A cross-calibration method for wide-swath imaging spectrometers based on radiative transfer simulation and observation angle deviation correction

By using a method based on radiative transfer simulation and observation angle deviation correction, the problems of synchronous nadir point dependence and angle deviation in cross-calibration of wide-swath imaging spectrometers were solved, achieving efficient cross-calibration and radiative consistency correction, and improving the collaborative application capability of multi-source remote sensing data.

CN123065593APending Publication Date: 2026-11-10SHANGHAI INSTITUTE OF TECHNICAL PHYSICS CHINESE ACADEMY OF SCIENCES
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611502050.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-24
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing cross-calibration methods for wide-swath imaging spectrometers rely on synchronous nadir observation data, which makes it difficult to frequently meet the acquisition conditions and cannot effectively handle radiation deviations over a wide angular range. This results in low calibration frequency and insufficient accuracy, limiting the collaborative application of multi-source remote sensing data.

Method used

By employing a method based on radiative transfer simulation and observation angle deviation correction, an empirical correction function is established through pixel-by-pixel forward modeling and statistical regression to compensate for systematic deviations between sensors, thereby achieving efficient cross-calibration of the wide-swath imaging spectrometer.

Benefits of technology

It improves the applicability and frequency of cross-calibration, enhances the radiation consistency of wide-angle sensors across the entire field of view, is suitable for a wider range of observation angles, and significantly improves the accuracy of quantitative applications.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123065593A_ABST
    Figure CN123065593A_ABST
Patent Text Reader

Abstract

This invention discloses a cross-calibration method for wide-swath imaging spectrometers based on radiative transfer simulation and observation angle deviation correction, belonging to the field of remote sensing satellite radiometric calibration technology. The method uses MODIS surface reflectance and aerosol optical thickness products as inputs to drive the MODTRAN radiative transfer model, performing pixel-by-pixel forward modeling of the apparent reflectance of the atmospheric top of the sensor to be calibrated. The simulated values ​​are compared pixel-by-pixel with the actual observed values ​​to calculate the residuals. The variation of the residuals with the observed zenith angle is analyzed, and a quadratic polynomial empirical correction function is established through statistical regression. This function is applied to the original observation data to achieve angle-dependent compensation for systematic deviations between sensors. This invention does not rely on synchronous nadir observation data and can effectively correct angle-related radiative deviations caused by atmospheric path differences and the BRDF effect, significantly improving the cross-calibration accuracy and applicability of wide-swath imaging spectrometers across the entire field of view.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of remote sensing satellite radiometric calibration technology, specifically relating to a cross-calibration method for wide-swath imaging spectrometers based on radiative transfer simulation and observation angle deviation correction. Background Technology

[0002] With the continuous development of Earth observation remote sensing technology, wide-swath imaging spectrometers, capable of covering large observation areas in a single pass, have become an important data source for global-scale surface monitoring and climate change research. However, these sensors typically have a field of view of ±50° or even larger, resulting in satellite observation zenith angle differences of tens of degrees between nadir points and edge pixels within the same image. This drastic change in observation zenith angle has two significant impacts: first, the path length of radiation propagation in the atmosphere increases rapidly with the observation zenith angle, significantly enhancing atmospheric attenuation, scattering, and absorption effects, causing varying degrees of atmospheric modulation of the apparent reflectance of pixels at different angles; second, the bidirectional reflection distribution function (BRDF) effect leads to fundamental differences in reflectance of the same ground object in different observation directions. The combination of these factors means that the observed radiation values ​​within wide-swath sensor images are not only controlled by the true reflectance of the ground object and the sensor calibration coefficients, but also strongly dependent on the observation angle, severely limiting the accuracy of quantitative applications of wide-swath remote sensing data and the collaborative analysis capabilities of multi-source remote sensing data.

[0003] Currently, cross-calibration is a common method for solving the problem of radiometric consistency between different sensors. Traditional cross-calibration methods typically rely on synchronous nadir observation data from two sensors. By selecting pixel pairs that are spatially and temporally matched and have similar observational geometry, a statistical regression relationship is established between the observations of the two sensors. However, this method has several significant drawbacks.

