optical remote sensing image, the samples corresponding to the optical features are geometrically binned, and the optical features are statistically analyzed or modeled within each geometric bin; alternatively, the optical features are normalized based on a geometric normalization function determined by the solar zenith angle, the observed zenith angle, and the relative azimuth angle corresponding to the ocean optical remote sensing image; or alternatively, the solar zenith angle, the observed zenith angle, and the relative azimuth angle corresponding to the ocean optical remote sensing image are used as model input features, and together with the optical features, are used to establish the statistical correspondence. Using the center position of each wind vector element as a reference, determine the aggregation window corresponding to the wind vector element; Based on the geometrically normalized optical features corresponding to the effective sea surface pixels within the aggregation window, at least one of the statistical measures, such as the mean, quantiles, and weighted mean, is determined as the optical feature value of the wind vector element scale. The weighted mean is determined based on distance weight and / or time weight. The distance weight is determined based on the distance between the center positions of the effective sea surface pixel and the wind vector element. The time weight is determined based on the time difference between the optical observation time of the marine optical remote sensing image to which the effective sea surface pixel belongs and the scatterometer observation time corresponding to the wind vector element.

[0080] In some embodiments, the mapping module 240 is used to: According to the polarization mode and incident angle, the samples in the paired sample set are divided into layers or bins; Within each layer or each sub-box, the statistical correspondence is established based on the mapping relationship between the wind vector element-scale optical eigenvalues ​​and the observed backscattering coefficients; Based on the statistical correspondence, the reference value of the backscattering coefficient corresponding to the optical eigenvalue at the wind vector element scale is determined; The statistical correspondence is established through at least one of linear regression, piecewise regression, monotonic constrained regression, nonlinear regression, binning statistics, and robust regression; the relative azimuth angle in the backscattering coefficient observation data of the scatterometer is used to participate in the stratification or binning, or as an input parameter for establishing the statistical correspondence.

[0081] In some embodiments, the determining module 260 is used to: The deviation term is determined based on the difference between the observed backscattering coefficient value and the reference backscattering coefficient value; The backscattering coefficient reference value and / or the deviation term are used as the external statistical reference quantity; Based on the external statistical reference, an objective function is constructed to estimate the bias parameter; wherein, the objective function includes the sum of residual loss values ​​corresponding to multiple samples respectively, the residual loss value is determined based on the residual between the candidate corrected backscattering coefficient and the reference value of the backscattering coefficient, and the candidate corrected backscattering coefficient is the backscattering coefficient obtained by correcting the observed value of the backscattering coefficient based on the bias parameter to be determined; The bias parameters are determined by minimizing the objective function; The bias parameter is used to correct the observed backscattering coefficients according to the following model: ; in, This represents the calibrated backscattering coefficient. This represents the observed value of the backscattering coefficient. Represents the multiplicative coefficient, This indicates the additive bias. Represents the geometrically related system term, Indicates the incident angle, Indicates the relative azimuth angle, This indicates the polarization mode.

[0082] The processing flow of each module in the device and the interaction flow between each module can be referred to the relevant descriptions in the above method embodiments, and will not be detailed here.

