Method and device for evaluating spectral relative calibration accuracy of spaceborne infrared hyperspectral detector

Abstract

The application provides a satellite-borne infrared hyperspectral detector spectral relative calibration precision evaluation method and device, and relates to the technical field of optical precision measurement. The method selects a spectral calibration precision highest detection element in a surface array detector as a reference detection element, and takes the observation spectrum of the reference detection element as a reference spectrum. The reference spectrum and the comparison spectrum are respectively converted into corresponding brightness temperature spectra through the inverse function of the Planck equation, and the spectral observation samples satisfying the brightness temperature difference threshold are screened. The spectral observation samples are converted into interferogram signals through inverse Fourier transform, and the interferogram is zero-filled and then subjected to Fourier transform to obtain a precise interpolation spectrum with fine spectral intervals. The spectral interval of a stable gas absorption band is selected as an evaluation window, the frequency offset of the comparison spectrum relative to the reference spectrum is calculated, and the spectral relative calibration deviation is obtained, which can significantly improve the efficiency and reliability of the spectral calibration precision evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of optical precision measurement, and particularly relates to a method and device for evaluating the relative calibration accuracy of a satellite-borne infrared hyperspectral detector. BACKGROUND

[0002] The satellite-borne infrared hyperspectral atmospheric detector is a precise quantitative application remote sensing instrument, and the spectral calibration accuracy of the observation data directly affects the spectral radiometric calibration accuracy, and also affects the accuracy of subsequent utilization of the observation spectrum to retrieve atmospheric temperature, relative humidity, trace gas concentration and other earth atmospheric physical parameters. At present, the evaluation of the spectral calibration accuracy of the infrared hyperspectral detector mainly adopts an absolute spectral calibration evaluation method, which needs to use an atmospheric radiation transfer forward model to simulate the spectral radiance of the top of the atmosphere in a clear sky scene. Then, the simulated spectrum is taken as the reference true value, and the spectral calibration accuracy of the observation spectrum is evaluated by statistically comparing the spectral line position deviation of the observation spectral sample relative to the simulated spectrum, which has the following defects:

[0003] (1) Time efficiency defect: The forward model used for absolute spectral calibration evaluation is generally developed by using a line-by-line integration method, such as the internationally recognized line-by-line radiative transfer model (LBLRTM). Since this model uses the spectral line parameters of the high-resolution atmospheric molecular absorption database HITRAN, the simulated atmospheric spectrum has accurate spectral line positions. However, although the line-by-line radiative transfer model has high calculation accuracy, the calculation speed is slow (for example, it takes about 10-15 s to calculate a spectrum with a spectral interval of 0.001 cm -1 -3000 cm -1 using LBLRTM), and the calculation of molecular absorption contribution for each channel and the calculation of high-precision transmittance for each layer of atmosphere also occupies a large amount of computer resources.

[0004] (2)Supporting data defects: When simulating the spectrum by using the atmospheric radiation transfer forward model, the atmospheric background field information such as atmospheric temperature / humidity, ground temperature / humidity, air pressure, trace gas content needs to be input as supporting data. The existing background field data mainly includes forecast data (such as EC, GFS, GRAPES data) or atmospheric state analysis data (such as ERA5, CRA data) released by the meteorological department. Among them, the accuracy of the forecast data decreases with the increase of the forecast time; the accuracy of the analysis data is high but has a certain time lag, and the time cannot be strictly synchronized with the satellite observation data. In addition, the spatial resolution of various background field data and the spatial resolution of real-time satellite observation data are not completely consistent. Therefore, the above shortcomings will lead to the fact that the simulated spectrum cannot be completely and stably matched with the observed spectrum, thereby affecting the evaluation of the spectral calibration accuracy. SUMMARY

[0005] Therefore, in view of the technical defects of low calculation timeliness, large calculation resource consumption and poor spatio-temporal matching of supporting data and observation data in the evaluation of the spectral calibration accuracy of the satellite-borne infrared hyperspectral atmospheric detector by using the absolute spectral calibration method, the purpose of the present application is to provide a spectral relative calibration accuracy evaluation method and device for a satellite-borne infrared hyperspectral detector, which is used for characterizing the spectral calibration accuracy and calibration stability of the instrument, and the spectral calibration accuracy of the instrument is evaluated by simply using the satellite instrument's own observation data and comparing the spectral positions with each other. Since the method of the present application does not rely on the reference spectrum simulated by the atmospheric radiation transfer model, it can overcome the defects of low calculation timeliness and poor spatio-temporal matching of observation data and simulated data existing in the use of the line-by-line integral radiation transfer model.

[0006] To achieve the above purpose, the present application provides the following technical solutions:

[0007] Based on the above purpose, in the first aspect, the present application provides a spectral relative calibration accuracy evaluation method for a satellite-borne infrared hyperspectral detector, comprising the following steps:

[0008] Selecting a detector element with the highest spectral calibration accuracy in a surface array detector as a reference detector element, and taking the observed spectrum of the reference detector element as a reference spectrum S ref ;

[0009] Converting the reference spectrum S ref and the comparison spectrum S ver into their respective corresponding brightness temperature spectra B ver through the inverse function of the Planck equation, and screening the spectral observation samples that meet the brightness temperature difference threshold;

[0010] Converting the spectral observation samples into an interferogram signal through inverse Fourier transform, performing Fourier forward transform on the zero-padded interferogram, and obtaining a fine interpolation spectrum with fine spectral intervals;

[0011] The spectral interval of the stable gas absorption band is selected as the evaluation window. The reference spectrum position of the precise interpolation is fixed within the selected evaluation window. The comparison spectrum is shifted relative to the reference spectrum, and then the cross-correlation coefficient and standard deviation of each step are calculated.

