A method and apparatus for spectral calibration and evaluation of very long-wavelength stars based on a radiative transfer model.

Abstract

This invention discloses a method and apparatus for evaluating on-board very long wave (VLW) spectral calibration based on a radiative transfer model, belonging to the field of remote sensing optics technology. This invention targets a wide-band VLW infrared hyperspectral sensor covering the wavelength range of 12-16 μm, analyzing the atmospheric absorption characteristics and sensor imaging spectral characteristics within this spectral range. The spectral calibration accuracy is evaluated through physical model simulation, including: simulation using the MODTRAN radiative transfer model; construction of a spectral drift model; solving for the center wavelength shift based on a spectral matching combined with an iterative optimization algorithm; and uncertainty analysis of the impact of factors such as the estimation error of the algorithm itself, sensor signal-to-noise ratio, inter-spectral response inconsistency, and atmospheric CO2 concentration on the estimation results. This invention is applied to evaluate the spectral calibration accuracy of on-orbit satellite data.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing optical technology, and specifically relates to a method and apparatus for very long-wavelength on-board spectral calibration and evaluation based on a radiative transfer model. Background Technology

[0002] Any object above absolute zero emits thermal radiation, the spectral distribution of which varies with temperature. The thermal radiation energy of objects at normal and low temperatures is primarily concentrated in the infrared band. In recent years, the strategic importance of the polar cryosphere has gradually increased, leading to a growing demand for polar observation. Due to the unique geographical environment and extreme climate conditions in the polar regions, the temperature of targets can reach minus 100 degrees Celsius. According to Planck's law, deep-cryogenic targets often have very weak radiation energy density, and the peak wavelength of radiation redshifts towards longer wavelengths as the temperature decreases. Very long wavelength infrared (VLWIR) (12-16 μm) hyperspectral imaging technology has become a key research tool for deep-cryogenic target detection. After satellite launch, during its on-orbit operation, the spectral performance of the sensor is affected by factors such as mechanical vibration, ambient radiation from space, and sunlight, resulting in attenuation. The most direct manifestation of this is the drift of the center wavelength of each sensor channel. Long-term on-orbit monitoring and spectral calibration of the VLWIR spaceborne infrared sensor are therefore essential.

[0003] Currently, the development of high spatial resolution, wide-spectrum, and wide-swath imaging very long-wavelength hyperspectral infrared sensors in China is still in its early stages. Although infrared calibration technology is relatively mature, internationally, infrared payloads have not yet broken through the calibration technology in the very long-wavelength cryogenic region. Spectral calibration, as a crucial part of calibration, also urgently needs further research. On-orbit spectral calibration methods mainly include onboard standard lamp spectral calibration and spectral matching calibration based on gas absorption characteristic peaks. The onboard standard lamp spectral calibration method requires the necessary instruments and equipment to be carried onboard, which places high demands on payload development. The spectral matching calibration method based on gas absorption characteristic peaks is one of the most commonly used methods, and it has shown superior performance on hyperspectral sensors with high spectral resolution. However, for hyperspectral sensors operating in the very long-wavelength range with limited spectral resolution, the application effect of this method lacks relevant research. Summary of the Invention

[0004] To address the lack of spectral calibration methods for broadband VL-IR hyperspectral sensors in existing technologies, this invention provides a method and apparatus for on-board VL-IR spectral calibration and evaluation based on a radiative transfer model. This invention targets broadband VL-IR hyperspectral sensors covering a wavelength range of 12-16 μm, analyzing atmospheric absorption characteristics and sensor imaging spectral characteristics within this spectral range, and evaluating the spectral calibration accuracy through physical model simulation.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model includes the following steps:

[0007] S1: The MODTRAN (MODerate spectral resolution radiative TRANsportmodel) radiative transfer model was used to simulate the radiance curve in the 12-16μm spectral range.

