Geometric simulation and registration method for infrared hyperspectral atmospheric sounding instrument

Abstract

The present application relates to the technical field of satellite remote sensing, and particularly relates to a kind of infrared hyperspectral atmospheric probe geometry simulation and registration method, by obtaining the detection time of HIRAS one pixel FOV, ground longitude, ground latitude and satellite position, the view vector of the pixel under the reference ellipsoid coordinate system is constructed;Rodriguez rotation formula is used, based on the size of pixel field of view, the imaging range of HIRAS pixel on the ground is simulated, the pixel of MERSI in the same platform is searched using KD-tree, according to the spectral convolution function and blackbody radiation formula, the equivalent brightness temperature of HIRAS pixel and the equivalent brightness temperature of MERSI simulation image are calculated respectively;The root mean square error of brightness temperature difference between HIRAS pixel and MERSI pixel is accumulated and counted by multiple rows, the root mean square error is arranged according to the FOR row timing, and the geometric positioning accuracy of HIRAS in the along-track and across-track directions is obtained.The present application realizes the fast search of the nearest pixel in the range of view vector, and realizes the accurate position determination of registration error.

Description

A geometric simulation and registration method for an infrared hyperspectral atmospheric sounder Technical Field

[0001] This invention relates to the field of satellite remote sensing technology, and in particular to a geometric simulation and registration method for an infrared hyperspectral atmospheric sounder. Background Technology

[0002] The L1 data from various meteorological satellite infrared payloads, such as the Hyperspectral Infrared Atmospheric Sounder (HIRAS) and the Medium Resolution Spectral Imager (MERSI), exhibit geometrical differences in location. This makes it difficult to accurately match data from different sources under a unified spatial reference during fusion applications or time-series analysis. L1 data, in particular, is the foundational data that has undergone geolocation and radiometric calibration. This limits the effectiveness of high-precision, consistent infrared data products in the in-depth application of core meteorological services such as numerical weather prediction models and climate analysis.

[0003] Medium-resolution imaging (MERSI) is a broadband imaging technique with resolutions of 250m and 1km. MERSI typically utilizes the 250m resolution visible light band for image registration with higher resolution images, achieving sub-pixel level geometric accuracy. Furthermore, the geometric relationship between MERSI's visible and infrared channels is stable, resulting in high geometric accuracy in the infrared band. HIRAS has a resolution of 14km per pixel's FOV (Field of View), significantly larger than a MERSI pixel, and HIRAS pixels are not continuously distributed, with gaps existing between pixel FOVs. Therefore, the infrared channels of MERSI can be used to verify the infrared pixel geometric positioning accuracy of HIRAS.

[0004] For example, patent application number 201910461383.5 discloses a method and apparatus for verifying the positioning consistency of infrared hyperspectral multi-pixel payloads. It discusses obtaining the deviation of low spatial resolution infrared hyperspectral multi-pixel payloads based on a high spatial resolution infrared hyperspectral image from the same platform. Some steps are similar to this patent, but the core process is different. This patent broadly provides a criterion that the angle between the view vector of the second pixel in the ECEF coordinate system and the view vector of the first pixel is less than 1 / 2. In actual operation, due to the large number of image pixels, performing a criterion for each pixel in a high-resolution image would require enormous computational time. This project provides a fast search scheme and proposes several strategies to reduce registration time. This patent application does not mention using the radiation brightness temperature difference between two matched samples as the basis for RMSE calculation, lacking considerable detail.

[0005] For example, the existing paper "Evaluation of HIRAS-II Positioning and Calibration Accuracy of FY-3E Same Platform Imager" discusses the method of positioning accuracy evaluation. However, this paper does not provide a method to accelerate pixel determination, nor does it mention how to use the view vector of the hyperspectral infrared atmospheric detector to simulate the process of pixel projection range on the ground.