[0083] This application also provides a computer device, such as... Figure 3 The diagram shown is a schematic representation of a computer device structure provided in an embodiment of this application, including: Processor 31 and memory 32; the memory 32 stores machine-readable instructions executable by the processor 31, and the processor 31 executes the machine-readable instructions stored in the memory 32. When the machine-readable instructions are executed by the processor 31, the processor 31 performs the following steps: Acquire marine optical remote sensing images and scatterometer backscattering coefficient observation data that meet preset matching conditions, and establish a paired sample set; wherein, the preset matching conditions include temporal matching conditions and spatial matching conditions, and the scatterometer backscattering coefficient observation data includes backscattering coefficient observation values, incident angle, relative azimuth angle, polarization mode, and wind vector element location information; The marine optical remote sensing image is filtered for effective pixels to obtain an effective sea surface pixel set, and optical features characterizing sea surface roughness changes are extracted based on the effective sea surface pixel set; wherein, the optical features include at least one of white cap features, flare features, and sea surface reflection texture features; Based on the observation geometric information corresponding to the marine optical remote sensing image, the optical features are geometrically normalized, and the geometrically normalized optical features are aggregated to the scale corresponding to the wind vector element to obtain the wind vector element scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer. Based on the optical eigenvalues ​​at the wind vector element scale and the backscattering coefficient observation data of the scatterometer, a statistical correspondence between the optical eigenvalues ​​and the observed backscattering coefficients is established, and a reference value for the backscattering coefficient is determined based on the statistical correspondence. Based on the backscattering coefficient reference value and the observed backscattering coefficient value, an external statistical reference quantity is constructed to characterize the system bias structure of the observed backscattering coefficient value; Based on the external statistical reference, bias parameters are determined for on-orbit calibration of the backscattering coefficient observations; wherein, the bias parameters include at least one of additive bias, multiplicative coefficients, and geometrically related system terms associated with the incident angle, the relative azimuth angle, and the polarization mode; The observed backscattering coefficients are corrected based on the bias parameters to obtain the calibrated backscattering coefficients.

[0084] The aforementioned memory 32 includes a main memory 321 and an external memory 322; the main memory 321, also known as internal memory, is used to temporarily store the computational data in the processor 31, as well as the data exchanged with external memory 322 such as a hard disk. The processor 31 exchanges data with the external memory 322 through the main memory 321.

[0085] The specific execution process of the above instructions can be referred to the backscattering coefficient of the scatterometer based on optical remote sensing characteristics described in the embodiments of this application. The steps for in-orbit calibration will not be elaborated here.

[0086] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program performs the backscattering coefficient of a scatterometer based on optical remote sensing characteristics described in the above-described method embodiments. The steps of the on-orbit calibration method. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0087] This application also provides a computer program product carrying program code. The program code includes instructions that can be used to execute the backscattering coefficient of a scatterometer based on optical remote sensing characteristics described in the above method embodiments. The steps of the on-orbit calibration method can be found in the above-described method implementation examples, and will not be repeated here.

[0088] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0089] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0092] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0093] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the technical scope disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A scatterometer backscattering coefficient based on optical remote sensing characteristics The on-orbit calibration method is characterized by... include: Acquire marine optical remote sensing images and scatterometer backscattering coefficients that meet preset matching conditions. The observation data is used to establish a paired sample set; wherein, the preset matching conditions include time matching conditions and spatial matching conditions, and the backscattering coefficient observation data of the scatterometer includes the backscattering coefficient observation value, incident angle, relative azimuth angle, polarization mode, and wind vector element position information; The marine optical remote sensing image is filtered for effective pixels to obtain an effective sea surface pixel set, and optical features characterizing sea surface roughness changes are extracted based on the effective sea surface pixel set; wherein, the optical features include at least one of white cap features, flare features, and sea surface reflection texture features; Based on the observation geometric information corresponding to the marine optical remote sensing image, the optical features are geometrically normalized, and the geometrically normalized optical features are aggregated to the scale corresponding to the wind vector element to obtain the wind vector element scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer. Based on the optical eigenvalues ​​at the wind vector element scale and the backscattering coefficient observation data of the scatterometer, a statistical correspondence between the optical eigenvalues ​​and the observed backscattering coefficients is established, and a reference value for the backscattering coefficient is determined based on the statistical correspondence. Based on the backscattering coefficient reference value and the observed backscattering coefficient value, an external statistical reference quantity is constructed to characterize the system bias structure of the observed backscattering coefficient value; Based on the external statistical reference, bias parameters are determined for on-orbit calibration of the backscattering coefficient observations; wherein, the bias parameters include at least one of additive bias, multiplicative coefficients, and geometrically related system terms associated with the incident angle, the relative azimuth angle, and the polarization mode; The observed backscattering coefficients are corrected based on the bias parameters to obtain the calibrated backscattering coefficients.