[0008] S2: Construct a spectral drift model for a very long wavelength hyperspectral sensor. Given a center wavelength shift, simulate the spectral response function of the very long wavelength hyperspectral sensor after spectral drift. Convolve the radiance curve obtained in S1 with the spectral response functions before and after spectral drift to obtain the reference spectrum and the measured spectrum.

[0009] S3: Perform spectral matching between the reference spectrum obtained in S2 and the measured spectrum, estimate the center wavelength offset using an iterative optimization algorithm, compare it with the center wavelength offset given in S2, and evaluate the accuracy.

[0010] S4: Conduct uncertainty analysis based on the estimation error of the algorithm itself, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the impact of atmospheric CO2 concentration on the estimation results;

[0011] S5: Applied to actual satellite data measured by atmospheric infrared detectors to further evaluate the accuracy of the method.

[0012] Further, S1 includes:

[0013] In the MODTRAN radiative transfer model, sensor parameters are set, including observation geometry, observation time, band range, and spectral resolution; atmospheric parameters are set, including aerosol optical thickness and water vapor content; surface characteristic parameters are set, including surface albedo and surface Lambertian body information; and the MODTRAN radiative transfer model is run to simulate the entrance pupil radiance of the ocean surface under clear sky as observed by the satellite at the zenith. .

[0014] Further, S2 includes:

[0015] S21: Based on the target sensor's spectral response function, including its band range, sampling interval, and full width at half maximum (FWHM), simulate the sensor's spectral response function using a Gaussian function. ;

[0016] S22: Radiance curve simulated by MODTRAN and sensor spectral response function Convolution simulates the spectral curve observed by the sensor, serving as a reference spectrum. ;

[0017] S23: Assuming that during the satellite sensor's on-orbit operation, only the center wavelength shift occurs, and the shift amount is consistent across all channels, a spectral drift model for the hyperspectral sensor is established to simulate the sensor radiance after spectral drift, which is then used as the measured spectrum. .

[0018] Further, S3 includes:

[0019] S31: Use spectral matching methods to compare the measured spectrum with the reference spectrum, and select the spectral angle as the evaluation standard to measure the similarity between the two spectra;

[0020] S32: Shift the center wavelength of the standard spectral response function by δ nm, and then adjust the radiance curve. Sampling was performed to obtain the observed spectrum after shifting the center wavelength of the target sensor by δ nm. ,calculate and The spectral angle between them;

[0021] S33: Constructing the cost loss function The simplex method is used to find the solution through iterative optimization. Extreme points; When taking the minimum value and The two spectral lines have the highest similarity at this point. Corresponding center wavelength offset This is the estimated center wavelength offset.

[0022] Further, S4 includes:

[0023] S41: Assuming that the center wavelength of the sensor drifts within the range of the left and right sampling interval width, different center wavelength offsets are given in step S2 to obtain multiple measured spectra. The center wavelength offsets are estimated separately in S3, and the residuals are calculated with the actual given center wavelength offsets to evaluate the uncertainty of the algorithm in S3.

[0024] S42: Assuming the sensor noise follows a Gaussian random distribution, give the simulated measured spectrum... Add random noise and evaluate the uncertainty caused by the sensor signal-to-noise ratio;

[0025] S43: Assuming that the inconsistency of the inter-spectral response follows a Gaussian random distribution and the relative magnitude relationship of the responses between each spectral band is fixed, add the inconsistency of the inter-spectral response to the simulated sensor measured gray value curve, and evaluate the algorithm uncertainty caused by the inconsistency of the inter-spectral response.

[0026] S44: Assuming there is a difference between the atmospheric CO2 concentration in the reference spectrum and the atmospheric CO2 concentration under actual observation conditions, assess the uncertainty caused by the difference in CO2 concentration;

[0027] S45: It is assumed that the estimation error of the algorithm in S3, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the influence of CO2 concentration on the algorithm are independent of each other, resulting in a combined uncertainty.