[0004] First, the conditions for acquiring synchronous nadir point observation data are extremely demanding. The two sensors need to pass over each other within a similar time frame and their observation areas need to overlap. At the same time, the difference in the observed zenith angle is required to be small. These conditions are difficult to meet frequently in actual satellite operation, resulting in a limited number of available calibration samples, low calibration frequency, and an inability to effectively track changes in the sensor's radiation response.

[0005] Second, traditional methods handle angular differences rather crudely. When there is a significant difference in the zenith angle observed by two sensors, the differences in atmospheric path length and the bidirectional reflection effect of the surface will introduce significant radiation bias. Especially for wide-swath imaging spectrometers, the range of observed zenith angles within a single image is extremely large, making pixel-by-pixel angle matching with the reference sensor extremely difficult. Even if pixels with relatively small angular differences are selected for calibration, there are still significant systematic radiation differences between the nadir point and edge pixels. Directly using a single linear regression model cannot effectively characterize this nonlinear bias that varies with angle, limiting the applicable angular range of cross-calibration.

[0006] Third, while some existing methods employ forward modeling based on radiative transfer models for cross-calibration, most simply compare simulation results with observed values ​​or perform linear regression. They fail to systematically analyze the intrinsic relationship between simulation residuals and observed geometric parameters, nor do they establish targeted empirical correction models to compensate for systematic biases. This is especially true for wide-swath sensors, where the large angular span within a single image often leads to non-linear variations in residuals with the observed zenith angle. Ignoring this trend will introduce significant systematic errors.

[0007] In addition, some researchers have attempted to normalize the observation angle using the BRDF model. However, the BRDF model itself relies on multi-angle observation data and its applicability varies across different land cover types, leading to significant uncertainties in practical applications. Therefore, there is an urgent need for a cross-calibration method that does not rely on synchronous nadir observations and can effectively handle large angular variations within a wide-swath sensor. This would improve the applicability, frequency, and accuracy of cross-calibration, supporting the quantitative collaborative application of multi-source remote sensing data. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a cross-calibration method for wide-swath imaging spectrometers based on radiative transfer simulation and observation angle deviation correction. By comparing forward modeling with actual observations pixel by pixel, the method analyzes the variation law of residuals with the observed zenith angle and establishes an empirical correction function, thereby achieving efficient compensation for systematic deviations between sensors.

[0009] To achieve the above objectives, the present invention adopts the following technical solution:

[0010] A cross-calibration method for wide-swath imaging spectrometers based on radiative transfer simulation and observation angle deviation correction includes the following steps:

[0011] Step S1: Acquire image data of the target area of ​​the sensor to be calibrated. The image data includes the observed apparent reflectance of each pixel, imaging time, latitude and longitude coordinates, solar zenith angle and satellite observation zenith angle information.

[0012] Step S2: Obtain the surface reflectance product MCD43C4 and aerosol optical depth (AOD) product MOD08D3 of the reference sensor MODIS (ModerateResolution Imaging Spectroradiometer) that match the time and region of the sensor to be calibrated, and extract the surface reflectance and AOD information corresponding to the pixels of the sensor to be calibrated from them.

[0013] Step S3: Based on the MODTRAN (MODerate resolution atmospheric TRANsmission) radiative transfer model, simulate radiance curves, atmospheric transmittance, and other information under different observation geometries and atmospheric conditions, and establish an atmospheric parameter lookup table. Using the surface reflectance and aerosol optical thickness information obtained in Step S2, simulate the apparent reflectance of the sensor to be calibrated.

[0014] Step S4: Analyze the relationship between the apparent reflectance observation value of the sensor to be calibrated obtained in step S1, the apparent reflectance of the sensor to be calibrated simulated in step S3, and the satellite observation angle of each pixel of the sensor to be calibrated obtained in step S2. Establish an empirical correction function through statistical regression method to compensate and correct the apparent reflectance observation value of each pixel of the sensor to be calibrated.

[0015] Further, step S1 includes:

[0016] Step S11: Obtain image data and auxiliary parameters of the target area of ​​the sensor to be calibrated, and extract the original digital quantization (DN) value, imaging time, latitude and longitude coordinates, solar zenith angle and satellite observation zenith angle information of each pixel;

[0017] Step S12: Perform radiometric calibration and cloud detection preprocessing on the acquired image data. Radiometric calibration is performed on the image data acquired in Step S11, converting DN values ​​to apparent reflectance; and cloud detection processing is performed to generate a cloud mask file, marking cloud, water, and ice / snow pixels in the image for subsequent filtering and exclusion.