2. The method according to claim 1, characterized in that, The process of acquiring marine optical remote sensing images and scatterometer backscattering coefficient observation data that meet preset matching conditions, and establishing a paired sample set, includes: For each wind vector element, determine the scatterometer observation time and center position corresponding to that wind vector element; Based on the time matching condition, a target optical remote sensing image is determined from the ocean optical remote sensing image whose time difference between the optical observation time and the scatterometer observation time is less than or equal to a preset time threshold. Based on the spatial matching conditions, the set of optical pixels corresponding to the wind vector element is determined from the target optical remote sensing image according to the center position of the wind vector element. Based on the wind vector element and the corresponding optical pixel set, the paired sample set is established.

3. The method according to claim 1, characterized in that, The effective pixel selection of the marine optical remote sensing image to obtain an effective sea surface pixel set includes: Based on the product quality marks and / or preset screening rules corresponding to the marine optical remote sensing images, generate pixel-level valid identifiers; Based on the pixel-level effective identifier, at least one of cloud pixels, cloud image pixels, land pixels, saturation pixels, stripe anomaly pixels, and observed geometric anomaly pixels in the marine optical remote sensing image is removed to obtain the effective sea surface pixel set. The method further includes: Based on the quality control information in the backscattering coefficient observation data of the scatterer, invalid backscattering coefficient observation data of the scatterer are removed; and / or, Backscattering coefficient observation data from scatterometers were excluded based on precipitation pollution identification; and / or, Determine the effective optical coverage for each wind vector element, and remove the corresponding wind vector element sample if the effective optical coverage is lower than the coverage threshold.

4. The method according to claim 1, characterized in that, The extraction of optical features characterizing sea surface roughness variations based on the effective sea surface pixel set includes at least one of the following: Based on the reflectance of each pixel in the effective sea surface pixel set in one or more visible light bands and / or near-infrared bands, the white hat discrimination value corresponding to each pixel is determined. Pixels whose white hat discrimination value meets the white hat discrimination condition are determined as white hat pixels. The white hat score is determined based on the number of white hat pixels and the number of effective sea surface pixels. Pixels that satisfy the flare observation geometry conditions are determined from the effective set of sea surface pixels. Based on the reflectance of the pixels that satisfy the flare observation geometry conditions, at least one of the following is determined: flare intensity statistics, flare gradient statistics, and flare region area fraction. Within a window of one or more spatial scales, at least one of the following is determined based on the reflectance of each pixel in the effective sea surface pixel set: local variance, spectral energy, and directional texture statistics.

5. The method according to claim 1, characterized in that, Based on the observation geometric information corresponding to the marine optical remote sensing image, the optical features are geometrically normalized, and the geometrically normalized optical features are aggregated to the scale corresponding to the wind vector element to obtain the wind vector element-scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer, including: Based on the solar zenith angle, the observed zenith angle, and the relative azimuth angle corresponding to the ocean optical remote sensing image, the samples corresponding to the optical features are geometrically binned, and the optical features are statistically analyzed or modeled within each geometric bin; alternatively, the optical features are normalized based on a geometric normalization function determined by the solar zenith angle, the observed zenith angle, and the relative azimuth angle corresponding to the ocean optical remote sensing image; or alternatively, the solar zenith angle, the observed zenith angle, and the relative azimuth angle corresponding to the ocean optical remote sensing image are used as model input features, and together with the optical features, are used to establish the statistical correspondence. Using the center position of each wind vector element as a reference, determine the aggregation window corresponding to the wind vector element; Based on the geometrically normalized optical features corresponding to the effective sea surface pixels within the aggregation window, at least one of the statistical measures, such as the mean, quantiles, and weighted mean, is determined as the optical feature value of the wind vector element scale. The weighted mean is determined based on distance weight and / or time weight. The distance weight is determined based on the distance between the center positions of the effective sea surface pixel and the wind vector element. The time weight is determined based on the time difference between the optical observation time of the marine optical remote sensing image to which the effective sea surface pixel belongs and the scatterometer observation time corresponding to the wind vector element.