[0028] Further, step S43 includes:

[0029] Assuming that there is a linear relationship between the gray values ​​and radiance of each spectral band of the sensor, the radiance curve is converted into a gray value curve.

[0030] Further, S5 includes:

[0031] S51: Select the atmospheric infrared detector image, find the satellite cloud product MYD35_L2 corresponding to the time and location, and after image registration, filter out the clear sky ocean pixels;

[0032] S52: Resample the pixel radiance curve and assign a center wavelength offset as the measured spectrum in S23. ;

[0033] S53: Estimate the center wavelength offset of each clear sky ocean pixel through S2~S3, and take the average value as the final center wavelength offset;

[0034] S54: Calculate the residual between the estimated center wavelength offset in S53 and the actual center wavelength offset given in S52, and evaluate the accuracy.

[0035] The present invention also provides a very long-wavelength on-board spectral calibration and evaluation device based on a radiative transfer model, comprising the following modules:

[0036] The simulation module uses the MODTRAN radiative transfer model to simulate the radiance curves in the 12-16μm spectral range.

[0037] The spectrum acquisition module constructs a spectral drift model for a very long wavelength hyperspectral sensor. Given a center wavelength shift, it simulates the spectral response function of the very long wavelength hyperspectral sensor after spectral drift. The radiance curve is convolved with the spectral response functions before and after spectral drift to obtain the reference spectrum and the measured spectrum.

[0038] The evaluation module performs spectral matching between the acquired reference spectrum and the measured spectrum, estimates the center wavelength shift using an iterative optimization algorithm, and compares it with the center wavelength shift given in the spectrum acquisition module to evaluate the accuracy.

[0039] The analysis module performs uncertainty analysis based on the estimation error of the algorithm itself, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the impact of atmospheric CO2 concentration on the estimation results.

[0040] The re-evaluation module is applied to measured satellite data from atmospheric infrared detectors to further evaluate the accuracy of the method.

[0041] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method for very long-wavelength star spectral calibration and evaluation based on a radiative transfer model.

[0042] The present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for very long-wavelength on-board spectral calibration and evaluation based on a radiative transfer model.

[0043] The beneficial effects of this invention are as follows:

[0044] 1. This invention is a spectral calibration method for broadband very long wave infrared hyperspectral sensors. This invention studies the spectral calibration of broadband very long wave infrared hyperspectral sensors with a wavelength range covering 12-16 μm and a spectral resolution on the order of 48 nm, and achieves high accuracy.

[0045] 2. This invention evaluates the spectral calibration accuracy from multiple perspectives through simulation. Using the MODTRAN radiative transfer model as the primary simulation method, it employs spectral matching combined with an iterative optimization algorithm to solve for the center wavelength shift and performs uncertainty analysis. Finally, the method is applied to satellite data from an atmospheric infrared detector for accuracy evaluation. This invention provides crucial support for evaluating the on-orbit spectral calibration accuracy of very long-wavelength hyperspectral sensors. Attached Figure Description

[0046] Figure 1 This is a flowchart of a very long-wavelength on-board spectral calibration and evaluation method based on a radiative transfer model, according to the present invention.

[0047] Figure 2 This is a schematic diagram of the simulated hyperspectral sensor radiance spectrum according to an embodiment of the present invention;

[0048] Figure 3 This is a graph showing the variation of the spectral angle of the reference spectrum and the measured spectrum with the center wavelength offset in an embodiment of the present invention. Detailed Implementation

[0049] 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.

[0050] like Figure 1 As shown, the very long-wavelength on-board spectral calibration and evaluation method based on the radiative transfer model proposed in this invention includes the following steps:

[0051] S1: Simulation was performed using the MODTRAN radiative transfer model to simulate the radiance curve in the 12-16μm spectral range;

[0052] S2: Construct a spectral drift model for a very long wavelength hyperspectral sensor. Given a center wavelength shift, simulate the spectral response function of the very long wavelength hyperspectral sensor after spectral drift. Convolve the radiance curve obtained in S1 with the spectral response functions before and after spectral drift to obtain the reference spectrum and the measured spectrum.