[0012] The mutual correlation coefficient and sample deviation standard deviation between the comparison spectrum and the reference spectrum at each step are iteratively calculated to determine the frequency shift step length corresponding to the maximum correlation and minimum deviation standard deviation. The frequency offset of the comparison spectrum relative to the reference spectrum is calculated to obtain the relative calibration deviation of the spectrum.

[0013] As a further solution of the present invention, the reference probe is preferably selected from the probe closest to the optical axis of the detector (such as the detector center probe). When selecting the reference probe, the probe with the highest spectral calibration accuracy can be selected based on the pre-launch instrument laboratory spectral calibration test or the post-launch absolute spectral calibration evaluation method, and the corresponding observed spectrum is used as the reference spectrum S for the relative spectral calibration accuracy evaluation. ref .

[0014] As a further solution of the present invention, the observed spectrum corresponding to the calibration accuracy detector to be evaluated other than the reference detector in the area array detector is the comparison spectrum S ver , by controlling the brightness temperature spectrum B ref and brightness temperature spectrum B ver The difference ΔB=|B ver -B ref The size of | determines the uniformity of the observed scene.

[0015] As a further solution of the present invention, when screening spectral observation samples that meet the brightness temperature difference threshold, screening satisfies ΔB=|B ver -B ref The observation samples with a depth of |<0.5K are selected, and scene data with cloud contamination or surface heterogeneity are excluded.

[0016] As a further solution of the present invention, when removing scene data with cloud contamination or surface heterogeneity, scene uniformity screening includes:

[0017] Observation scenes with cloud cover ≥5% or surface heterogeneity were excluded;

[0018] Only clear sky ocean area data are retained as valid samples.

[0019] As a further solution of the present invention, the reference spectrum S ref and comparison spectrum S ver The inverse function of Planck equation is converted into the corresponding brightness temperature spectrum B ver When , the brightness temperature spectrum expression is derived according to the inverse function of Planck equation, where the expression of the inverse function of Planck equation is:

[0020]

[0021] The deduced brightness temperature spectral expression is:

[0022]

[0023] In the formula, Planck constant is h = 6.626 x 10 -34 J·s, light speed is c = 2.998 x 10 -34 m / s, Boltzmann constant is k B = 1.38065 x 10 -23 J / K, variable σ is wave number cm -1 .

[0024] As a further scheme of the present application, when the interferogram is zero-padded, the length after zero-padding of the interferogram needs to reach 2 15 The above, σ fine = 0.001 cm -1 , then the relative frequency shift detection sensitivity reaches 1.25 ppm at 800 cm -1 .

[0025] As a further scheme of the present application, when the spectral interval of the stable gas absorption band is selected as the evaluation window, at least one spectral interval in the stable gas absorption band is selected as the evaluation window, including one of the following wave bands:

[0026] Long-wave CO2 absorption band: 680 ± 10 cm -1 , 710 ± 10 cm -1 , 755 ± 10 cm -1 ;

[0027] Medium-wave water vapor absorption band: 1480 ± 40 cm -1 , 1560 ± 40 cm -1 , 1640 ± 40 cm -1 ;

[0028] Medium-short-wave CO2 absorption band: 2250 ± 25 cm -1 , 2305 ± 25 cm -1 .

[0029] As a further scheme of the present application, the cross-correlation coefficient and the sample deviation standard deviation between the comparison spectrum and the reference spectrum are calculated at each step in the iteration, the corresponding frequency shift step when the correlation is maximum and the deviation standard deviation is minimum is determined, and when the relative frequency shift amount is calculated, the frequency shift amount Δσ of the comparison spectrum relative to the reference spectrum is calculated through the precise spectral interval σ fine and the moving step m, and the relative frequency shift amount is:

[0030]

[0031] wherein, The correlation coefficient calculation formula is:

[0032]

[0033] In the formula, i is the spectral channel number, n is the number of spectral channels, And D are the mean and standard deviation of the corresponding spectral sample respectively; the deviation standard deviation calculation formula defined based on the deviation of the reference spectrum and the comparison spectrum is:

[0034]

[0035] In a second aspect, the present application provides a device for evaluating the spectral relative calibration accuracy of a spaceborne infrared hyperspectral detector, comprising the following components:

[0036] The reference pixel selection module is used to select the pixel with the highest spectral calibration accuracy from the area array detector as the reference pixel, and obtain the observation spectrum thereof as the reference spectrum;

[0037] The comparison sample screening module converts the reference spectrum and the comparison spectrum into brightness temperature spectra, screens the uniform scene data through a brightness temperature difference threshold, and eliminates cloud pollution and surface heterogeneity samples;

[0038] The spectral precision interpolation module performs inverse Fourier transform on the screened reference spectrum and comparison spectrum to generate an interferogram signal, and after zero point filling, performs Fourier transform to generate a precision interpolated spectrum with fine spectral intervals;

[0039] The evaluation window selection module automatically matches the spectral interval of the stable gas absorption band as the evaluation window, fixes the position of the precision interpolated reference spectrum in the selected evaluation window, and then steps and translates the comparison spectrum relative to the reference spectrum, and calculates the correlation coefficient and the deviation standard deviation at each step;

[0040] The frequency shift calculation module determines the correlation maximum and the deviation standard deviation minimum corresponding to the frequency shift step length by iteratively calculating the correlation coefficient and the sample deviation standard deviation between the comparison spectrum and the reference spectrum at each step, calculates the frequency offset of the comparison spectrum relative to the reference spectrum, and obtains the spectral relative calibration deviation.

[0041] As a further scheme of the present application, the reference pixel selection module comprises:

[0042] The pre-launch calibration database stores the precision evaluation data set of the spectral calibration accuracy of each pixel of the infrared hyperspectral detector in the pre-launch laboratory environment;

[0043] The dynamic updating unit selects a spectral calibration element with the highest spectral calibration precision based on a pre-launch instrument laboratory spectral calibration test or a post-launch absolute spectral calibration evaluation method, and takes the corresponding observation spectrum as a reference spectrum for relative spectral calibration precision evaluation, and optimizes the selection priority of the reference spectral element.