[0018] Step S13: Filter valid pixels based on cloud detection results and construct a dataset to be calibrated. Based on the cloud mask file generated in step S12, remove cloud, water, and ice / snow pixels, retain valid land pixels under clear sky conditions, and construct a dataset to be calibrated; at the same time, record the imaging time, latitude and longitude coordinates, solar zenith angle, and satellite observation zenith angle information corresponding to each valid pixel as input parameters for subsequent radiative transfer simulation.

[0019] Further, step S2 includes:

[0020] Step S21: Acquire the MODIS surface reflectance product MCD43C4 and the global aerosol optical thickness product MOD08D3 that match the target area and imaging time of the sensor to be calibrated, and use them as surface end-member input data and atmospheric state input data, respectively.

[0021] Step S22: Spatial matching is performed between the reference sensor product data and the pixels of the sensor to be calibrated, and the surface reflectance and aerosol optical thickness information corresponding to each pixel are extracted. Using the latitude and longitude coordinates and imaging time of each effective pixel in the dataset to be calibrated in step S13 as the matching benchmark, the surface reflectance value of the corresponding band of each pixel is extracted from the MCD43C4 product, and the aerosol optical thickness value of each pixel is extracted from the MOD08D3 product, to construct a reference dataset that corresponds one-to-one with the pixels to be calibrated.

[0022] Further, step S3 includes:

[0023] Step S31: Using the MODTRAN radiative transfer model as the simulation engine, set the value range and step size of parameters such as different atmospheric modes, solar zenith angle, satellite observation zenith angle, and aerosol optical thickness, and drive MODTRAN to simulate the radiance curves and atmospheric transmittance information under various parameter combinations one by one, and construct an atmospheric parameter lookup table covering various observation geometries and atmospheric conditions.

[0024] Step S32: Determine the atmospheric model based on the imaging time and latitude and longitude coordinates of each effective pixel of the sensor to be calibrated obtained in step S13. Combine the actual observation geometric parameters (solar zenith angle, satellite observation zenith angle) and the atmospheric state parameters (aerosol optical thickness) obtained in S22 as the lookup index, perform multidimensional interpolation in the atmospheric parameter lookup table established in step S31 to obtain the atmospheric correction parameters corresponding to each pixel. Then, combine the surface reflectance extracted in step S22 to calculate the simulated value of the apparent reflectance of the top of the atmosphere of each pixel in the corresponding band through the radiative transfer equation.

[0025] Further, step S4 includes:

[0026] Step S41: Pair the apparent reflectance of each pixel at the top of the atmosphere obtained from step S32 with the actual apparent reflectance of each pixel obtained after radiometric calibration in step S12, and calculate the residual value between the two.

[0027] Step S42: Take the (observed zenith angle, residual value) data pairs of all valid pixels obtained in step S41 as a sample set, and use the satellite observed zenith angle of each pixel as the independent variable and the corresponding residual as the dependent variable to statistically analyze the variation of the residual with the observed zenith angle. Preferably, a grouped statistical method is used: the pixels are grouped according to the observed zenith angle interval, the mean and standard deviation of the residuals of each group are calculated, and a scatter plot or curve of the residuals changing with the observed zenith angle is plotted.

[0028] Step S43: Select a quadratic polynomial, use the least squares method to perform curve fitting on the sample data, determine the function parameters through regression, and obtain the empirical correction function;

[0029] Step S44: Apply the empirical correction function established in step S43 to the original observed apparent reflectance data of the sensor to be calibrated, and perform compensation correction on the apparent reflectance observation value of each effective pixel to obtain the corrected apparent reflectance.

[0030] In a second aspect, the present invention provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction.

[0031] Thirdly, the present invention provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enable the processor to implement the aforementioned method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction.

[0032] The beneficial effects of this invention are as follows:

[0033] (1) This invention does not rely on the synchronous nadir observation data of two sensors. It only requires the surface reflectance product and aerosol product of the reference sensor as input to achieve cross-calibration of the sensor to be calibrated, which greatly reduces the dependence of cross-calibration on data conditions and improves the operability and frequency of calibration.

