ROLO model improvement method combining foundation and meteorological satellite moon observation data

Abstract

The invention belongs to the field of remote sensor calibration, and provides an ROLO model improvement method combining foundation and meteorological satellite lunar observation data, which comprises the following steps of: acquiring a meteorological satellite lunar observation image data set and a foundation lunar observation image data set, and respectively correcting the meteorological satellite lunar observation image data set and the foundation lunar observation image data set; obtaining standard meteorological satellite observation moon radiance and foundation observation moon radiance, and calculating observation moon irradiance; calculating geometric parameters of observation points, the earth, the moon and the sun at the observation moments of the foundation and the meteorological satellite, and calculating the moon irradiance of the ROLO model; polynomial fitting is adopted for the relative deviation between the observed moon irradiance of different wave bands and different moon phase angles and the moon irradiance of the ROLO model, and the ROLO model is improved. The method is not affected by the atmosphere, and errors caused by atmospheric correction of ROLO model foundation observation data are directly avoided; and the phase angle dependence of the ROLO model can be corrected, and the prediction precision of the ROLO model can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensor calibration, and in particular to a ROLO model improvement method combining ground-based and meteorological satellite lunar observation data. Background Art

[0002] Remote sensor calibration is the premise of quantification of remote sensing products. The accuracy of remote sensor calibration largely determines the reliability of remote sensing data and the depth and breadth of application. In the current on-orbit calibration method of remote sensors, for satellites equipped with an onboard calibrator, calibration is performed using the onboard calibrator. However, the degradation problem of the onboard calibrator itself may cause the calibration results to be difficult to accurately reflect the actual on-orbit status of the satellite. For satellites not equipped with an onboard calibrator, calibration is usually performed using methods such as cross calibration, pseudo-invariant calibration point calibration, and on-site calibration. However, these calibration methods rely on the consistency of reference targets with known reflectivity; in addition, these calibration methods are also plagued by severe atmospheric attenuation. Affected by the earth's atmosphere, many imager channels are difficult or even impossible to calibrate.

[0003] Compared with the above calibration methods, the calibration method using the moon as a reference target has attracted increasing attention due to its significant advantages. The moon's surface has no atmosphere and is not affected by the atmosphere; the radiation characteristics of the moon are very stable, and the irradiance changes by no more than 10% per year. -8 The moon provides a calibration reference for most Earth-orbiting satellites every month, and the moon has suitable response characteristics for any Earth-orbiting satellite, which allows lunar images to collect sufficient signal-to-noise ratio without overexposure. Studies have shown that the calibration method based on the moon is very effective and has important practical application value.

[0004] The accuracy of the lunar-based calibration method depends on the accuracy of the lunar radiation model. In order to establish a common calibration reference for spacecraft, the United States Geological Survey implemented a lunar ground-based observation program called the "RObotic Lunar Observatory" program. Through 8 years of observations, the ROLO lunar irradiance model (abbreviated as ROLO model) was developed. The ROLO model predicts the equivalent reflectivity of the lunar disk in 32 wavelength bands (from 350nm to 2500nm) based on various observation geometric conditions. To date, the ROLO model is the most widely used model for predicting lunar reflectivity. However, the absolute accuracy of the ROLO model is only 5-10%, which is mainly caused by the following factors: (1) The ROLO model uses ground-based observation data and cannot completely eliminate the atmospheric effect; (2) The error caused by the introduction of the Apollo spectrum and the interpolation of the original 32 bands into the hyperspectral spectrum; (3) The error caused by the lunar phase angle dependence of the ROLO model. Summary of the invention

[0005] In view of this, the present invention aims to provide a ROLO model improvement method combining ground-based and meteorological satellite lunar observation data to solve the problem of insufficient absolute accuracy of the ROLO model.

[0006] To achieve the above object, the technical solution created by the present invention is implemented as follows: A method for improving the ROLO model by combining ground-based and meteorological satellite lunar observation data comprises the following steps: S1: Acquire meteorological satellite lunar observation image datasets and ground-based lunar observation image datasets; S2: Correct the meteorological satellite lunar observation image dataset and the ground-based lunar observation image dataset respectively to obtain the standard meteorological satellite observation lunar radiance and ground-based observation lunar radiance; S3: integrating and summing the ground-based observed lunar radiance of all pixels within the lunar boundary in the ground-based lunar observation image to obtain the ground-based observed irradiance, and integrating and summing the meteorological satellite observed lunar radiance of all pixels within the lunar boundary in the meteorological satellite lunar observation image to obtain the meteorological satellite observed irradiance. The ground-based observed irradiance and the meteorological satellite observed irradiance are collectively referred to as the observed lunar irradiance. S4: Solve the geometric parameters of the observation point, the Earth, the Moon, and the Sun at the time of ground-based and meteorological satellite observations, and calculate the lunar irradiance of the ROLO model; S5: The ROLO model is improved by using polynomial fitting to find the relative deviation between the observed lunar irradiance in different bands and different lunar phases and the ROLO model lunar irradiance.