[0044] Compared with the prior art, the spectral relative calibration precision evaluation method and device for the satellite-borne infrared hyperspectral detector proposed by the application significantly improves the efficiency and reliability of the spectral calibration precision evaluation through the optimization of the reference spectral element selection, scene screening, spectral interpolation and frequency shift detection, and has the following beneficial effects:

[0045] 1. The dependence on external simulation data is reduced, and the spectral calibration precision evaluation efficiency is improved.

[0046] The application takes the observation spectrum of the high-precision spectral element of the instrument itself as a reference, replaces the absolute spectral calibration precision evaluation carried out by the traditional LBLRTM simulation spectrum, does not need to simulate the atmospheric spectrum by the line-by-line radiative transfer model (LBLRTM), reduces the time consumption of single calibration precision evaluation, eliminates the uncertainty of external data, avoids the simulation deviation caused by the input error of the atmospheric background field, is especially suitable for the areas with low simulation spectral precision such as the polar region and complex terrain, realizes the rapid calibration evaluation through the automatic process, meets the timeliness requirement of meteorological disaster monitoring, and supports the real-time / near real-time processing of the on-orbit observation data.

[0047] 2. The scene screening and quality control optimization, the precise spectral interpolation and the high-sensitivity frequency shift detection are realized, and the 1ppm level precision detection is realized.

[0048] The application screens the uniform scene through the brightness temperature difference threshold, combines the short-wave infrared channel cloud detection and the geographic mask to remove the interference samples, removes the cloud pollution and the surface heterogeneity samples, ensures the stability of the spectral characteristics in the evaluation window, guarantees the significance and sensitivity of the spectral frequency shift detection through the brightness temperature spectrum comparison, generates the fine interpolation spectrum through the zero-filling of the interferogram, calculates the frequency shift deviation, captures the slight shift of the compared spectrum center wavelength, and adapts to different area array detector layouts.

[0049] These aspects or other aspects of the application will be more apparent in the following description of the embodiments. It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the application. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the following will briefly introduce the drawings needed to be used in the exemplary embodiments or the related art description. The drawings are used to provide further understanding of the present application, and constitute a part of the specification, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation to the present application. In the drawings:

[0051] Figure 1 A flow chart of a spectral relative calibration accuracy evaluation method of a spaceborne infrared hyperspectral detector according to an embodiment of the present application.

[0052] Figure 2 A schematic diagram of a 3x3 layout array detector used in a spectral relative calibration accuracy evaluation method of a spaceborne infrared hyperspectral detector according to an embodiment of the present application.

[0053] Figure 3 A screened comparison spectral sample S ver and a reference spectrum S ref distribution comparison schematic diagram.

[0054] Figure 4 A spectral graph after precise interpolation and evaluation window interception in a spectral relative calibration accuracy evaluation method of a spaceborne infrared hyperspectral detector according to an embodiment of the present application.

[0055] Figure 5 A curve graph of a correlation coefficient of a comparison spectrum and a reference spectrum with a moving step in a spectral relative calibration accuracy evaluation method of a spaceborne infrared hyperspectral detector according to an embodiment of the present application.

[0056] Figure 6 A curve graph of a standard deviation of a comparison spectrum and a reference spectrum with a moving step in a spectral relative calibration accuracy evaluation method of a spaceborne infrared hyperspectral detector according to an embodiment of the present application. DETAILED DESCRIPTION

[0057] The present application will be further described below in conjunction with the drawings and specific embodiments. It should be noted that the following described embodiments or technical features can be combined with each other to form new embodiments without conflict.

[0058] In order to make the objectives, technical solutions and advantages of the present application clearer, the following will further specifically explain the embodiments of the present application with reference to the drawings, and it should be understood that the specific embodiments described herein are only used to explain the present application, and do not limit the present application.

[0059] It should be noted that all the expressions of "first" and "second" in the embodiments of the present application are used to distinguish two non-identical entities or non-identical parameters with the same name, and "first" and "second" are only for the convenience of description, and should not be understood as a limitation on the embodiments of the present application. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, the process, method, system, product or equipment inherently includes other steps or units.

[0060] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0061] The flowchart shown in the drawings is only an example and does not necessarily include all the contents and operations / steps, and does not necessarily be executed in the order described. For example, some operations / steps can be decomposed, combined or partially merged, so the actual execution order can be changed according to the actual situation.

[0062] Some embodiments of the present application will be described in detail below in combination with the drawings. The following embodiments and features in the embodiments can be combined with each other without conflict.

[0063] In view of the technical defects of low calculation timeliness, large calculation resource consumption and poor spatio-temporal matching of support data and observation data existing in the absolute spectral calibration method for evaluating the spectral calibration accuracy of a spaceborne infrared hyperspectral atmospheric sounder, the present application provides a spaceborne infrared hyperspectral sounder spectral relative calibration accuracy evaluation method and device, which is used to characterize the spectral calibration accuracy and calibration stability of the instrument, and the spectral calibration accuracy of the instrument is evaluated by simply using satellite instrument observation data and comparing the spectral positions with each other. Since the method of the present application does not depend on the reference spectrum simulated by the atmospheric radiation transfer model, the defects of low calculation timeliness and poor spatio-temporal matching of observation data and simulated data existing in the line-by-line integral radiation transfer model can be overcome.

[0064] Referring to Figure 1 The embodiments of the present application provide a spaceborne infrared hyperspectral sounder spectral relative calibration accuracy evaluation method, which comprises the following steps:

[0065] Step S10, selecting a spectral calibration accuracy highest detection element in a surface array detector as a reference detection element, and taking the observation spectrum of the reference detection element as a reference spectrum S ref .