[0034] (2) This invention uses pixel-by-pixel radiative transfer forward modeling to finely consider the differences in observation geometry, atmospheric state and surface features of each pixel, which has higher simulation accuracy than the traditional regional averaging method.

[0035] (3) This invention establishes an empirical correction function for angular deviation by systematically analyzing the variation law of simulation residuals with the observed zenith angle. This function is particularly effective in compensating for nonlinear radiation deviation caused by the large span of the observed zenith angle in a single image of a wide-swath imaging spectrometer (ranging from 0° to more than 60° from the nadir point to the edge). It can effectively correct the angle-related systematic radiation deviation caused by the difference in atmospheric path length and the BRDF effect on the ground surface, making the cross-calibration method applicable to a wider range of observation angles and significantly improving the radiation consistency of the wide-swath sensor across the entire field of view.

[0036] (4) This invention organically combines the radiative transfer physical model with the statistical regression method, which has both the rationality of the physical mechanism and the flexibility of statistical correction. It can be widely used for radiation consistency correction between multispectral sensors on different satellite platforms. Attached Figure Description

[0037] Figure 1 This is a flowchart of a cross-calibration method for a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction according to the present invention.

[0038] Figure 2 This is a scatter plot of the residual between the actual apparent reflectance and the simulated apparent reflectance of this invention as a function of the satellite-observed zenith angle.

[0039] Figure 3 This is a curve showing the fitting of the mean of the residuals between the actual apparent reflectance and the simulated apparent reflectance of this invention, after calculating the mean and standard deviation of the residuals in each interval, and the variation of the mean of the residuals in each interval with the zenith angle observed by the satellite. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of the invention described below can be combined with each other as long as they do not conflict with each other. The invention will be further described below with reference to the accompanying drawings and specific embodiments, including illustrative examples and descriptions. This is merely an explanation of the invention and is not intended to limit it.

[0041] like Figure 1 As shown, the present invention proposes a cross-calibration method for wide-swath imaging spectrometers based on radiative transfer simulation and observation angle deviation correction, comprising the following steps:

[0042] Step S1: Acquire image data of the target area of ​​the sensor to be calibrated. The image data includes the observed apparent reflectance of each pixel, imaging time, latitude and longitude coordinates, solar zenith angle and satellite observation zenith angle information.

[0043] Step S2: Obtain the surface reflectance product MCD43C4 and AOD product MOD08D3 of the reference sensor MODIS that match the time and region of the sensor to be calibrated, and extract the surface reflectance and AOD information corresponding to the pixels of the sensor to be calibrated.

[0044] Step S3: Based on the MODTRAN radiative transfer model, simulate radiance curves, atmospheric transmittance, and other information under different observation geometries and atmospheric conditions, and establish an atmospheric parameter lookup table. Using the surface reflectance and aerosol optical thickness obtained in Step S2, simulate the apparent reflectance of the sensor to be calibrated;

[0045] Step S4: Analyze the relationship between the apparent reflectance observation value of the sensor to be calibrated obtained in Step S1, the simulated apparent reflectance of the sensor to be calibrated obtained in Step S3, and the satellite observation angle of each pixel of the sensor to be calibrated obtained in Step S2. Establish an empirical correction function through statistical regression and compensate for the apparent reflectance observation value of each pixel of the sensor to be calibrated.

[0046] A further preferred embodiment of the present invention is that step S1 includes:

[0047] Step S11: Acquire image data and auxiliary parameters of the target area of ​​the sensor to be calibrated. Acquire image data of the target area of ​​the sensor to be calibrated, and extract the original DN value, imaging time, latitude and longitude coordinates, and solar zenith angle of each pixel. and satellite observation zenith angle information;

[0048] Step S12: Perform radiometric calibration and cloud detection preprocessing on the acquired image data. Perform radiometric calibration on the image data acquired in step S11, converting the DN values ​​into sensor entrance pupil radiance.

[0049] For the visible and near-infrared short-wavelength spectral bands, the DN value is converted to radiance using the following formula:

[0050] ,

[0051] in, The equivalent radiance at the entrance pupil of the satellite payload channel, in units of Gain is the scaling factor gain, in units of... Bias is the scaling factor bias, in units of , Indicates wavelength.