[0007] Furthermore, in step S1, the process of obtaining the meteorological satellite lunar observation image dataset is as follows: Use STK software to simulate the orbit of meteorological satellites, calculate the time when the moon appears in the satellite's cold sky field of view based on the geometric relationship between the satellite, the earth, the sun, and the moon, and collect meteorological satellite lunar observation images.

[0008] Furthermore, in step S1, the process of obtaining the ground-based lunar observation image dataset is as follows: A hyperspectral imager is used on the ground to observe the moon and obtain ground-based lunar observation images; the observation range of the hyperspectral imager is 380-1000nm, and the spectral resolution is ≥1nm.

[0009] Furthermore, images with a lunar phase angle in the range of [-92°, 92°] and a complete lunar region are screened from the collected meteorological satellite lunar observation data to form a meteorological satellite lunar observation image dataset.

[0010] Furthermore, in step S2, the meteorological satellite lunar observation image data set is corrected to obtain the standard meteorological satellite lunar radiation brightness as follows: S21a: geometric correction of the lunar radiance observed by meteorological satellites, correcting the lunar radiance observed by meteorological satellites to the standard distance;

[0011] In the formula, represents the lunar radiance observed by meteorological satellites after geometric correction, It represents the lunar radiance observed by meteorological satellites before geometric correction. represents the distance between the sun and the moon, represents the distance between the meteorological satellite and the moon, represents the average distance between the Earth and the Moon, and 1 AU represents the average distance between the Earth and the Sun; S22a: Use effective wavelengths instead of spectral channels;

[0012] In the formula, represents the effective wavelength, represents the starting wavelength of the spectral channel, represents the end wavelength of the spectral channel, represents the spectral response function, represents the lunar soil reflectivity reference curve, represents the reference solar spectrum irradiance curve, represents the lunar solid angle; S23a: First, a threshold algorithm is used to extract a rectangular area containing the moon from the meteorological satellite lunar observation image, and then a Sobel edge detection algorithm is used to detect the lunar boundary in the extracted rectangular area; S24a: Select the average radiance value of the area from the 15th pixel to the 30th pixel width outside the edge of the moon as the night sky noise value, and subtract the night sky noise value from the meteorological satellite lunar observation image;

[0013] In the formula, represents the lunar radiance observed by meteorological satellites after correction for night sky noise, Represents the night sky noise value.

[0014] Furthermore, in step S2, the ground-based lunar observation image dataset is corrected to obtain the standard ground-based lunar radiance as follows: S21b: geometric correction of the lunar radiance observed by ground-based observations, and correction of the lunar radiance observed by ground-based observations to the standard distance;

[0015] In the formula, represents the ground-based lunar radiance after geometric correction, represents the ground-based lunar radiance before geometric correction, represents the distance between the sun and the moon, represents the distance between the ground position and the moon, represents the average distance between the Earth and the Moon, and 1 AU represents the average distance between the Earth and the Sun; S22b: Use the LBLRTM model to simulate atmospheric transmittance and perform atmospheric correction on the ground-based lunar radiance observations based on the atmospheric transmittance. ; In the formula, represents the ground-based observed lunar radiance after atmospheric correction, represents the atmospheric transmittance; S23b: firstly, a threshold algorithm is used to extract a rectangular region containing the moon from the ground-based lunar observation image, and then a Sobel edge detection algorithm is used to detect the lunar boundary in the extracted rectangular region; S24b: selecting a pure night sky area of ​​the ground-based lunar observation image that does not include the moon as a night sky noise value, and subtracting the night sky noise value from the ground-based lunar observation image;

[0016] In the formula, represents the ground-based observation of the moon’s radiance after correction for night sky noise, Represents the night sky noise value.