[0053] S3: Perform spectral matching between the reference spectrum obtained in S2 and the measured spectrum, estimate the center wavelength offset using an iterative optimization algorithm, compare it with the center wavelength offset given in S2, and evaluate the accuracy.

[0054] S4: Conduct uncertainty analysis based on the estimation error of the algorithm itself, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the impact of atmospheric CO2 concentration on the estimation results;

[0055] S5: Applied to actual satellite data measured by atmospheric infrared detectors to further evaluate the accuracy of the method.

[0056] A further preferred embodiment of the present invention is that step S1 includes:

[0057] In the MODTRAN radiative transfer model, sensor parameters are set, including observation geometry, observation time, band range, and spectral resolution. Atmospheric parameters are also set, including aerosol optical thickness and water vapor content. Surface characteristic parameters are set, including surface albedo and surface Lambertian information. The MODTRAN radiative transfer model is then run to simulate the entrance pupil radiance of the ocean surface under clear sky conditions as observed by a satellite at the zenith. ,like Figure 2 As shown, Indicates wavelength.

[0058] A further preferred embodiment of the present invention is that step S2 includes the following steps:

[0059] S21: Based on the target sensor's spectral response function, including its band range, sampling interval, and full width at half maximum (FWHM), a Gaussian function is used to simulate the sensor's spectral response function. :

[0060] ;

[0061] in, , The center wavelength is denoted by FWHM, and the half-width at half-maximum (FWHM) is the target sensor's half-maximum width.

[0062] S22: Radiance curve simulated by MODTRAN and sensor spectral response function Convolution, simulating the spectral curve observed by the sensor. ;

[0063] ;

[0064] S23: Assuming that during the satellite sensor's on-orbit operation, only the center wavelength shift occurs, and the shift amount is consistent across all channels, a spectral drift model for the hyperspectral sensor is established to simulate the sensor radiance after spectral drift, which is then used as the measured spectrum. :

[0065] ;

[0066] in, This is the actual offset of the center wavelength.

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

[0068] S31: Use spectral matching to compare the measured spectrum with the reference spectrum. Select the spectral angle (SA) as the evaluation standard to measure the similarity between the two spectra. The smaller the spectral angle, the higher the similarity between the two spectra.

[0069] S32: Shift the center wavelength of the standard spectral response function by δ nm, and then adjust the radiance curve. Sampling was performed to obtain the observed spectrum after shifting the center wavelength of the target sensor by δ nm. :

[0070] ;

[0071] Where δ is the center wavelength shift.

[0072] calculate and Spectral angles between:

[0073]

[0074] in, Let i represent the number of bands to be calibrated, where i represents the i-th band. For the first The center wavelength of the band.

[0075] S33: Constructing the cost loss function The simplex method is used to find the solution through iterative optimization. Extreme point. The two spectral lines have the highest similarity when the minimum value is taken. This is the estimated center wavelength offset, such as Figure 3 As shown;

[0076] ;

[0077] A further preferred embodiment of the present invention is that step S4 specifically includes the following steps:

[0078] S41: Assuming the sensor center wavelength drifts within the range of the left and right sampling intervals, different center wavelength offsets are given in step S2. Multiple measured spectra were obtained, and the center wavelength shift was estimated using S3, comparing it with the actual given center wavelength shift. Calculate the residuals and evaluate the uncertainty of the algorithm itself in step S3. ;

[0079] S42: Assuming the sensor noise follows a Gaussian random distribution, give the simulated measured spectrum... Add random noise Evaluate the uncertainty caused by the sensor signal-to-noise ratio. ;

[0080] This represents a mean of 0 and a variance of 0. The Gaussian random distribution function, where , Representing the Each band, For the first The center wavelength of the band, Represents the number of sensor bands, and SNR is the signal-to-noise ratio. Take the value from the middle.