[0006] Therefore, proposing a geometric simulation and registration method for infrared hyperspectral atmospheric detectors based on KD-tree search is of great practical significance. This method enables rapid search for the nearest pixel within the range of the view vector, thereby improving the registration speed and error calculation accuracy of the entire image. Summary of the Invention

[0007] To address the aforementioned issues, this invention provides a geometric simulation and registration method for an infrared hyperspectral atmospheric sounder. By constructing a KD-tree search method, it achieves rapid searching for the nearest pixel within the view vector range, thereby avoiding brute-force comparison of all pixels and reducing the time complexity from O(N) to an average of O(log N), thus improving the registration speed of the entire image. Furthermore, it sets sampling windows and step sizes. When the positioning accuracy of the original image is poor, a larger matching window and step size can be set to reduce registration time. When the accuracy after image correction is high, a smaller window and step size can be set to improve registration accuracy. The invention also clarifies that the simulated image and the original image are compared using temperature differences, and employs a multi-row overlay method to improve the accuracy of error calculation, achieving precise location determination of registration errors without the need for filtering the original data.

[0008] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0009] This invention provides a geometric simulation and registration method for an infrared hyperspectral atmospheric sounder, comprising the following steps:

[0010] The detection time, ground longitude, ground latitude, and satellite position of a single pixel (FOV) from the Infrared Hyperspectral Atmospheric Sounder (HIRAS) are obtained, and the view vector of the pixel in the reference ellipsoidal coordinate system is constructed.

[0011] The imaging range of HIRAS pixels on the ground is simulated based on the pixel field of view using the Rodriguez rotation formula. The imaging range is formed by calculating the intersection points of multiple conical vectors with the ground to form a polygonal region.

[0012] Based on the imaging range, KD-tree is used to search for pixels of the same platform medium resolution imager (MERSI), including constructing an index of MERSI latitude and longitude data, reading block images, calculating MERSI pixel view vectors, and filtering pixels located within the HIRAS imaging range.

[0013] Based on the spectral convolution function and the blackbody radiation formula, the equivalent brightness temperature of the HIRAS pixel and the equivalent brightness temperature of the MERSI simulation image are calculated respectively.

[0014] The MERSI image is translated and registered by accumulating the root mean square error of the brightness temperature difference between HIRAS and MERSI pixels across multiple rows.

[0015] The root mean square error is arranged in a time sequence according to FOR (Field Of Regard) to obtain the geometric positioning accuracy of HIRAS in the along-track and across-track directions.

[0016] As a further aspect of the present invention, constructing the view vector of the pixel in the reference ellipsoidal coordinate system includes the following steps:

[0017] Based on the detection time of the obtained pixel , ground longitude Ground latitude and satellite position Convert latitude and longitude to ground point coordinates in the Earth-fixed system. , The three-dimensional coordinates in the Earth-fixed coordinate system are:

[0018] ;

[0019] ;

[0020] ;

[0021] in, ;

[0022] In the formula, Elevation represents the height of the observation point relative to the reference ellipsoid. The radius of curvature of the circumpolar region is related to latitude. The semi-major axis of the reference ellipsoid is the WGS84 ellipsoid. The square of the first eccentricity of the ellipsoid; Indicates matrix transpose; Represents ground coordinates;

[0023] Calculate and normalize the view vector of the HIRAS FOV (Hyperspectral Infrared Atmospheric Sounder Field of View). In the formula, This represents the unit direction vector from the satellite to the ground point, used to represent the view vector of the HIRAS FOV; The satellite position is represented by the three-dimensional coordinate vector of the satellite in the Earth-fixed coordinate system. ; Coordinates of ground points under the Earth-Fixed System .

[0024] As a further aspect of the present invention, the Rodriguez rotation formula is used to simulate the imaging range of HIRAS pixels on the ground based on the pixel field of view size, including the following steps:

[0025] Based on the size of the pixel's field of view angle Construct any initial vector satisfy ;

[0026] Using Rodriguez's rotation formula Construct a conical vector that wraps around the view vector, where, The rotation angle is set to 10° to obtain 36 vectors; It is a conical vector;

[0027] All conical vectors, satellite positions, and ground intersection points together form a collinear equation, and the ground points satisfy the ellipsoidal equation of the reference ellipsoid.

[0028] The ground latitude and longitude corresponding to multiple conical vectors are solved sequentially, and the polygon formed by connecting multiple sets of latitude and longitude is represented as the imaging range of the FOV on the ground.

[0029] As a further aspect of the present invention, the collinearity equation is:

[0030] ;

[0031] The equation of the ellipsoid is:

[0032] In the formula, The coordinates of the ground intersection point represent the three-dimensional coordinates of the intersection point of the conical vector and the Earth's ellipsoid. This is the three-dimensional coordinate vector of the satellite's position; It is a scaling factor; With reference to the semi-major axis of the ellipsoid, The semi-minor axis of the ellipsoid Elevation.