[0017] Furthermore, in step S3, the ground-based lunar observation radiance of all pixels within the lunar boundary in the ground-based lunar observation image is integrated and summed to obtain the ground-based observation irradiance, and the calculation formula is as follows: ; In the formula, represents the ground-based observation irradiance, represents the lunar radiance observed by meteorological satellites after correction for night sky noise, represents the solid angle of the pixel, and Respectively represent the number of rows and columns of meteorological satellite lunar observation images and ground-based lunar observation images; The meteorological satellite observed lunar radiation brightness of all pixels within the lunar boundary in the meteorological satellite lunar observation image is integrated and summed to obtain the meteorological satellite observed irradiance. The calculation formula is as follows: ; In the formula, represents the irradiance observed by meteorological satellite, Represents the ground-based lunar radiance after correction for night sky noise.

[0018] Furthermore, step S4 specifically includes the following steps: S41: solving geometric parameters of the observation point, the earth, the moon, and the sun at the ground-based and meteorological satellite observation times, inputting the geometric parameters into the ROLO model, and calculating the lunar albedo of the ROLO model corresponding to the ground-based and meteorological satellite observation times; wherein the geometric parameters include the moon phase angle, the longitude of the sun in the lunar coordinate system, the longitude of the observation point in the lunar coordinate system, and the latitude of the observation point in the lunar coordinate system;

[0019]

[0020] In the formula, represents the lunar albedo of the ROLO model, Indicates the band, Indicates the moon phase angle, represents the longitude of the sun in the lunar coordinate system, represents the latitude of the observation point in the lunar coordinate system, represents the longitude of the observation point in the lunar coordinate system, , , , , , , , , , , , , Respectively represent coefficient terms; observation points are ground locations or meteorological satellites; S42: Calculate the lunar irradiance of the ROLO model according to the lunar albedo of the ROLO model;

[0021] In the formula, represents the lunar irradiance of the ROLO model, represents the distance correction factor, It represents the solid angle of the moon relative to the earth at the standard distance. Represents solar radiation intensity.

[0022] Furthermore, in step S5, the calculation formula for the relative deviation between the observed lunar irradiance and the ROLO model lunar irradiance is:

[0023] In the formula, It represents the relative deviation between the observed lunar irradiance at the k-band and the lunar irradiance of the ROLO model. The k-band selects the common band of the meteorological satellite lunar observation image, the ground-based lunar observation image and the ROLO model, specifically any one of the bands of 470nm, 510nm, 640nm, 855nm, 1610nm and 2255nm. represents the observed lunar irradiance at the k-band, represents the ROLO model lunar irradiance at the k-band.

[0024] Furthermore, in step S5, a polynomial is used to fit the relative deviation of each k-band at different moon phase angles, so as to correct the phase angle dependence of the ROLO model at different k-bands and different moon phase angles, thereby improving the ROLO model. , ; ,

[0025]

[0026] ,

[0027] In the formula, and They represent the lunar irradiance of the improved ROLO model under negative and positive lunar phase angles, respectively; and They represent the improvement coefficients of the ROLO model under negative and positive moon phase angles, respectively; It represents the solid angle of the moon relative to the earth at standard distance; represents the solar radiation intensity; represents the lunar albedo of the ROLO model, represents the distance correction factor; and Both represent coefficients and are fixed values.

[0028] Compared with the prior art, the present invention can achieve the following technical effects: (1) The lunar data observed by meteorological satellites are not affected by the atmosphere, which directly avoids the errors caused by the atmospheric correction of the ground-based observation data of the ROLO model; the hyperspectral data observed by ground-based observations have better spectral resolution than ROLO data and can better reflect spectral information. Combining the lunar irradiance observed by meteorological satellites and ground-based observations can break through the limitations of ROLO data and integrate the advantages of data from two different platforms.

[0029] (2) The combination of lunar irradiance from meteorological satellites and ground-based observations can also correct the phase angle dependence of the ROLO model. Model correction coefficients are assigned to different bands and different lunar phases of the ROLO model to improve the ROLO model and enhance its prediction accuracy. The average relative deviation between the improved ROLO model and the observed data is reduced to 0.27%-0.39%. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A flow chart of a method for improving a ROLO model combining ground-based and meteorological satellite lunar observation data according to an embodiment of the present invention; Figure 2 A schematic diagram of the relative deviation between the observed lunar irradiance and the ROLO model lunar irradiance according to an embodiment of the present invention; Figure 3 A schematic diagram of the relative deviation between the observed lunar irradiance and the improved ROLO model lunar irradiance described in the embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and do not constitute a limitation of the invention.