[0066] In this step, the reference detector is selected from the detectors closest to the optical axis of the detector. When selecting the reference detector, the detector with the highest spectral calibration accuracy is selected based on the pre-launch instrument laboratory spectral calibration test or post-launch absolute spectral calibration evaluation method, and the corresponding observation spectrum is taken as the reference spectrum S ref . Among them, according to the basic principle of spectral calibration of infrared hyperspectral detector, the detectors close to the optical axis of the instrument have relatively high calibration accuracy, and the detectors far from the optical axis have low calibration accuracy due to the large off-axis effect. Therefore, the reference spectrum S ref may select the spectrum of the central detector (such as the 5th detector of the 3x3 layout area detector used by JPSS / CrIS, FY-3 / HIRAS, etc., see Figure 2 ). Due to the differences in actual instruments and the complexity of spectral calibration, for safety, the spectrum of the detector with the highest absolute spectral calibration evaluation accuracy is taken as the reference spectrum S ref .

[0067] Step S20, converting the reference spectrum S ref and the comparison spectrum S ver into their respective brightness temperature spectra B ver through the inverse function of Planck's equation, and screening the spectral observation samples that meet the brightness temperature difference threshold.

[0068] In this step, the observation spectrum corresponding to the detector to be evaluated for calibration accuracy other than the reference detector in the area detector is the comparison spectrum S ver , and the uniformity of the observation scene is determined by controlling the difference ΔB = |B ref -B ver | between the brightness temperature spectrum B ver and the brightness temperature spectrum B ref .

[0069] Among them, when screening the spectral observation samples that meet the brightness temperature difference threshold, the observation samples that meet ΔB = |B ver -B ref | <0.5K are screened, and the cloud pollution or surface heterogeneity scene data is removed.

[0070] Among them, when removing cloud pollution or surface heterogeneity scene data, the scene uniformity screening includes:

[0071] Remove observation scenes with cloud coverage ≥5% or surface heterogeneity;

[0072] Only clear sky ocean area data is retained as valid samples.

[0073] As mentioned in step S10, in the 3x3 layout array detector, the radiation spectrum data of a large square area of 9 circular areas is measured at one time, also known as the measured spectrum of 9 detectors, and in the infrared hyperspectral atmospheric detection application, the spatial resolution requirement is not high, and it can be considered that the detection results of the 9 circular areas represent the results of the entire square area. As described above, if the LBLRTM software simulation data is used as the reference true value to detect the accuracy of the 9 detector measured spectrum, 9 times of LBLRTM simulation are required, which is very time-consuming in terms of computing resources and computing time; and the accuracy of the LBLRTM simulated spectrum also depends on the accuracy of the EC, ERA5 and other background field data input by the software. Therefore, there is great uncertainty in evaluating the accuracy of the instrument in the polar observation spectrum by using the LBLRTM simulated spectrum.

[0074] In this embodiment, instead of frequently evaluating the 9 detector measured spectrum by using the LBLRTM simulated spectrum, the mutual consistency or mutual accuracy, i.e., the relative calibration accuracy, between the 9 detector measured spectrums is evaluated, and then only one of the 9 detectors with the most accurate measured data and the most stable working performance is selected as the reference spectrum S ref As mentioned above, the 5th detector in the center of the detector is closest to the optical axis, and its spectral measurement accuracy is also the highest, so the 5th detector is selected as the reference spectrum S ref The spectrums of the remaining 8 detectors are compared with the 5th detector, and the spectrums of the remaining 8 detectors are the comparison spectrums S ver to be evaluated. It should be noted that the relative calibration accuracy evaluation scheme proposed in this application requires that the earth scene observed by the 9 detectors be relatively uniform, and the following uniformity judgment condition needs to be met, so that the data samples of the 9 detectors that can simultaneously observe the uniform earth scene are selected from many observation scenes.

[0075] Due to the complexity of the earth's atmospheric scene, the same resident observation of the observation area by the area array detector can form a large spectral difference between different detectors due to the complex observation scene in the field of view. For the infrared hyperspectral detector, the main reason is that the cloud blocks the observation of the atmospheric path of the detector, and therefore the reference spectrum and the comparison spectrum should be selected from the observation samples of the uniform scene such as clear sky and ocean. For simplicity, in the comparison spectrum sample screening, the reference spectrum S ref and the comparison spectrum S ver are converted into their respective brightness temperature spectrums B ver through the inverse function of the Planck equation, and the brightness temperature spectrum expression is derived according to the inverse function of the Planck equation, wherein the expression of the inverse function of the Planck equation is:

[0076]

[0077] The derived brightness temperature spectrum expression is:

[0078]

[0079] where h = 6.626 x 10 -34 J s is the Planck constant, c = 2.998 x 10 -34 m / s is the speed of light, and k B = 1.38065 x 10 -23 J / K is the Boltzmann constant, and σ is the wave number cm -1 .

[0080] Then the uniformity of the observed scene is determined by controlling the size of the difference ΔB = |B ver -B ref |, where |·| denotes taking the absolute value. If ΔB is greater than or equal to 0.5K, it means that B ver relatively B ref deviates greatly, since B ver and B ref correspond to the brightness temperature observation values of two adjacent regions, therefore if ΔB is greater than or equal to 0.5K, it is determined that the brightness temperature of the two regions is not uniform, and 0.5K is a threshold limit for judging the uniformity of adjacent regions. If ΔB is 0K, it means that the radiation of the adjacent regions is very consistent and very uniform, between 0 and 0.5K, it can be considered to be approximately uniform, and equal to or greater than 0.5K is considered to be unacceptable uniformity.