[0033] As a further aspect of the present invention, using a KD-tree to search for pixels in a medium-resolution imager on the same platform includes the following steps:

[0034] Read MERSI imagery and latitude / longitude data captured within the same time period, build an index for all latitude / longitude data using a KD-tree, and find the nearest MERSI latitude / longitude coordinates to the target FOV view vector using a nearest neighbor search to obtain the image-side coordinates. ;

[0035] Based on the image coordinates, segmented images are read according to the reading range. The view vector of each MERSI pixel is calculated based on the satellite orbital position at the pixel imaging time. ;

[0036] Calculate view vectors and The angle between them, and filter those with an angle less than that of the HIRAS view vector. If the pixel is smaller than the HIRAS pixel FOV size If the MERSI pixel is within the HIRAS imaging range, calculate all pixels within the block image range and retain the pixels that meet the conditions.

[0037] As a further aspect of the present invention, the formula for calculating the reading range is:

[0038] ;

[0039] in, For the range to be read, This indicates the satellite's orbital altitude; for example, FY3F is 836km. This refers to the resolution of the corresponding MERSI band, such as the 1km resolution of CH22; This is the scaling factor, usually around 2, used to ensure a balance between memory overflow and speed. The pixel field of view is the size.

[0040] As a further aspect of the present invention, the MERSI pixels of the MERSI image are: .

[0041] As a further aspect of the present invention, calculating the equivalent brightness temperature of HIRAS pixels and the equivalent brightness temperature of MERSI simulation images includes the following steps:

[0042] For HIRAS pixels, the spectral radiance of the HIRAS pixel is calculated using the continuous spectrum measured by HIRAS and combined with the MERSI spectral curve, employing a spectral convolution function: In the formula, The spectral radiance of a pixel, For the target at wavelength Spectral radiance at that location, wavelength Spectral response value at; and The band range represents the starting and ending wavelengths of the spectral band for integration.

[0043] The equivalent brightness temperature of HIRAS pixels is inverted using the blackbody radiation formula: In the formula, The equivalent brightness temperature of a HIRAS pixel. and The radiation constant is related to Planck's constant. The wave number corresponding to this band. The measured spectral radiance;

[0044] The temperature of each MERSI pixel is calculated using the blackbody radiation formula. Temperature inversion is then performed on all MERSI pixels within the HIRAS imaging range to calculate the equivalent brightness temperature of the MERSI simulation image. .

[0045] As a further aspect of the present invention, translational registration of MERSI images includes the following steps:

[0046] Comparison of HIRAS pixel equivalent brightness temperature Equivalent brightness temperature of MERSI pixels Differences as measured values ;

[0047] The window range set when MERSI pixels are translated. Internal translation, where, To register the window size, the step size is k, and the brightness temperature difference is calculated once for each translation.

[0048] The root mean square error (RMSE) of the brightness temperature difference at FOV is accumulated using a multi-row overlay method to obtain... The root mean square error matrix of size is calculated, and the position with the minimum value of the root mean square error matrix is ​​the optimal registration position.

[0049] Record the latitude and longitude information of the best registration position on MERSI, which is the actual latitude and longitude position of FOV in HIRAS at this time series. Compare it with the initial latitude and longitude to calculate the positioning error of the pixel.

[0050] Compared with existing technologies, the geometric simulation and registration method for infrared hyperspectral atmospheric detectors provided by this invention has the following advantages:

[0051] This invention improves the efficiency and accuracy of simulation verification. By employing a KD-tree search method for nearest pixel search, it achieves rapid searching for the nearest pixel within the view vector range. Based on the search results, it finds effective registration pixels within the field of view, improving the registration speed of the entire scene image. By introducing configurable sampling window size and step size, a dynamic balance between registration accuracy and efficiency is achieved. When the positioning accuracy of the original image is poor, a larger window and step size can be set for rapid registration; when the positioning accuracy is high, a smaller window and step size are used to improve detail accuracy. This invention also clarifies that the comparison between the simulated image and the original image is based on temperature difference. Brightness temperature inversion makes the registration results more reliable. By using multi-row superposition of the root mean square error of the brightness temperature difference, the precise location of the registration error is determined. Precise location of the registration error can be achieved without filtering the original data, realizing high-precision simulation of the HIRAS pixel imaging range.