[0032] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0033] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0034] like Figure 1 As shown, the ROLO model improvement method provided by the embodiment of the present invention that combines ground-based and meteorological satellite lunar observation data includes the following steps: S1: Acquire meteorological satellite lunar observation image datasets and ground-based lunar observation image datasets.

[0035] The process of obtaining the meteorological satellite lunar observation image dataset is as follows: Use STK software to simulate the orbit of meteorological satellites, calculate the time when the moon appears in the satellite's cold sky field of view based on the geometric relationship between the satellite, the earth, the sun, and the moon, and collect meteorological satellite lunar observation images.

[0036] Meteorological satellites include but are not limited to Himawari-8. In Satellite Tool Kit 10 (STK 10), the orbit of Himawari-8 is simulated based on the two-line orbital data (Two-Line Orbital Element, TLE) of the Himawari Imager (AHI) carried by Himawari-8. Based on the geometric relationship between satellite, earth, sun and moon, the time when the moon appears in the satellite's cold sky field of view is calculated, and the moon imaging time and the corresponding moon phase angle are obtained.

[0037] The meteorological satellite lunar observation images with the lunar phase angle in the range of [-92°, 92°] are screened to ensure that the illuminated part of the moon is greater than 50%, and at the same time consistent with the lunar phase angle range of the ROLO model. The meteorological satellite lunar observation images with complete lunar regions are retained, thereby eliminating the lunar images in the following situations: (1) Part of the Moon is outside the field of view, and only part of it is imaged within the field of view; (2) The Moon is partially obscured by the Earth; (3) The Moon is cut off by the boundary between the scanning strips; the Moon is within 0.1° of the Earth’s periphery; (4) A combination of any two or more of the above three situations.

[0038] The process of obtaining the ground-based lunar observation image dataset is as follows: A hyperspectral imager is used on the ground to observe the moon and obtain ground-based lunar observation images; the observation range of the hyperspectral imager is 380-1000nm, and the spectral resolution is ≥1nm.

[0039] When using a hyperspectral imager to observe the moon, choose a night far away from the city with a wide field of view, clear sky and no clouds, and a wind speed of less than level 3 (<5.4m / s). During the observation process, the atmospheric conditions are monitored in real time through mobile meteorological stations, which are used to perform atmospheric correction using meteorological data.

[0040] S2: Correct the meteorological satellite lunar observation image dataset and the ground-based lunar observation image dataset respectively to obtain the standard meteorological satellite observation lunar radiance and ground-based observation lunar radiance.

[0041] The process of correcting the meteorological satellite lunar observation image dataset to obtain the standard meteorological satellite lunar radiation brightness is as follows: S21a: Geometric correction.

[0042] Make geometric corrections to the lunar radiance observed by meteorological satellites and correct the lunar radiance observed by meteorological satellites to the standard distance.

[0043]

[0044] In the formula, It represents the lunar radiance observed by meteorological satellite after geometric correction; It represents the lunar radiance observed by meteorological satellites before geometric correction; Indicates the distance between the sun and the moon; Indicates the distance between the weather satellite and the moon; represents the average distance between the Earth and the Moon, ; 1 AU represents the average distance between the Earth and the Sun, .

[0045] S22a: Use effective wavelengths instead of spectral channels.

[0046] The effective wavelength is used as a rough approximation of the spectral channel (ensuring that the ratio of the spectral channel bandwidth to the effective wavelength is ≤ 0.2), so that the spectral data between different platforms can be compared and integrated.

[0047]

[0048] In the formula, Indicates the effective wavelength; Indicates the starting wavelength of the spectral channel; Indicates the end wavelength of the spectral channel; represents the spectral response function (SRF); represents the lunar soil reflectivity reference curve, Represents the reference solar spectrum irradiance curve, in units of ; Represents the lunar solid angle in units of .

[0049] S23a: Lunar boundary extraction.

[0050] Firstly, the threshold algorithm is used to extract the rectangular area containing the moon from the meteorological satellite lunar observation image, and then the Sobel edge detection algorithm is used to detect the lunar boundary in the extracted rectangular area.

[0051] The connected area with pixel values ​​greater than the threshold is set as the rough range of the moon, and the moon is placed in the center to form a rectangular area that includes the moon and part of the surrounding night sky background. Then the Sobel edge detection algorithm is used to further refine the boundary of the moon.

[0052] S24a: Night sky noise correction.