[0081] Measured spectrum of the sensor after adding noise for ;

[0082] Each experiment was repeated 100 times to avoid the influence of random noise.

[0083] S43: Assuming the non-uniformity of the inter-spectral responses conforms to a Gaussian random distribution, and the relative magnitudes of the responses between each spectral band are fixed, provide the simulated measured grayscale curves. Increasing inter-spectral response inconsistency and evaluating the algorithm uncertainty caused by inter-spectral response inconsistency. ;

[0084] Due to the inconsistency in spectral response, the actual gray value of the pixel for:

[0085] ;

[0086] in, Let be the response rate of each spectral band. Assuming that the inter-spectral responses are not uniformly distributed according to a Gaussian random distribution, and that the relative magnitudes of the responses between the spectral bands are fixed, then:

[0087] ;

[0088] in, The representative mean is variance is The Gaussian random distribution function, where, Then you can get The inter-spectral response has a pixel non-uniformity of less than 10%, and the relative magnitude relationship of the inter-spectral response remains unchanged.

[0089] S44: Assuming a difference between the atmospheric CO2 concentration in the reference spectrum and the actual observed atmospheric CO2 concentration, assess the uncertainty caused by the difference in CO2 concentration. .

[0090] ;

[0091] in, The atmospheric CO2 concentration under actual observation conditions is At that time, the entrance pupil radiance simulated by the MODTRAN radiative transfer model;

[0092] S45: It is assumed that the algorithm estimation error, sensor signal-to-noise ratio, inter-spectral response inconsistency, and the influence of CO2 concentration on the algorithm in S3 are independent of each other, resulting in a combined uncertainty. :

[0093] ;

[0094] in, These represent the algorithm estimation error, sensor signal-to-noise ratio, inter-spectral response inconsistency, and uncertainty caused by CO2 concentration in S3, respectively.

[0095] A further preferred embodiment of the present invention is that step S43 includes:

[0096] Assuming a linear relationship between grayscale values ​​and radiance in each spectral band of the sensor, the radiance curve is converted into a grayscale value curve. .

[0097] ;

[0098] in, It refers to the radiance at each wavelength. and These are the radiation calibration coefficients for each waveband. These are the grayscale values ​​at each band.

[0099] A further preferred embodiment of the present invention is that step S5 includes the following steps:

[0100] S51: Select the atmospheric infrared detector image, find the satellite cloud product MYD35_L2 corresponding to the time and location, and after image registration, filter out the clear sky ocean pixels.

[0101] S52: Pixel radiance curve of clear sky ocean Perform spectral resampling and give a center wavelength offset As the measured spectrum in step S23 ;

[0102] ;

[0103] S53: Estimate the center wavelength offset of each clear-sky ocean pixel and take the average value as the final center wavelength offset. ;

[0104] S54: Calculate the residual between the estimated center wavelength offset and the actual center wavelength offset to assess the accuracy.

[0105] residual= ;

[0106] Figure 2 This is a simulated hyperspectral sensor radiance spectrum according to an embodiment of the present invention. The horizontal axis represents wavelength in micrometers (μm), and the vertical axis represents radiance in watts (W). cm -2 sr -1 μm -1As the wavelength increases, the radiance decreases due to the strong absorption of CO2, with a small peak around 15 μm. Limited by the spectral resolution of the target sensor at 48 nm, the sensor cannot capture clear CO2 absorption characteristic lines, retaining only some peaks and troughs.

[0107] Figure 3 This describes the changes in the reference spectrum and the measured spectrum with respect to the center wavelength shift in this embodiment of the invention. The horizontal axis represents the assumed center wavelength shift of the reference spectrum, in nanometers (nm), and the vertical axis represents the spectral angle between the reference spectrum after the assumed center wavelength shift and the measured spectrum, in degrees (°). The actual center wavelength shift of the measured spectrum is 4 nm. When the assumed center wavelength shift of the reference spectrum is... When the wavelength is nm, the spectral angle between the two spectral lines is the smallest, and the center wavelength shift estimated by the method of this invention is: nm.