[0052] These or other aspects of the invention will become more apparent from the following description of embodiments. It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. In the drawings:

[0054] Figure 1 is a flowchart of a geometric simulation and registration method for an infrared hyperspectral atmospheric detector according to the present invention.

[0055] Figure 2 is a schematic diagram of the overlap between the FOV conical surface of the infrared hyperspectral atmospheric detector and the mid-resolution imager in the geometric simulation and registration method of the infrared hyperspectral atmospheric detector of the present invention.

[0056] Figure 3 is a schematic diagram showing the correspondence between the simulated HIRAS imaging range and the MERSI image obtained in the geometric simulation and registration method of an infrared hyperspectral atmospheric detector according to the present invention.

[0057] Figure 4 is a schematic diagram of the calculated position and registration position in the geometric simulation and registration method of an infrared hyperspectral atmospheric detector according to the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0059] The technical solutions in the exemplary embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described exemplary embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] Referring to Figures 1 to 4, embodiments of this application also provide a geometric simulation and registration method for an infrared hyperspectral atmospheric sounder, comprising the following steps:

[0061] Step S10: Obtain the detection time, ground longitude, ground latitude and satellite position of a pixel (FOV) of the Infrared Hyperspectral Atmospheric Sounder (HIRAS), and construct the view vector of the pixel in the reference ellipsoidal coordinate system.

[0062] This step involves constructing the view vector of the pixel in the reference ellipsoidal coordinate system, including the following steps:

[0063] Step S101: Based on the obtained detection time of the pixel , ground longitude Ground latitude and satellite position Convert latitude and longitude to ground point coordinates in the Earth-fixed system. , The three-dimensional coordinates in the Earth-fixed coordinate system are:

[0064] ;

[0065] ;

[0066] ;

[0067] in, ;

[0068] In the formula, Elevation represents the height of the observation point relative to the reference ellipsoid. The radius of curvature of the circumpolar region is related to latitude. The semi-major axis of the reference ellipsoid is the WGS84 ellipsoid. The square of the first eccentricity of the ellipsoid; Indicates matrix transpose; Represents ground coordinates;

[0069] Step S102: Calculate the view vector of HIRAS FOV (Hyperspectral Infrared Atmospheric Sounder Field of View) and normalize it.

[0070] Among them, view vector In the formula, This represents the unit direction vector from the satellite to the ground point, used to represent the view vector of the HIRAS FOV; The satellite position is represented by the three-dimensional coordinate vector of the satellite in the Earth-fixed coordinate system. ; Coordinates of ground points in the Earth-Fixed System .

[0071] Step S20: Using the Rodriguez rotation formula, the imaging range of HIRAS pixels on the ground is simulated based on the pixel field of view size. The imaging range is formed by calculating the intersection points of multiple conical vectors with the ground to form a polygonal region.

[0072] In this step, the Rodriguez rotation formula is used to simulate the imaging range of HIRAS pixels on the ground based on the pixel field of view size, including the following steps:

[0073] Step S201: Based on the pixel field of view angle size Construct any initial vector satisfy .

[0074] Step S202: The root is rotated using the Rodrigues formula. Construct a conical vector that wraps around the view vector, where, The rotation angle is set to 10° to obtain 36 vectors; It is a conical vector.

[0075] Among them, the initial vector about the axis of rotation Rotation angle To obtain the next vector, when the interval is 10°, 36 conical vectors can be obtained.

[0076] Step S203: All conical vectors, satellite positions, and ground intersection points together form a collinear equation, and the ground points satisfy the ellipsoid equation of the reference ellipsoid.

[0077] The collinear equation is:

[0078] ;

[0079] The equation of the ellipsoid is:

[0080] In the formula, The coordinates of the ground intersection point represent the three-dimensional coordinates of the intersection point of the conical vector and the Earth's ellipsoid. This is the three-dimensional coordinate vector of the satellite's position; It is a scaling factor; With reference to the semi-major axis of the ellipsoid, The semi-minor axis of the ellipsoid Elevation.