[0081] The present application takes the observation spectrum of the instrument itself with high-precision detection as a reference, replaces the absolute spectral calibration precision evaluation relying on the traditional LBLRTM simulated spectrum, does not need to simulate the atmospheric spectrum by line-by-line integral radiation transfer model (LBLRTM), reduces the time consumption of single calibration precision evaluation, eliminates the uncertainty of external data, avoids the simulation deviation caused by the input error of atmospheric background field, is especially suitable for the regions with low simulation spectrum precision such as the polar region and complex terrain, realizes rapid calibration evaluation through an automatic process, meets the timeliness demand of meteorological disaster monitoring, and supports real-time / near real-time processing of in-orbit observation data.

[0082] Step S30: converting the spectral observation sample into an interferogram signal through inverse Fourier transform, performing Fourier forward transform on the interferogram after zero padding, and obtaining a precise interpolation spectrum with fine spectral intervals.

[0083] In this step, when the interferogram is zero-padded, the length of the interferogram after zero padding needs to reach 2 15 The above, the spectral interval after interpolation reaches σ fine = 0.001 cm -1 , and the relative frequency shift detection sensitivity reaches 1.25ppm at 800cm -1 .

[0084] In this embodiment, in the spectrum precision interpolation, the spectrum calibration accuracy evaluation parameter generally adopts the spectrum position (wave number σ: cm -1 ), that is:

[0085]

[0086] Where Δσ represents the deviation of the spectral position, and Δσ is the spectrum to be compared S ver The frequency coordinate σ ver and the reference spectrum S as the true value ref The frequency coordinate σ ref The difference Δσ=σ ver -σ ref ; σ0 represents the reference spectrum frequency position. For a well-calibrated spectrum, the relative deviation is a very small value (about 10 -6 ~10 -5 magnitude), so multiply by a factor of 10 6 A more moderate value will be obtained, and the unit of ρ is ppm (1ppm=1×10 -6 ).

[0087] The sampling interval of the spectral calibration of the infrared hyperspectral atmospheric sounder is relatively wide. For example, the sampling interval of the spectral calibration of CrIS is 0.625cm. -1 , relative σ0=700cm -1 A position shift of one sampling interval will produce approximately 0.625 / 700×10 6 ≈893ppm relative deviation. In order to evaluate the spectral position offset of several ppm, it is necessary to perform precise interpolation on the calibration spectrum. Precision interpolation uses the zero-filling technique in Fourier transform spectroscopy. The specific steps include:

[0088] ①Convert the observed spectrum into an interferogram signal through inverse Fourier transform;

[0089] ②Add a large number of zero-value signals on both sides of the interference pattern;

[0090] ③ Perform Fourier transform on the zero-filled interference pattern to obtain a precise spectrum with fine spectral intervals.

[0091] Since the above interpolation method only changes the sampling interval of the original spectrum without changing the spectral resolution, it is called sinc interpolation. For example, in order to resolve a subtle spectral shift of more than 0.05ppm, the interferogram should be zero-filled to at least 2 15 Length, 700cm -1 The fine spectral spacing obtained at the position is about 1 / (0.625 -1 ×2 15) / 700≈0.027ppm of the relative frequency offset.

[0092] Step S40, select the spectral interval of stable gas absorption band as the evaluation window, fix the reference spectrum position of the precise interpolation in the selected evaluation window, perform position pull-off between the comparison spectrum and the reference spectrum, then step translation and calculate the cross-correlation coefficient and deviation standard deviation of each step; when the correlation coefficient reaches the maximum and the deviation standard deviation reaches the minimum,.

[0093] In this step, the spectral coverage of the infrared hyperspectral atmospheric detector is mainly distributed in the 600-3000 cm -1 (3.3-16.6 μm) infrared emission band, which contains the stable 15 μm carbon dioxide absorption band, the wide 6.25 μm water vapor absorption band, and the relatively transparent and non-absorbing 8-10 μm “infrared window region”. If the spectral calibration accuracy evaluation window band is selected in the “infrared window region”, the difference of the surface characteristics (such as the surface emissivity and the surface temperature) will affect the spectral shift estimation; if it is selected in the lower tropospheric water vapor band, the spatial and temporal distribution of water vapor will also affect the spectral calibration accuracy evaluation.

[0094] In this embodiment, when the spectral interval of the stable gas absorption band is selected as the evaluation window, at least one spectral interval in the stable gas absorption band is selected as the evaluation window, including one of the following bands:

[0095] Long-wave CO2 absorption band: 680±10 cm -1 , 710±10 cm -1 , 755±10 cm -1 ;

[0096] Medium-wave water vapor absorption band: 1480±40 cm -1 , 1560±40 cm -1 , 1640±40 cm -1 ;

[0097] Medium-short wave CO2 absorption band: 2250±25 cm -1 , 2305±25 cm -1 .

[0098] After the selected evaluation window, the precise interpolation spectrum can be intercepted according to the window.

[0099] Step S50, iteratively calculate the cross-correlation coefficient and sample deviation standard deviation between the comparison spectrum and the reference spectrum at each step, determine the frequency shift step corresponding to the maximum correlation and the minimum deviation standard deviation, calculate the frequency offset of the comparison spectrum relative to the reference spectrum, and obtain the spectral relative calibration deviation.

[0100] In the step, the cross-correlation coefficient and the sample deviation standard deviation between the comparison spectrum and the reference spectrum are iteratively calculated, the corresponding frequency shift step when the correlation is maximum and the deviation standard deviation is minimum is determined, the relative frequency shift amount of the comparison spectrum relative to the reference spectrum is estimated, and the correlation degree of the spectrum lines in the evaluation window region of the two is used, that is, the fixed reference spectrum position is used, then the position of the comparison spectrum is continuously shifted in the certain frequency shift interval (such as [-10ppm, +10ppm]) with the precise interpolation spectrum sampling interval as the unit step, and the cross-correlation coefficient and the sample deviation standard deviation of the two are iteratively calculated, when the cross-correlation coefficient reaches the maximum and the deviation standard deviation reaches the minimum, the frequency shift amount of the comparison spectrum relative to the reference spectrum is calculated through the precise spectrum interval fine and the moving step m, and the relative frequency shift amount is:

[0101]

[0102] wherein, is the average wave number of the evaluation window, and the cross-correlation coefficient calculation formula is:

[0103]

[0104] In the formula, i is the spectrum channel number, n is the spectrum channel number, and D are the average value and the standard deviation of the corresponding spectrum sample respectively; the deviation standard deviation calculation formula defined based on the sample deviation of the reference spectrum and the comparison spectrum is:

[0105]

[0106] The present application selects the uniform scene through the brightness temperature difference threshold value, combines the short-wave infrared channel cloud detection and the geographical mask to remove the interference samples, removes the cloud pollution and the ground heterogeneity samples, ensures the stability of the spectrum characteristics in the evaluation window, guarantees the significance and the sensitivity of the spectrum frequency shift detection through the brightness temperature spectrum comparison, fills zero to generate the fine interpolation spectrum for the interference diagram, calculates the frequency shift deviation, can capture the small shift of the center wavelength of the comparison spectrum, and is suitable for different area array detector layouts.

[0107] The present application realizes the evaluation of the relative calibration accuracy of the spectrum based on the observation spectrum data sample screening and quality control of the spaceborne infrared hyperspectral detector. Figure 3 The display is the example of the comparison spectrum and the reference spectrum after the brightness temperature conversion and the quality control, the whole spectrum is composed of the long-wave infrared (650-1095cm -1 ), the medium-wave infrared (1210-1750cm -1 ) and the short-wave infrared (2155-2550cm -1 ) separated three-section spectrum, and the spectrum sampling interval is 0.625cm -1After precise interpolation and window truncation of the three spectra, the fine spectra corresponding to the three windows are as follows: Figure 4 As shown, the spectral ranges of the three windows are 650~700cm -1 1260~1320cm -1 2280~2380cm -1 , the corresponding fine spectrum sampling interval is 0.0005cm -1 Fix the reference spectrum of the selected window, move the comparison spectrum left and right 20 steps, calculate the correlation coefficient and standard deviation of the comparison spectrum and the reference spectrum every time you move one step, and get the distribution of the two with the step length as shown below: Figure 5 and Figure 6 shown.

[0108] The relative spectral calibration accuracy evaluation method compares the spectra of two detector pairs observing a uniform scene, eliminating the need to simulate theoretical spectra using the line-by-line integrated radiative transfer model (LBLRTM), thereby greatly reducing the amount of data calculation. Figure 3 Select appropriate window areas in long wave, medium wave and short wave, and perform precise interpolation on the brightness temperature spectra of the two detectors, as shown in Figure 4 shown.

[0109] By using one of the brightness temperature spectra as the reference spectrum and the other as the comparison spectrum, and moving the reference spectrum, the maximum correlation coefficient method and the minimum deviation standard deviation method can be used to detect the relative spectral frequency shift. Then, the comparison spectrum and the reference spectrum are exchanged and the same operation is performed. The curve distribution of the deviation standard deviation and the correlation coefficient with the moving step size is as follows: Figure 5 and Figure 6 shown.

[0110] according to Figure 5 and Figure 6 According to the corresponding change relationship, find the step size of the comparison spectrum movement when the correlation coefficient is maximum and / or the standard deviation is minimum, and use equation (2) to calculate the frequency shift of the comparison spectrum relative to the reference spectrum, which is the relative spectral calibration accuracy of the two. Figure 5 and Figure 6 As shown, when the comparison spectrum moves two steps to the right relative to the reference spectrum, the correlation coefficient between the two reaches the maximum and the standard deviation reaches the minimum, so the relative spectral frequency shift between the two is calculated to be approximately ρ = 1.4706 ppm.

[0111] As a high-precision quantitative remote sensing instrument, the spaceborne infrared hyperspectral detector implements synchronous observation of regional scenes by using a planar detector composed of multiple detection elements. In quantitative remote sensing applications, the spectral calibration accuracy of each detection element needs to reach <5ppm, and the spectral calibration of the detection elements also needs to be highly consistent (relative deviation not more than 2-3ppm). The current spectral calibration accuracy evaluation technology mainly uses the atmospheric spectrum simulated by a line-by-line integral radiation transfer model as the reference true value, and then the spectral line position deviation of the observed spectrum relative to the simulated spectrum is compared to achieve the evaluation. However, due to the slow calculation speed, large resource consumption, and poor temporal and spatial matching of atmospheric background support data of the line-by-line integral model, the above absolute spectral calibration evaluation technology has the disadvantages of complex process, complex calculation, and many implementation constraints.

[0112] The present application no longer uses simulated spectrum as reference true value, but uses the detection element with the highest spectral calibration accuracy of the instrument and its corresponding observed spectrum as reference detection element and reference spectrum. The relative calibration accuracy between the detection elements is evaluated by comparing the relative frequency shift of the observed spectrum of other detection elements relative to the reference spectrum. Therefore, once the absolute spectral calibration accuracy of the reference detection element is obtained through pre-launch spectral calibration test or post-launch absolute spectral calibration accuracy evaluation, the absolute calibration accuracy of the compared detection element and the consistency of the spectral calibration accuracy between the detection elements can be inferred based on the relative calibration accuracy between the detection elements. The main technical progress of the present application is to break away from the excessive dependence on the simulated spectrum of the line-by-line integral radiation transfer model, simplify the calculation process of spectral calibration accuracy evaluation, and improve the calculation efficiency.

[0113] It should be noted that the key technical points of the present application are a complete set of implementation schemes for spectral calibration accuracy evaluation by spectral relative frequency shift detection using the observed data of the spaceborne infrared hyperspectral atmospheric detector, including the observation data screening method and the evaluation window selection method.