[0108] The present invention also provides a very long-wavelength on-board spectral calibration and evaluation device based on a radiative transfer model, comprising the following modules:

[0109] The simulation module uses the MODTRAN radiative transfer model to simulate the radiance curves in the 12-16μm spectral range.

[0110] The spectrum acquisition module constructs a spectral drift model for a very long wavelength hyperspectral sensor. Given a center wavelength shift, it simulates the spectral response function of the very long wavelength hyperspectral sensor after spectral drift. The radiance curve is convolved with the spectral response functions before and after spectral drift to obtain the reference spectrum and the measured spectrum.

[0111] The evaluation module performs spectral matching between the acquired reference spectrum and the measured spectrum, estimates the center wavelength shift using an iterative optimization algorithm, and compares it with the center wavelength shift given in the spectrum acquisition module to evaluate the accuracy.

[0112] The analysis module performs uncertainty analysis based on the estimation error of the algorithm itself, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the impact of atmospheric CO2 concentration on the estimation results.

[0113] The re-evaluation module is applied to measured satellite data from atmospheric infrared detectors to further evaluate the accuracy of the method.

[0114] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method for very long-wavelength star spectral calibration and evaluation based on a radiative transfer model.

[0115] The present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for very long-wavelength on-board spectral calibration and evaluation based on a radiative transfer model.

[0116] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0117] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0118] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0119] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0120] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0121] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model, characterized in that, Includes the following steps: S1: Simulation was performed using the MODTRAN radiative transfer model to simulate the radiance curve in the 12-16μm spectral range; S2: Construct a spectral drift model for a very long wavelength hyperspectral sensor. Given a center wavelength shift, simulate the spectral response function of the very long wavelength hyperspectral sensor after spectral drift. Convolve the radiance curve obtained in S1 with the spectral response functions before and after spectral drift to obtain the reference spectrum and the measured spectrum. S3: Perform spectral matching between the reference spectrum obtained in S2 and the measured spectrum, estimate the center wavelength offset using an iterative optimization algorithm, compare it with the center wavelength offset given in S2, and evaluate the accuracy. S4: Conduct uncertainty analysis based on the estimation error of the algorithm itself, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the impact of atmospheric CO2 concentration on the estimation results; S5: Applied to actual satellite data measured by atmospheric infrared detectors to further evaluate the accuracy of the method.

2. The method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model according to claim 1, characterized in that, S1 includes: In the MODTRAN radiative transfer model, sensor parameters are set, including observation geometry, observation time, band range, and spectral resolution; atmospheric parameters are set, including aerosol optical thickness and water vapor content; surface characteristic parameters are set, including surface albedo and surface Lambertian body information; and the MODTRAN radiative transfer model is run to simulate the entrance pupil radiance of the ocean surface under clear sky as observed by the satellite at the zenith. .

3. The method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model according to claim 1, characterized in that, S2 includes: S21: Based on the target sensor's spectral response function, including its band range, sampling interval, and full width at half maximum (FWHM), simulate the sensor's spectral response function using a Gaussian function. ; S22: Radiance curve simulated by MODTRAN and sensor spectral response function Convolution simulates the spectral curve observed by the sensor, serving as a reference spectrum. ; S23: Assuming that during the satellite sensor's on-orbit operation, only the center wavelength shift occurs, and the shift amount is consistent across all channels, a spectral drift model for the hyperspectral sensor is established to simulate the sensor radiance after spectral drift, which is then used as the measured spectrum. .