[0081] Step S204: Solve the ground latitude and longitude corresponding to multiple conical vectors in sequence, and connect the multiple sets of latitude and longitude in sequence to form a polygon. The polygon represents the imaging range of the FOV on the ground.

[0082] By combining the above collinearity equation and ellipsoid equation formula, the ground latitude and longitude corresponding to multiple conical vectors can be solved sequentially. The polygon formed by connecting multiple sets of latitude and longitude can be represented as the imaging range of the FOV on the ground, thereby realizing the geometric positioning simulation of HIRAS pixels.

[0083] Step S30: Based on the imaging range, use KD-tree to search for pixels of the same platform Medium Resolution Imager (MERSI), including constructing an index of MERSI latitude and longitude data, reading block images, calculating MERSI pixel view vectors, and filtering pixels located within the HIRAS imaging range.

[0084] This step involves using a KD-tree to search for pixels in a medium-resolution imager on the same platform, including the following steps:

[0085] Step S301: Read MERSI images and latitude / longitude data captured in the same time period, build an index for all latitude / longitude data using a KD-tree, and find the MERSI latitude / longitude closest to the target FOV view vector using nearest neighbor search to obtain the image coordinates. .

[0086] Step S302: Read the segmented image according to the image coordinates within the reading range, wherein the calculation formula for the reading range is:

[0087] ;

[0088] in, For the range to be read, This indicates the satellite's orbital altitude; for example, FY3F is 836km. This refers to the resolution of the corresponding MERSI band, such as the 1km resolution of CH22; This is the scaling factor, usually around 2, used to ensure a balance between memory overflow and speed. The pixel field of view is the size.

[0089] Step S303: Calculate the view vector of the MERSI pixel based on the satellite orbital position at the pixel imaging time. .

[0090] Step S304: Calculate the view vector and The angle between them, and filter those with an angle less than that of the HIRAS view vector. If the pixel is smaller than the HIRAS pixel FOV size If the MERSI pixel is within the HIRAS imaging range, calculate all pixels within the block image range and retain the pixels that meet the conditions.

[0091] Wherein, the MERSI pixels of the MERSI image are .

[0092] Step S40: Calculate the equivalent brightness temperature of the HIRAS pixel and the equivalent brightness temperature of the MERSI simulation image based on the spectral convolution function and the blackbody radiation formula.

[0093] This step involves calculating the equivalent brightness temperature of HIRAS pixels and the equivalent brightness temperature of the MERSI simulation image, including the following steps:

[0094] Step S401: For HIRAS pixels, using the continuous spectrum measured by HIRAS and combined with the MERSI spectral curve, calculate the spectral radiance of the HIRAS pixel using the spectral convolution function: In the formula, The spectral radiance of a pixel, For the target at wavelength Spectral radiance at that location, wavelength Spectral response value at; and The band range represents the starting and ending wavelengths of the spectral band for integration.

[0095] Step S402: Invert the equivalent brightness temperature of the HIRAS pixel according to the blackbody radiation formula: In the formula, The equivalent brightness temperature of a HIRAS pixel. and The radiation constant is related to Planck's constant. The wave number corresponding to this band. The measured spectral radiance;

[0096] Step S403: Calculate the temperature of each MERSI pixel using the blackbody radiation formula, perform temperature inversion on all MERSI pixels within the HIRAS imaging range, and calculate the equivalent brightness temperature of the MERSI simulation image. .

[0097] Step S50: Perform translation registration on the MERSI image by accumulating the root mean square error of the brightness temperature difference between HIRAS pixels and MERSI pixels across multiple rows.

[0098] This step involves translational registration of the MERSI image, including the following steps:

[0099] Step S501: Compare the equivalent brightness temperature of HIRAS pixels. Equivalent brightness temperature of MERSI pixels Differences as measured values ;

[0100] Step S502: Set the window range for MERSI pixels during translation. Internal translation, where, To register the window size, the step size is k, and the brightness temperature difference is calculated once for each translation.

[0101] Step S503: Accumulate and statistically analyze the root mean square error (RMSE) of the FOV brightness temperature difference using a multi-row overlay method to obtain... The root mean square error matrix of size is calculated, and the position with the minimum value of the root mean square error matrix is ​​the optimal registration position.