[0052] Use the following formula to convert radiance to reflectance:

[0053] ,

[0054] Where λ is the wavelength. This represents the actual observed value of apparent reflectance. The zenith angle of the sun. The solar irradiance at the top of the atmosphere can be calculated using the following formula based on the reference solar spectrum and the Earth-Sun distance:

[0055] ,

[0056] in, Let be the average solar irradiance when the Earth-Sun distance is 1, which is a constant. D is the Earth-Sun distance, which can be calculated from the accumulated solar radiation over an annual period. Estimate.

[0057] For the infrared spectral band, the DN value is converted to radiance using the following formula:

[0058] ,

[0059] in, , , These are the calibration coefficients. Typically, for the infrared spectrum, Planck's formula is also needed to convert radiance to brightness temperature. :

[0060] ,

[0061] Among them, intermediate quantity , h is Planck's constant, 6.626. 10 -34 J s; c is the speed of light, 2.998 10 8 m s -1 k is the Boltzmann constant, 1.381. 10 -23 J K -1 .

[0062] Perform cloud detection processing to generate cloud mask files, marking cloud, water, and ice / snow pixels in the image for subsequent filtering and exclusion;

[0063] Step S13: Filter valid pixels based on cloud detection results and construct a dataset to be calibrated. Based on the cloud mask file generated in Step S12, remove cloud, water, and ice / snow pixels, retaining valid land pixels under clear sky conditions to construct the dataset to be calibrated; simultaneously record the imaging time, latitude and longitude coordinates, and solar zenith angle corresponding to each valid pixel. and satellite observation zenith angle Information is used as input parameters for subsequent radiative transfer simulations.

[0064] A further preferred embodiment of the present invention is that step S2 includes the following steps:

[0065] Step S21: Acquire the MODIS surface reflectance product MCD43C4 and the global aerosol optical thickness product MOD08D3 that match the target area and imaging time of the sensor to be calibrated, and use them as surface end-member input data and atmospheric state input data, respectively.

[0066] Step S22: Using the latitude and longitude coordinates and imaging time of each effective pixel in the dataset to be calibrated in Step S13 as the matching benchmark, extract the surface reflectance values ​​of the corresponding bands of each pixel from the MCD43C4 product. The aerosol optical thickness value τ corresponding to each pixel is extracted from the MOD08D3 product to construct a reference dataset that corresponds one-to-one with the pixels to be calibrated.

[0067] A further preferred embodiment of the present invention is that step S3 includes the following steps:

[0068] Step S31: Using the MODTRAN radiative transfer model as the simulation engine, set the value range and step size of parameters such as different atmospheric models, solar zenith angle, satellite observation zenith angle, and aerosol optical thickness. Drive MODTRAN to simulate the radiance curves and atmospheric transmittance information under various parameter combinations one by one, and construct an atmospheric parameter lookup table covering various observation geometries and atmospheric conditions.

[0069] ,

[0070] in, This is atmospheric path radiation, which is the portion of light that is scattered back to the satellite by the atmosphere without reaching the ground. This part of the signal mainly carries aerosol information. Total atmospheric transmittance; The atmospheric hemispherical albedo; This is an atmospheric model, determined by pixel imaging time and latitude / longitude coordinates; The solar zenith angle; For satellite observation of zenith angle; It is AOD.

[0071] Step S32: Determine the atmospheric model based on the imaging time and latitude / longitude coordinates of each effective pixel of the sensor to be calibrated obtained in step S13. Combined with actual observational geometric parameters (solar zenith angle) Satellite observation zenith angle Using the atmospheric state parameters (aerosol optical thickness τ) obtained in step S22 as the lookup index, multidimensional interpolation is performed in the atmospheric parameter lookup table established in step S31 to obtain the atmospheric correction parameters corresponding to each pixel. Combined with the surface reflectance extracted in step S22 Simulated values ​​of the apparent reflectance at the top of the atmosphere for each pixel in the corresponding band were calculated using the radiative transfer equation:

[0072] ;

[0073] in, This represents the simulated apparent reflectance of a pixel at the top of the atmosphere in the corresponding band.