4. The method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model according to claim 1, characterized in that, S3 includes: S31: Use spectral matching methods to compare the measured spectrum with the reference spectrum, and select the spectral angle as the evaluation standard to measure the similarity between the two spectra; S32: Shift the center wavelength of the standard spectral response function by δ nm, and then adjust the radiance curve. Sampling was performed to obtain the observed spectrum after shifting the center wavelength of the target sensor by δ nm. ,calculate and The spectral angle between them; S33: Constructing the cost loss function The simplex method is used to find the solution through iterative optimization. Extreme points; When taking the minimum value and The two spectral lines have the highest similarity at this point. Corresponding center wavelength offset This is the estimated center wavelength offset.

5. The method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model according to claim 1, characterized in that, S4 includes: S41: Assuming that the center wavelength of the sensor drifts within the range of the left and right sampling interval width, different center wavelength offsets are given in step S2 to obtain multiple measured spectra. The center wavelength offsets are estimated separately in S3, and the residuals are calculated with the actual given center wavelength offsets to evaluate the uncertainty of the algorithm in S3. S42: Assuming the sensor noise follows a Gaussian random distribution, give the simulated measured spectrum... Add random noise and evaluate the uncertainty caused by the sensor signal-to-noise ratio; S43: Assuming that the inconsistency of the inter-spectral response follows a Gaussian random distribution and the relative magnitude relationship of the responses between each spectral band is fixed, add the inconsistency of the inter-spectral response to the simulated sensor measured gray value curve, and evaluate the algorithm uncertainty caused by the inconsistency of the inter-spectral response. S44: Assuming there is a difference between the atmospheric CO2 concentration in the reference spectrum and the atmospheric CO2 concentration under actual observation conditions, assess the uncertainty caused by the difference in CO2 concentration; S45: It is assumed that the estimation error of the algorithm in S3, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the influence of CO2 concentration on the algorithm are independent of each other, resulting in a combined uncertainty.

6. The method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model according to claim 5, characterized in that, Step S43 includes: Assuming that there is a linear relationship between the gray values ​​and radiance of each spectral band of the sensor, the radiance curve is converted into a gray value curve.

7. The method for on-board spectral calibration and evaluation of very long-wavelength satellites based on a radiative transfer model according to claim 1, characterized in that, S5 includes: S51: Select the atmospheric infrared detector image, find the satellite cloud product MYD35_L2 corresponding to the time and location, and after image registration, filter out the clear sky ocean pixels; S52: Resample the pixel radiance curve and assign a center wavelength offset as the measured spectrum in S23. ; S53: Estimate the center wavelength offset of each clear sky ocean pixel through S2~S3, and take the average value as the final center wavelength offset; S54: Calculate the residual between the estimated center wavelength offset in S53 and the actual center wavelength offset given in S52, and evaluate the accuracy.

8. A very long-wavelength on-board spectral calibration and evaluation device based on a radiative transfer model, characterized in that, Includes the following modules: The simulation module uses the MODTRAN radiative transfer model to simulate the radiance curves in the 12-16μm spectral range. The spectrum acquisition module constructs a spectral drift model for a very long wavelength hyperspectral sensor. Given a center wavelength shift, it simulates the spectral response function of the very long wavelength hyperspectral sensor after spectral drift. The radiance curve is convolved with the spectral response functions before and after spectral drift to obtain the reference spectrum and the measured spectrum. The evaluation module performs spectral matching between the acquired reference spectrum and the measured spectrum, estimates the center wavelength shift using an iterative optimization algorithm, and compares it with the center wavelength shift given in the spectrum acquisition module to evaluate the accuracy. The analysis module performs uncertainty analysis based on the estimation error of the algorithm itself, the sensor signal-to-noise ratio, the inconsistency of inter-spectral response, and the impact of atmospheric CO2 concentration on the estimation results. The re-evaluation module is applied to measured satellite data from atmospheric infrared detectors to further evaluate the accuracy of the method.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the very long-wavelength on-board spectral calibration and evaluation method based on the radiative transfer model as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the very long-wavelength on-board spectral calibration and evaluation method based on the radiative transfer model as described in any one of claims 1 to 7.

Images