[0102] Step S504: Record the latitude and longitude information of the best registration position on MERSI, which is the actual latitude and longitude position of FOV in HIRAS at this time series. Compare it with the initial latitude and longitude to calculate the positioning error of the pixel.

[0103] Step S60: Arrange the root mean square error in a time sequence according to FOR (Field Of Regard) to obtain the geometric positioning accuracy of HIRAS in the along-track and across-track directions.

[0104] In this step, arranging the error according to the FOR line sequence allows us to obtain the geometric positioning accuracy of HIRAS in the along-rail and across-rail directions, which is beneficial for subsequent positioning accuracy verification and correction.

[0105] This invention improves the efficiency and accuracy of simulation verification. By employing a KD-tree search method for nearest pixel search, it achieves rapid searching for the nearest pixel within the view vector range. Based on the search results, it finds effective registration pixels within the field of view, improving the registration speed of the entire scene image. By introducing configurable sampling window size and step size, a dynamic balance between registration accuracy and efficiency is achieved. When the positioning accuracy of the original image is poor, a larger window and step size can be set for rapid registration; when the positioning accuracy is high, a smaller window and step size are used to improve detail accuracy. This invention also clarifies that the comparison between the simulated image and the original image is based on temperature difference. Brightness temperature inversion makes the registration results more reliable. By using multi-row superposition of the root mean square error of the brightness temperature difference, the precise location of the registration error is determined. Precise location of the registration error can be achieved without filtering the original data, realizing high-precision simulation of the HIRAS pixel imaging range.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A geometric simulation and registration method for an infrared hyperspectral atmospheric sounder, characterized in that, Includes the following steps: The detection time, ground longitude, ground latitude, and satellite position of a HIRAS pixel's field of view are obtained, and the view vector of the pixel in the reference ellipsoidal coordinate system is constructed. The Rodrigues rotation formula is used to simulate the imaging range of the HIRAS pixel on the ground based on the pixel's field of view. The imaging range is formed by calculating the intersection points of multiple conical vectors with the ground to form a polygonal region. Based on the imaging range, a KD-tree is used to search for MERSI pixels on the same platform. This includes constructing an index of MERSI latitude and longitude data, reading segmented images, calculating the view vectors of MERSI pixels, and filtering pixels located within the HIRAS imaging range. The equivalent brightness temperature of HIRAS pixels and the equivalent brightness temperature of the MERSI simulation image are calculated using the spectral convolution function and the blackbody radiation formula. The MERSI image is then translated and registered by accumulating the root mean square error of the brightness temperature difference between HIRAS and MERSI pixels across multiple rows. The root mean square error is arranged in a FOR row time sequence to obtain the geometric positioning accuracy of HIRAS in the along-track and cross-track directions.

2. The geometric simulation and registration method for an infrared hyperspectral atmospheric detector as described in claim 1, characterized in that, Constructing the view vector of the pixel in the reference ellipsoidal coordinate system includes the following steps: based on the obtained detection time of the pixel... , ground longitude Ground latitude and satellite position Convert latitude and longitude to ground point coordinates in the Earth-fixed system. ， The three-dimensional coordinates in the Earth-fixed coordinate system are: ； ； ;in, In the formula, Elevation represents the height of the observation point relative to the reference ellipsoid. The radius of curvature of the circumpolar region is related to latitude. The semi-major axis of the reference ellipsoid is the WGS84 ellipsoid. Let the square of the first eccentricity of the ellipsoid be ; calculate the view vector of the HIRAS FOV and normalize it, where the view vector is . In the formula, This represents the unit direction vector from the satellite to the ground point, used to represent the view vector of the HIRAS FOV; The satellite position is represented by the three-dimensional coordinate vector of the satellite in the Earth-fixed coordinate system. ； Coordinates of ground points in the Earth-Fixed System 。 3. The geometric simulation and registration method for an infrared hyperspectral atmospheric sounder as described in claim 2, characterized in that, Using the Rodriguez rotation formula, the imaging range of HIRAS pixels on the ground is simulated based on the pixel field of view size, including the following steps: [The steps are described in the original text, but the provided text is incomplete and cannot be translated accurately.] Construct any initial vector satisfy The Rodriguez rotation formula is used. Construct a conical vector that wraps around the view vector, where, The rotation angle is set to 10° to obtain 36 vectors; The vectors are conical vectors; all conical vectors, satellite positions, and ground intersections together form a collinear equation, and the ground points satisfy the ellipsoid equation of the reference ellipsoid; the ground latitude and longitude corresponding to multiple conical vectors are solved sequentially, and the polygon formed by connecting multiple sets of latitude and longitude is represented as the imaging range of the FOV on the ground.