[0074] A further preferred embodiment of the present invention is that step S4 includes the following steps:

[0075] Step S41: Calculate the apparent reflectance of each pixel at the top of the atmosphere obtained from step S32. The actual observed apparent reflectance of each pixel obtained after S12 radiometric calibration Perform pixel-by-pixel pairing and calculate the residual value between them:

[0076] ,

[0077] in, Let i be the residual of the i-th pixel. Let be the actual TOA reflectance of the i-th pixel. Let be the simulated TOA reflectance of the i-th pixel.

[0078] Step S42: Calculate the zenith angle of all valid pixels obtained in step S41. residual value The data pairs are used as a sample set, with the satellite observation zenith angle of each pixel as the independent variable and the corresponding residual as the dependent variable. The variation of the residual with the observation zenith angle is statistically analyzed. Preferably, a grouped statistical method is used: pixels are grouped according to the observation zenith angle interval, the mean and standard deviation of the residuals of each group are calculated, and a scatter plot or curve of the residuals as a function of the observation zenith angle is plotted.

[0079] Step S43: Select a quadratic polynomial, use the least squares method to perform curve fitting on the sample data, determine the function parameters through regression, and obtain the empirical correction function:

[0080] ,

[0081] in, To observe the zenith angle The corresponding residual prediction values; m, n, and p are the empirical coefficients of the fitted quadratic polynomial.

[0082] Step S44: Apply the empirical correction function established in step S43 to the raw observed apparent reflectance data of the sensor to be calibrated, and compensate and correct the actual observed value of each effective pixel to obtain the corrected apparent reflectance:

[0083] ,

[0084] in, Let be the TOA reflectance of the i-th pixel after correction.

[0085] Figure 2 This refers to the observation zenith angle of all effective pixels in the embodiments of the present invention. residual value Using data pairs as the sample set, a scatter plot of the residuals versus the observed zenith angle is plotted, with the satellite-observed zenith angle of each pixel as the independent variable and the corresponding residuals as the dependent variable. The horizontal axis represents the satellite-observed zenith angle in degrees (°), and the vertical axis represents the dimensionless residuals obtained by subtracting the simulated apparent reflectance from the actual apparent reflectance of the satellite. The residuals approximately follow a quadratic polynomial distribution with respect to the solar zenith angle.

[0086] Figure 3 This embodiment of the invention employs a grouped statistical method, grouping pixels according to the observed zenith angle interval, calculating the mean and standard deviation of the residuals for each group, plotting the variation of the mean residuals with the observed zenith angle, and fitting the data using a quadratic polynomial. The fitting effect is good, with R... 2 It is 0.9846.

[0087] In a second aspect, the present invention provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction.

[0088] Thirdly, the present invention provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enable the processor to implement the aforementioned method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction.

[0089] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A cross-calibration method for wide-swath imaging spectrometers based on radiative transfer simulation and observation angle deviation correction, characterized in that, include: Step S1: Obtain image data of the target area of ​​the sensor to be calibrated. The image data includes the observed apparent reflectance of each pixel, imaging time, latitude and longitude coordinates, solar zenith angle and satellite observation zenith angle information. Step S2: Obtain the surface reflectance product and aerosol optical thickness product of the reference sensor that matches the time and region of the sensor to be calibrated, and extract the surface reflectance and aerosol optical thickness information corresponding to each pixel from them. Step S3: Based on the radiative transfer model, simulate the radiance curves and atmospheric transmittance information under different observation geometry and atmospheric conditions, establish an atmospheric parameter lookup table, and use the surface reflectance and aerosol optical thickness information obtained in step S2 to simulate the apparent reflectance simulation value of the sensor to be calibrated. Step S4: Analyze the observed apparent reflectance values ​​of the sensor to be calibrated obtained in step S1, the simulated apparent reflectance values ​​of the sensor to be calibrated obtained in step S3, and the satellite observation zenith angle information of each pixel of the sensor to be calibrated obtained in step S1. Establish an empirical correction function through statistical regression to compensate and correct the observed apparent reflectance values ​​of each pixel of the sensor to be calibrated obtained in step S1.