6. The method according to claim 1, characterized in that, The process of establishing a statistical correspondence between the optical features at the wind vector element scale and the observed backscattering coefficient data from the scatterometer, and determining a reference value for the backscattering coefficient based on the statistical correspondence, includes: According to the polarization mode and incident angle, the samples in the paired sample set are divided into layers or bins; Within each layer or each sub-box, the statistical correspondence is established based on the mapping relationship between the wind vector element-scale optical eigenvalues ​​and the observed backscattering coefficients; Based on the statistical correspondence, the reference value of the backscattering coefficient corresponding to the optical eigenvalue at the wind vector element scale is determined; The statistical correspondence is established through at least one of linear regression, piecewise regression, monotonic constrained regression, nonlinear regression, binning statistics, and robust regression; the relative azimuth angle in the backscattering coefficient observation data of the scatterometer is used to participate in the stratification or binning, or as an input parameter for establishing the statistical correspondence.

7. The method according to claim 1, characterized in that, The step of constructing an external statistical reference quantity to characterize the systematic bias structure of the observed backscattering coefficients based on the reference value and the observed backscattering coefficients, and determining bias parameters for on-orbit calibration of the observed backscattering coefficients based on the external statistical reference quantity, includes: The deviation term is determined based on the difference between the observed backscattering coefficient value and the reference backscattering coefficient value; The backscattering coefficient reference value and / or the deviation term are used as the external statistical reference quantity; Based on the external statistical reference, an objective function is constructed to estimate the bias parameter; wherein, the objective function includes the sum of residual loss values ​​corresponding to multiple samples respectively, the residual loss value is determined based on the residual between the candidate corrected backscattering coefficient and the reference value of the backscattering coefficient, and the candidate corrected backscattering coefficient is the backscattering coefficient obtained by correcting the observed value of the backscattering coefficient based on the bias parameter to be determined; The bias parameters are determined by minimizing the objective function; The bias parameter is used to correct the observed backscattering coefficients according to the following model: ； in, This represents the calibrated backscattering coefficient. This represents the observed value of the backscattering coefficient. Represents the multiplicative coefficient, This indicates the additive bias. Represents the geometrically related system term, Indicates the incident angle, Indicates the relative azimuth angle, This indicates the polarization mode.

8. A backscattering coefficient of a scatterometer based on optical remote sensing characteristics The on-orbit calibration device is characterized in that... include: The acquisition module is used to acquire marine optical remote sensing images and scatterometer backscattering coefficients that meet preset matching conditions. The observation data is used to establish a paired sample set; wherein, the preset matching conditions include time matching conditions and spatial matching conditions, and the backscattering coefficient observation data of the scatterometer includes the backscattering coefficient observation value, incident angle, relative azimuth angle, polarization mode, and wind vector element position information; The filtering module is used to filter the effective pixels of the marine optical remote sensing image to obtain an effective sea surface pixel set, and extract optical features characterizing sea surface roughness changes based on the effective sea surface pixel set; wherein, the optical features include at least one of white hat features, flare features, and sea surface reflection texture features; The aggregation module is used to perform geometric normalization processing on the optical features based on the observation geometric information corresponding to the marine optical remote sensing image, and aggregate the geometrically normalized optical features to the scale corresponding to the wind vector element to obtain the wind vector element scale optical feature value corresponding to the backscattering coefficient observation data of the scatterometer. The mapping module is used to establish a statistical correspondence between the optical features and the observed backscattering coefficient based on the wind vector element scale optical feature value and the backscattering coefficient observation data of the scatterometer, and to determine the backscattering coefficient reference value based on the statistical correspondence. The construction module is used to construct an external statistical reference quantity to characterize the system bias structure of the backscattering coefficient observation value based on the backscattering coefficient reference value and the backscattering coefficient observation value; The determination module is used to determine, based on the external statistical reference, bias parameters for on-orbit calibration of the backscattering coefficient observations; wherein the bias parameters include at least one of additive bias, multiplicative coefficient, and geometrically related system terms associated with the incident angle, the relative azimuth angle, and the polarization mode; The correction module is used to correct the observed backscattering coefficient based on the bias parameter to obtain the calibrated backscattering coefficient.

9. A computer device, characterized in that, include: A processor and a memory, the memory storing machine-readable instructions executable by the processor, the processor executing the machine-readable instructions stored in the memory, wherein when the machine-readable instructions are executed by the processor, the processor performs the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a computer device, performs the steps of the method as described in any one of claims 1 to 7.

Images