4. The geometric simulation and registration method for an infrared hyperspectral atmospheric sounder as described in claim 3, characterized in that, The collinear equation is: The equation of the ellipsoid is: In the formula, The coordinates of the ground intersection point represent the three-dimensional coordinates of the intersection point of the conical vector and the Earth's ellipsoid. This is the three-dimensional coordinate vector of the satellite's position; It is a scaling factor; With reference to the semi-major axis of the ellipsoid, The semi-minor axis of the ellipsoid Elevation.

5. The geometric simulation and registration method for an infrared hyperspectral atmospheric sounder as described in claim 4, characterized in that, Searching for pixels from a medium-resolution imager on the same platform using a KD-tree involves the following steps: reading MERSI images and latitude / longitude data captured during the same time period; building an index for all latitude / longitude data using a KD-tree; and finding the nearest neighbor MERSI latitude / longitude coordinates to the target FOV view vector using a nearest neighbor search to obtain the image-side coordinates. Based on the image coordinates, the image is read in blocks according to the reading range. The view vector of the MERSI pixel is calculated based on the satellite orbital position at the pixel imaging time. ; Calculate view vector and The angle between them, and filter those with an angle less than that of the HIRAS view vector. If the pixel is smaller than the HIRAS pixel FOV size If the MERSI pixel is within the HIRAS imaging range, calculate all pixels within the block image range and retain the pixels that meet the conditions.

6. The geometric simulation and registration method for an infrared hyperspectral atmospheric sounder as described in claim 5, characterized in that, The formula for calculating the reading range is: ;in, For the range to be read, Indicates the satellite's orbital altitude; This represents the resolution of the corresponding MERSI band. This refers to the scaling ratio; The pixel field of view is the size.

7. The geometric simulation and registration method for an infrared hyperspectral atmospheric sounder as described in claim 6, characterized in that, The MERSI pixels of the MERSI image are 。 8. The geometric simulation and registration method for an infrared hyperspectral atmospheric sounder as described in claim 7, characterized in that, The calculation of the equivalent brightness temperature of HIRAS pixels and the equivalent brightness temperature of the MERSI simulation image includes the following steps: For HIRAS pixels, the spectral radiance of the HIRAS pixels is calculated using the continuous spectrum measured by HIRAS and combined with the MERSI spectral curve, employing a spectral convolution function. In the formula, The spectral radiance of a pixel, For the target at wavelength Spectral radiance at that location, wavelength Spectral response value at; and The band range represents the start and end wavelengths of the integrated spectral band; the equivalent brightness temperature of the HIRAS pixel is inverted according to the blackbody radiation formula: In the formula, The equivalent brightness temperature of a HIRAS pixel. and The radiation constant is related to Planck's constant. The wave number corresponding to this band. To measure the spectral radiance, the temperature of each MERSI pixel was calculated using the blackbody radiation formula. Temperature inversion was performed on all MERSI pixels within the HIRAS imaging range to calculate the equivalent brightness temperature of the MERSI simulation image. 。 9. The geometric simulation and registration method for an infrared hyperspectral atmospheric sounder as described in claim 8, characterized in that, Translation registration of MERSI images includes the following steps: comparing the equivalent brightness temperature of HIRAS pixels. Equivalent brightness temperature of MERSI pixels Differences as measured values ; Set the window range for MERSI pixels during translation Internal translation, where, To register the window size, a step size of k is used, and the brightness temperature difference is calculated after each translation. The root mean square error of the FOV brightness temperature difference is accumulated and statistically analyzed using a multi-row overlay method to obtain... The root mean square error matrix of size is calculated, and the minimum position of the root mean square error matrix is ​​the optimal registration position. The latitude and longitude information of the optimal registration position on MERSI is recorded, which is the actual latitude and longitude position of FOV in HIRAS at this time series. It is compared with the initial latitude and longitude to calculate the positioning error of the pixel.

Images