[0053] The average radiance value of the area from the 15th pixel to the 30th pixel width outside the edge of the moon is selected as the night sky noise value, and this night sky noise value is subtracted from the meteorological satellite lunar observation image;

[0054] In the formula, represents the lunar radiance observed by meteorological satellites after correction for night sky noise, Represents the night sky noise value.

[0055] The process of correcting the ground-based lunar observation image dataset to obtain the standard ground-based observation lunar radiance is as follows: S21b: Geometric correction.

[0056] Make geometric corrections to the lunar radiance observed from the ground, and correct the lunar radiance observed from the ground to the standard distance;

[0057] In the formula, represents the ground-based lunar radiance after geometric correction, represents the ground-based lunar radiance before geometric correction, represents the distance between the sun and the moon, represents the distance between the ground position and the moon, It represents the average distance between the Earth and the Moon, and 1 AU represents the average distance between the Earth and the Sun.

[0058] S22b: Atmospheric correction.

[0059] The LBLRTM (Line-By-Line Radiative Transfer Model) model is used to simulate atmospheric transmittance and eliminate atmospheric effects. Multi-source meteorological and atmospheric data are input into the LBLRTM model to simulate the atmospheric transmittance of 380nm-1000nm, and atmospheric correction is performed on the ground-based observed lunar radiance to obtain the atmospherically corrected ground-based observed lunar radiance, that is, the ground-based observed lunar radiance at the top of the atmosphere.

[0060] Meteorological and atmospheric data include temperature, air pressure, wind speed, water vapor profile, aerosol profile and carbon dioxide content. The temperature, air pressure and wind speed are derived from the synchronous detection data of mobile meteorological stations, and the water vapor profile, aerosol profile and carbon dioxide content are derived from the ERA5 dataset.

[0061] ; In the formula, represents the ground-based observed lunar radiance after atmospheric correction, Indicates atmospheric transmittance.

[0062] S23b: Lunar boundary extraction.

[0063] Firstly, a threshold algorithm is used to extract the rectangular area containing the moon from the ground-based lunar observation image, and then the Sobel edge detection algorithm is used to detect the lunar boundary in the extracted rectangular area.

[0064] S24b: Night sky noise correction.

[0065] A pure night sky area of ​​the ground-based lunar observation image that does not include the moon is selected as a night sky noise value, and the night sky noise value is subtracted from the ground-based lunar observation image;

[0066] In the formula, represents the ground-based observation of the moon’s radiance after correction for night sky noise, Represents the night sky noise value.

[0067] S3: The ground-based observed lunar radiation brightness of all pixels within the lunar boundary in the ground-based lunar observation image is integrated and summed to obtain the ground-based observed irradiance, and the meteorological satellite observed lunar radiation brightness of all pixels within the lunar boundary in the meteorological satellite lunar observation image is integrated and summed to obtain the meteorological satellite observed irradiance. The ground-based observed irradiance and the meteorological satellite observed irradiance are collectively referred to as the observed lunar irradiance.

[0068] The ground-based lunar observation irradiance is obtained by integrating and summing the ground-based lunar observation brightness of all pixels within the lunar boundary in the ground-based lunar observation image. The calculation formula is as follows: ; In the formula, represents the ground-based observation irradiance, represents the lunar radiance observed by meteorological satellites after correction for night sky noise, represents the solid angle of the pixel, and They represent the number of rows and columns of meteorological satellite lunar observation images and ground-based lunar observation images respectively.

[0069] The meteorological satellite observed lunar radiation brightness of all pixels within the lunar boundary in the meteorological satellite lunar observation image is integrated and summed to obtain the meteorological satellite observed irradiance. The calculation formula is as follows: ; In the formula, represents the irradiance observed by meteorological satellite, Represents the ground-based lunar radiance after correction for night sky noise.

[0070] Ground-based irradiance observation Observing irradiance with meteorological satellites Observed lunar irradiance , that is, in the following content, the lunar irradiance is observed Refers to the irradiance observed by meteorological satellites Or it refers to ground-based observation irradiance .

[0071] S4: Solve the geometric parameters of the observation point, the Earth, the Moon, and the Sun at the time of ground-based and meteorological satellite observations, and calculate the lunar irradiance of the ROLO model.

[0072] Step S4 specifically includes the following steps: S41: Calculate the lunar albedo for the ROLO model.