2. The method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in claim 1, characterized in that, Step S1 specifically includes: acquiring image data and auxiliary parameters of the target area of ​​the sensor to be calibrated; extracting the original digital quantization value, imaging time, latitude and longitude coordinates, solar zenith angle and satellite observation zenith angle information of each pixel; performing radiometric calibration processing on the acquired image data, converting the digital quantization value into an apparent reflectance observation value, and performing cloud detection processing to generate a cloud mask file; filtering effective pixels based on the cloud detection results, constructing a dataset to be calibrated, and simultaneously recording the imaging time, latitude and longitude coordinates, solar zenith angle and satellite observation zenith angle information corresponding to each effective pixel.

3. The method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in claim 1, characterized in that, Step S2 specifically includes: acquiring the surface reflectance product and global aerosol optical thickness product of the reference sensor that match the target area and imaging time of the sensor to be calibrated, and using them as the surface end-member input data and atmospheric state input data, respectively; using the latitude and longitude coordinates and imaging time of each effective pixel in the dataset to be calibrated as the matching benchmark, extracting the surface reflectance value of the corresponding band of each pixel from the surface reflectance product, and extracting the aerosol optical thickness value of each pixel from the aerosol optical thickness product, to construct a reference dataset that corresponds one-to-one with the pixels to be calibrated.

4. The method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in claim 1, characterized in that, Step S3 specifically includes: using the radiative transfer model as a simulation engine, setting different atmospheric modes, solar zenith angle, satellite observation zenith angle, and aerosol optical thickness parameters with varying value ranges and step sizes, driving the radiative transfer model to simulate radiance curves and atmospheric transmittance information under various parameter combinations, and constructing an atmospheric parameter lookup table covering multiple observation geometries and atmospheric conditions; determining the atmospheric mode based on the imaging time and latitude / longitude coordinates of each effective pixel of the sensor to be calibrated, using the solar zenith angle, satellite observation zenith angle obtained from step S1, and aerosol optical thickness obtained from step S2 as query indexes, performing multidimensional interpolation in the atmospheric parameter lookup table to obtain the atmospheric correction parameters corresponding to each pixel, and then, combining the surface reflectance extracted from step S2, calculating the simulated value of the apparent reflectance of the top of the atmosphere for each pixel in the corresponding band using the radiative transfer equation.

5. The method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in claim 2, characterized in that, Step S4 specifically includes: pairing the simulated apparent reflectance values ​​of each pixel at the top of the atmosphere obtained in step S3 with the observed apparent reflectance values ​​of each pixel obtained after radiometric calibration in step S1, and calculating the residual values ​​between the two; using the data pairs of satellite observation zenith angles and corresponding residual values ​​of all effective pixels as a sample set, statistically analyzing the variation law of residuals with the observed zenith angle, with the satellite observation zenith angle of each pixel as the independent variable and the corresponding residual as the dependent variable; using a quadratic polynomial and the least squares method to perform curve fitting on the sample data, regression to determine the function parameters, and obtaining an empirical correction function; applying the empirical correction function to the observed apparent reflectance values ​​of the sensor to be calibrated obtained in step S1, and compensating and correcting the observed apparent reflectance values ​​of each effective pixel to obtain the corrected apparent reflectance.

6. The method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in claim 5, characterized in that, The statistical analysis of the variation of residuals with the observed zenith angle specifically adopts a grouped statistical method: the pixels are grouped according to the observed zenith angle interval, the mean and standard deviation of the residuals of each group are calculated, and a scatter plot or curve of the residuals as a function of the observed zenith angle is plotted.

7. The method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in claim 5, characterized in that, The residual value is the difference between the observed apparent reflectance value and the simulated apparent reflectance value of each pixel; the quadratic polynomial is expressed as a quadratic function of the observed zenith angle, including quadratic coefficients, linear coefficients and a constant term.

8. The method for cross-calibration of a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in claim 1, characterized in that, The compensation correction specifically involves substituting the satellite observation zenith angle corresponding to each effective pixel into the empirical correction function to calculate the residual prediction value, and then subtracting the residual prediction value from the apparent reflectance observation value of that pixel to obtain the corrected apparent reflectance.

9. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When one or more programs are executed by the one or more processors, the one or more processors implement the cross-calibration method for a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed by a processor, enable the processor to implement the cross-calibration method for a wide-swath imaging spectrometer based on radiative transfer simulation and observation angle deviation correction as described in any one of claims 1-8.