[0114] The existing technology mainly calculates the simulated spectrum by using the atmospheric radiation transfer model, and then detects the correlation frequency shift between the observed spectrum and the simulated spectrum. Since spectral simulation requires a large amount of calculation and depends on the accuracy of external input atmospheric background field data, it consumes a large amount of computing resources and reduces the timeliness. The present application is based on the observed data of the spaceborne instrument, and does not need to use the atmospheric radiation transfer model, so it has good timeliness, small calculation amount, and can quickly verify the spectral relative calibration accuracy. Further, the stability of the calibration can be evaluated by constructing the time series of the spectral relative frequency shift.

[0115] In some embodiments, the present application needs to convert the radiance spectrum into a brightness temperature spectrum by the inverse Planck function before screening the comparison spectrum sample part, and then set the brightness temperature deviation threshold control. Alternatively, the above-mentioned brightness temperature conversion can not be performed, and the radiance deviation threshold control can be set.

[0116] In the spectral precision interpolation part of the present application, the spectrum is first inversely Fourier transformed, then the interference graph is zero-filled, and then the zero-filled interference graph is Fourier transformed to obtain the interpolation spectrum. Alternatively, the program that can be used is to first Fourier transform the spectrum, then zero-fill the interference graph, and then inversely Fourier transform the zero-filled interference graph to obtain the interpolation spectrum. The two programs are essentially the same.

[0117] Other alternative solutions that appear to meet at least one of the above conditions are considered to conflict with the present application and should not be protected.

[0118] It should be noted that the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present application, and are not for limiting purposes. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be executed synchronously or asynchronously, for example, in multiple modules.

[0119] It should be understood that although the above steps are described in a certain order, these steps are not necessarily executed in the above-mentioned order. Unless explicitly stated herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, part of the steps of the present embodiment can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0120] In a second aspect of the embodiments of the present application, the present application also provides a device for evaluating the relative spectral calibration accuracy of a spaceborne infrared hyperspectral detector, comprising:

[0121] The reference pixel selection module is used to select the pixel with the highest spectral calibration accuracy from the area array detector as the reference pixel, and obtain the observation spectrum as the reference spectrum;

[0122] The comparison sample screening module converts the reference spectrum and the comparison spectrum into a brightness temperature spectrum, screens the uniform scene data through a brightness temperature difference threshold, and eliminates cloud pollution and ground heterogeneity samples;

[0123] Spectrum precision interpolation module: inverse Fourier transform is performed on the screened reference spectrum and the comparison spectrum to generate an interferogram signal, and after zero point filling, a Fourier transform is performed to generate a precision interpolation spectrum with fine spectral intervals;

[0124] Evaluation window selection module: a stable gas absorption band database is built in, the spectral range of the stable gas absorption band is automatically matched as the evaluation window, the position of the precision interpolated reference spectrum is fixed in the selected evaluation window, the position of the comparison spectrum relative to the reference spectrum is pulled off, then the comparison spectrum is stepped and translated, and the cross-correlation coefficient and the sample deviation standard deviation of each step are calculated;

[0125] Frequency shift calculation module: the cross-correlation coefficient and the sample deviation standard deviation between the comparison spectrum and the reference spectrum at each step are calculated by iteration, the frequency shift step corresponding to the maximum correlation and the minimum deviation standard deviation is determined, the frequency offset of the comparison spectrum relative to the reference spectrum is calculated, and the spectral relative calibration deviation is obtained.

[0126] In the embodiment, the reference element selection module comprises:

[0127] Pre-launch calibration database: store the precision evaluation data set of the spectral calibration accuracy of each element of the infrared hyperspectral detector in the pre-launch laboratory environment;

[0128] Dynamic updating unit: select the element with the highest spectral calibration accuracy based on the pre-launch instrument laboratory spectral calibration test or post-launch absolute spectral calibration evaluation method, and use the corresponding observation spectrum as the reference spectrum for relative spectral calibration accuracy evaluation, and optimize the reference element selection priority.

[0129] Through the above detailed steps, the satellite-borne infrared hyperspectral detector spectral relative calibration accuracy evaluation device of the present application is used to execute the steps of the satellite-borne infrared hyperspectral detector spectral relative calibration accuracy evaluation method in the above embodiment, which will not be described here.

[0130] The above is an exemplary embodiment disclosed by the present application, but it should be noted that various changes and modifications can be made without departing from the scope of the embodiments disclosed by the present application as defined in the claims. The functions, steps and / or actions of the method claims described herein do not need to be performed in any particular order. In addition, although the elements of the embodiments disclosed by the present application can be described or claimed in singular form, they can also be understood as plural unless explicitly limited to singular.

[0131] It should be understood that, as used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising", or "includes" and / or "including" when used herein, specify the presence of stated features, integers, steps, operations, elements, and / or components but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. The above-mentioned embodiments of the present application are only illustrative and not intended to limit the scope of the present application (including the claims). The technical features of the above-mentioned embodiments or different embodiments can be combined, and there are many other variations of different aspects of the present application as described above. In order to be brief, they are not provided in detail. Therefore, any omission, modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

[0132] It should be understood that, as used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising", or "includes" and / or "including" when used herein, specify the presence of stated features, integers, steps, operations, elements, and / or components but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. The above-mentioned embodiments of the present application are only illustrative and not intended to limit the scope of the present application (including the claims). The technical features of the above-mentioned embodiments or different embodiments can be combined, and there are many other variations of different aspects of the present application as described above. In order to be brief, they are not provided in detail. Therefore, any omission, modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for evaluating the relative calibration accuracy of a spaceborne infrared hyperspectral detector, characterized in that: The method comprises the following steps: Selecting a detector with the highest spectral calibration accuracy in the area array detector as a reference detector, and using the observed spectrum of the reference detector as the reference spectrum; The reference spectrum and the comparison spectrum are converted into their corresponding brightness temperature spectra through the inverse function of the Planck equation, and the spectral observation samples that meet the brightness temperature difference threshold are selected to participate in the spectral relative calibration accuracy evaluation; The reference spectrum and comparison spectrum observation samples are converted into interferogram signals through inverse Fourier transform, and the interferogram is zero-filled and then forward Fourier transform is performed to obtain precise interpolation spectra with fine spectral intervals. The spectral interval of the stable gas absorption band is selected as the evaluation window. The reference spectrum position of the precise interpolation is fixed within the selected evaluation window. The comparison spectrum is shifted relative to the reference spectrum, and then the cross-correlation coefficient and standard deviation of each step are calculated. The mutual correlation coefficient and sample deviation standard deviation between the comparison spectrum and the reference spectrum at each step are iteratively calculated to determine the frequency shift step length corresponding to the maximum correlation and minimum deviation standard deviation. The frequency offset of the comparison spectrum relative to the reference spectrum is calculated to obtain the relative calibration deviation of the spectrum.

2. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 1, wherein: The reference probe is preferably selected from the probe closest to the optical axis of the detector. When selecting the reference probe, the probe with the highest spectral calibration accuracy is selected based on the pre-launch instrument laboratory spectral calibration test or the post-launch absolute spectral calibration evaluation method, and the corresponding observed spectrum is used as the reference spectrum S for the relative spectral calibration accuracy evaluation. ref .

3. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 2, wherein: The observed spectrum corresponding to the calibration accuracy detector to be evaluated other than the reference detector in the array detector is the comparison spectrum S ver , by controlling the brightness temperature spectrum B ref and brightness temperature spectrum B ver The difference ΔB=|B ver -B ref The size of | determines the uniformity of the observed scene.

4. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 3, wherein: When screening spectral observation samples that meet the brightness temperature difference threshold, the samples that meet ΔB=|B ver -B ref The observation samples with a depth of |<0.5K are selected, and scene data with cloud contamination or surface heterogeneity are excluded.

5. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 4, wherein: When removing scene data with cloud contamination or surface heterogeneity, scene uniformity screening includes: Observation scenes with cloud cover ≥5% or surface heterogeneity were excluded; Only clear sky ocean area data are retained as valid samples.

6. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 4, wherein: The reference spectrum S ref and comparison spectrum S ver The inverse function of Planck equation is converted into the corresponding brightness temperature spectrum B ver When , the brightness temperature spectrum expression is derived according to the inverse function of Planck equation, where the expression of the inverse function of Planck equation is: The derived brightness temperature spectrum expression is: Where, Planck constant is h = 6.626 × 10 -34 J·s, the speed of light is c=2.998×10 -34 m / s, and the Boltzmann constant is k B =1.38065×10 -23 J / K, variable σ is the wave number cm -1 .

7. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 1, wherein: When zero-filling the interference pattern, the length of the interference pattern after zero-filling must reach 2 15 Above, the spectral interval after interpolation reaches σ fine =0.001cm -1 , then the relative frequency shift detection sensitivity is 800cm -1 It reaches 1.25ppm.

8. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 1, wherein: When selecting a spectral interval of a stable gas absorption band as an evaluation window, select at least one spectral interval in the stable gas absorption band as the evaluation window, including one of the following bands: Long-wave CO2 absorption band: 680±10cm -1 、710±10cm -1 、755±10cm -1 ; Medium-wave water vapor absorption band: 1480±40cm -1 、1560±40cm -1 、1640±40cm -1 ; Medium and short wave CO2 absorption band: 2250±25cm -1 、2305±25cm -1 .

9. The method for evaluating relative calibration accuracy of a spaceborne infrared hyperspectral detector according to claim 6, wherein: Iteratively calculate the correlation coefficient and sample deviation standard deviation between each step comparison spectrum and the reference spectrum, determine the frequency shift step corresponding to the maximum correlation and minimum deviation standard deviation, and calculate the relative frequency shift. The frequency offset Δσ of the comparison spectrum relative to the reference spectrum is calculated by the precise spectrum interval σ fine Calculated by the moving step length m, the relative frequency shift is: in, To evaluate the window wave number mean, the cross-correlation coefficient is calculated as follows: Where i is the spectral channel number, n is the number of spectral channels, and D are the mean and standard deviation of the corresponding spectral samples respectively; the calculation formula for the deviation standard deviation defined based on the deviation of the two spectral samples of the reference spectrum and the comparison spectrum is:

10. A device for evaluating relative calibration accuracy of a satellite-borne infrared hyperspectral detector, characterized in that: The device is used to perform the method for evaluating the relative calibration accuracy of a spaceborne infrared hyperspectral detector according to any one of claims 1 to 9, comprising: Reference detector selection module: used to select the detector with the highest spectral calibration accuracy from the array detector as the reference detector, and obtain its observed spectrum as the reference spectrum; Comparison sample screening module: converts reference spectra and comparison spectra into brightness temperature spectra, filters uniform scene data through brightness temperature difference threshold, and removes cloud pollution and surface heterogeneity samples; Spectral precision interpolation module: Perform inverse Fourier transform on the screened reference spectrum and comparison to generate interferogram signals, and then generate a precise interpolation spectrum with fine spectral intervals through forward Fourier transform after zero filling; Evaluation window selection module: Built-in stable gas absorption band database, automatically matches the spectral interval of stable gas absorption bands as the evaluation window, fixes the reference spectrum position of precise interpolation within the selected evaluation window, deviates the position of the comparison spectrum relative to the reference spectrum, then steps and translates it, and calculates the cross-correlation coefficient and standard deviation of each step; Frequency shift calculation module: It iteratively calculates the correlation coefficient and sample deviation standard deviation between the comparison spectrum and the reference spectrum at each step, determines the frequency shift step corresponding to the maximum correlation and minimum deviation standard deviation, calculates the frequency offset of the comparison spectrum relative to the reference spectrum, and obtains the relative calibration deviation of the spectrum.

Images