[0073] Solve the geometric parameters of the observation point, the earth, the moon, and the sun at the time of ground-based and meteorological satellite observation, input the geometric parameters into the ROLO model, and calculate the lunar albedo of the ROLO model corresponding to the ground-based and meteorological satellite observation times; the geometric parameters include the moon phase angle, the longitude of the sun in the lunar coordinate system, the longitude of the observation point in the lunar coordinate system, and the latitude of the observation point in the lunar coordinate system;

[0074]

[0075] In the formula, represents the lunar albedo of the ROLO model, Indicates the band, Indicates the moon phase angle, represents the longitude of the sun in the lunar coordinate system, represents the latitude of the observation point in the lunar coordinate system, represents the longitude of the observation point in the lunar coordinate system, , , , , , , , , , , , , They represent the coefficient terms respectively.

[0076] Ground-based observation points refer to locations on the ground, and meteorological satellite observation points refer to the meteorological satellites themselves.

[0077] Moon albedo for the ROLO model It is actually the multispectral lunar albedo of the ROLO model. In order to obtain the hyperspectral ROLO model lunar irradiance, the average near-side hyperspectral reflectance of the M3 data is used as a reference instead of the Apollo measured reflectance, and the multispectral lunar albedo of the ROLO model is interpolated and fitted to obtain the hyperspectral lunar albedo of the ROLO model.

[0078] M3 data: lunar surface reflectivity data in the 430-3000nm band provided by the Moon Mineralogy Mapper M3.

[0079] The original ROLO model uses the laboratory hyperspectral albedo of the Apollo samples to interpolate the ROLO multispectral albedo to obtain the ROLO hyperspectral albedo. However, the spectrum of the Apollo samples measured in the laboratory may not be able to describe the most realistic features of the lunar surface. As a result, the ROLO model easily underestimates the overall radiation intensity of the moon. In the present invention, the average near-side reflectance spectrum of M3 is used to adjust the ROLO model. M3 obtains 85 bands of high-spectral, high-spatial-resolution lunar data, covering more than 95% of the lunar surface. Compared with the laboratory reflectance measurement results of Apollo, which may be interfered with, the reflectance spectrum of M3 data more accurately reflects the true reflectance of the entire moon. Therefore, the average near-side reflectance spectrum of M3 data is more suitable for fitting the ROLO hyperspectral albedo.

[0080] S42: Calculate the lunar irradiance of the ROLO model.

[0081] According to the hyperspectral lunar albedo of the ROLO model, the formula for calculating the lunar irradiance of the ROLO model is as follows:

[0082] In the formula, represents the lunar irradiance of the ROLO model, represents the distance correction factor, It represents the solid angle of the moon relative to the earth at the standard distance. Represents solar radiation intensity.

[0083] S5: The ROLO model is improved by using polynomial fitting to find the relative deviation between the observed lunar irradiance in different bands and different lunar phases and the ROLO model lunar irradiance.

[0084] The calculation formula for the relative deviation between the observed lunar irradiance and the ROLO model lunar irradiance is:

[0085] In the formula, It represents the relative deviation between the observed lunar irradiance at the k-band and the lunar irradiance of the ROLO model. The k-band selects the common bands of the meteorological satellite lunar observation images, the ground-based lunar observation images and the ROLO model. The common bands are 470nm, 510nm, 640nm, 855nm, 1610nm and 2255nm. The k-band can select any one of these six bands. represents the observed lunar irradiance at the k-band (ground-based observation irradiance or meteorological satellite observation irradiance), represents the ROLO model lunar irradiance at the k-band.

[0086] like Figure 2 As shown in the figure, as the moon phase angle increases, the error of the ROLO model increases accordingly. Near the moon phase angle of ±92°, the error of the ROLO model reaches 2%-8%. And this moon phase angle dependence is related to the wavelength. The moon phase angle dependence is the smallest at 640nm and gradually increases towards 470nm and 2255nm.

[0087] The relative deviation of each k-band under different lunar phase angles is fitted by polynomials to correct the phase angle dependence of the ROLO model under different k-bands and different lunar phase angles, thus improving the ROLO model.

[0088] , ; ,

[0089]

[0090] ,

[0091] In the formula, and They represent the lunar irradiance of the improved ROLO model under negative and positive lunar phase angles, respectively; and They represent the improvement coefficients of the ROLO model under negative and positive moon phase angles, respectively; It represents the solid angle of the moon relative to the earth at standard distance; represents the solar radiation intensity; represents the distance correction factor; and All represent coefficients and are fixed values ​​(see Tables 1 and 2).

[0092] Table 1 Coefficients in the improved ROLO model Value

[0093] Table 2 Coefficients in the improved ROLO model Value

[0094] The present invention uses a random sampling method to divide the observed lunar irradiance into a modeling data set and a verification data set. The modeling data set is used to fit the relative deviation of each k-band at different lunar phase angles, correct the phase angle dependence of the ROLO model at different k-bands and different lunar phase angles, and improve the ROLO model. The verification data set is used to verify the relative deviation between the observed lunar irradiance and the lunar irradiance of the improved ROLO model.

[0095] like Figure 3 As shown in the figure, near the lunar phase angle of ±92°, the error of the improved ROLO model is reduced to 0.27%-0.39%, and the relative deviations of different lunar phase angles are evenly distributed around 0, indicating that the prediction accuracy of the improved ROLO model is significantly improved and the error of phase angle dependence is reduced.

[0096] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0097] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for improving the ROLO model by combining ground-based and meteorological satellite lunar observation data, characterized in that: The steps include: S1: Acquire meteorological satellite lunar observation image datasets and ground-based lunar observation image datasets; S2: Correct the meteorological satellite lunar observation image dataset and the ground-based lunar observation image dataset respectively to obtain the standard meteorological satellite observation lunar radiance and ground-based observation lunar radiance; S3: integrating and summing the ground-based observed lunar radiance of all pixels within the lunar boundary in the ground-based lunar observation image to obtain the ground-based observed irradiance, and integrating and summing the meteorological satellite observed lunar radiance of all pixels within the lunar boundary in the meteorological satellite lunar observation image to obtain the meteorological satellite observed irradiance. The ground-based observed irradiance and the meteorological satellite observed irradiance are collectively referred to as the observed lunar irradiance. S4: Solve the geometric parameters of the observation point, the Earth, the Moon, and the Sun at the time of ground-based and meteorological satellite observations, and calculate the lunar irradiance of the ROLO model; S5: The ROLO model is improved by using polynomial fitting to find the relative deviation between the observed lunar irradiance in different bands and different lunar phases and the ROLO model lunar irradiance.

2. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 1 is characterized in that: In step S1, the process of obtaining the meteorological satellite lunar observation image dataset is as follows: Use STK software to simulate the orbit of meteorological satellites, calculate the time when the moon appears in the satellite's cold sky field of view based on the geometric relationship between the satellite, the earth, the sun, and the moon, and collect meteorological satellite lunar observation images.

3. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 1 is characterized in that: In step S1, the process of obtaining the ground-based lunar observation image dataset is as follows: A hyperspectral imager is used on the ground to observe the moon and obtain ground-based lunar observation images; the observation range of the hyperspectral imager is 380nm-1000nm, and the spectral resolution is ≥1nm.

4. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 2 is characterized in that: Images with a complete lunar region and a lunar phase angle in the range of [-92°, 92°] were screened from the collected meteorological satellite lunar observation data to form a meteorological satellite lunar observation image dataset.

5. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 1 is characterized in that: In step S2, the meteorological satellite lunar observation image data set is corrected to obtain the standard meteorological satellite lunar radiance brightness as follows: S21a: geometric correction of the lunar radiance observed by meteorological satellites, correcting the lunar radiance observed by meteorological satellites to the standard distance; In the formula, represents the lunar radiance observed by meteorological satellites after geometric correction, It represents the lunar radiance observed by meteorological satellites before geometric correction. represents the distance between the sun and the moon, represents the distance between the meteorological satellite and the moon, represents the average distance between the Earth and the Moon, and 1 AU represents the average distance between the Earth and the Sun; S22a: Use effective wavelengths instead of spectral channels; In the formula, represents the effective wavelength, represents the starting wavelength of the spectral channel, represents the end wavelength of the spectral channel, represents the spectral response function, represents the lunar soil reflectivity reference curve, represents the reference solar spectrum irradiance curve, represents the lunar solid angle; S23a: First, a threshold algorithm is used to extract a rectangular area containing the moon from the meteorological satellite lunar observation image, and then a Sobel edge detection algorithm is used to detect the lunar boundary in the extracted rectangular area; S24a: Select the average radiance value of the area from the 15th pixel to the 30th pixel width outside the edge of the moon as the night sky noise value, and subtract the night sky noise value from the meteorological satellite lunar observation image; In the formula, represents the lunar radiance observed by meteorological satellites after correction for night sky noise, Represents the night sky noise value.

6. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 1 is characterized in that: In step S2, the ground-based lunar observation image dataset is corrected to obtain the standard ground-based lunar radiance as follows: S21b: geometric correction of the lunar radiance observed by ground-based observations, and correction of the lunar radiance observed by ground-based observations to the standard distance; In the formula, represents the ground-based lunar radiance after geometric correction, represents the ground-based lunar radiance before geometric correction, represents the distance between the sun and the moon, represents the distance between the ground position and the moon, represents the average distance between the Earth and the Moon, and 1 AU represents the average distance between the Earth and the Sun; S22b: Use the LBLRTM model to simulate atmospheric transmittance and perform atmospheric correction on the ground-based lunar radiance observations based on the atmospheric transmittance. ； In the formula, represents the ground-based observed lunar radiance after atmospheric correction, represents the atmospheric transmittance; S23b: firstly, a threshold algorithm is used to extract a rectangular region containing the moon from the ground-based lunar observation image, and then a Sobel edge detection algorithm is used to detect the lunar boundary in the extracted rectangular region; S24b: selecting a pure night sky area of ​​the ground-based lunar observation image that does not include the moon as a night sky noise value, and subtracting the night sky noise value from the ground-based lunar observation image; In the formula, represents the ground-based lunar radiance after night sky noise correction, Represents the night sky noise value.

7. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 1 is characterized in that: In step S3, the ground-based lunar observation radiance of all pixels within the lunar boundary in the ground-based lunar observation image is integrated and summed to obtain the ground-based observation irradiance. The calculation formula is as follows: ； In the formula, represents the ground-based observation irradiance, represents the lunar radiance observed by meteorological satellites after correction for night sky noise, represents the solid angle of the pixel, and Respectively represent the number of rows and columns of meteorological satellite lunar observation images and ground-based lunar observation images; The meteorological satellite observed lunar radiation brightness of all pixels within the lunar boundary in the meteorological satellite lunar observation image is integrated and summed to obtain the meteorological satellite observed irradiance. The calculation formula is as follows: ； In the formula, represents the irradiance observed by meteorological satellite, Represents the ground-based lunar radiance after correction for night sky noise.

8. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 1 is characterized in that: Step S4 specifically includes the following steps: S41: solving geometric parameters of the observation point, the earth, the moon, and the sun at the ground-based and meteorological satellite observation times, inputting the geometric parameters into the ROLO model, and calculating the lunar albedo of the ROLO model corresponding to the ground-based and meteorological satellite observation times; wherein the geometric parameters include the moon phase angle, the longitude of the sun in the lunar coordinate system, the longitude of the observation point in the lunar coordinate system, and the latitude of the observation point in the lunar coordinate system; In the formula, represents the lunar albedo of the ROLO model, Indicates the band, Indicates the moon phase angle, represents the longitude of the sun in the lunar coordinate system, represents the latitude of the observation point in the lunar coordinate system, represents the longitude of the observation point in the lunar coordinate system, , , , , , , , , , , , , Respectively represent coefficient terms; observation points are ground locations or meteorological satellites; S42: Calculate the lunar irradiance of the ROLO model according to the lunar albedo of the ROLO model; In the formula, represents the lunar irradiance of the ROLO model, represents the distance correction factor, It represents the solid angle of the moon relative to the earth at the standard distance. Represents solar radiation intensity.

9. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 1, characterized in that: In step S5, the calculation formula for the relative deviation between the observed lunar irradiance and the ROLO model lunar irradiance is: In the formula, It represents the relative deviation between the observed lunar irradiance at the k-band and the lunar irradiance of the ROLO model. The k-band selects the common band of the meteorological satellite lunar observation image, the ground-based lunar observation image and the ROLO model, specifically any one of the bands of 470nm, 510nm, 640nm, 855nm, 1610nm and 2255nm. represents the observed lunar irradiance at k-band, represents the ROLO model lunar irradiance at the k-band.

10. The ROLO model improvement method combining ground-based and meteorological satellite lunar observation data according to claim 9 is characterized in that: In step S5, a polynomial is used to fit the relative deviation of each k-band at different moon phase angles, and the phase angle dependence of the ROLO model at different k-bands and different moon phase angles is corrected to improve the ROLO model; ， ； ， ， In the formula, and They represent the lunar irradiance of the improved ROLO model under negative and positive lunar phase angles, respectively; and They represent the improvement coefficients of the ROLO model under negative and positive moon phase angles, respectively; It represents the solid angle of the moon relative to the earth at standard distance; represents the solar radiation intensity; represents the lunar albedo of the ROLO model, represents the distance correction factor; and Both represent coefficients and are fixed values.

Images