A method and related equipment for correcting pointing deviation in optical remote sensing satellite imaging

Abstract

This invention discloses a method and related equipment for correcting pointing deviation in optical remote sensing satellite imaging, relating to the field of satellite attitude control technology. The method includes: selecting a reference star sensor; using gyroscope data to compensate for time errors in the measurements of each star sensor, obtaining time-uniformed measurement data; updating the transformation relationship between the star sensor measurement coordinate systems accordingly, thereby unifying the attitude determination reference coordinate system with the reference star sensor measurement coordinate system; controlling the satellite to perform a multi-parameter imaging task based on this unified coordinate system to obtain actual imaging data; comparing this data with a simulation model to determine the error angle; establishing a mathematical model based on multiple sets of error angles; generating pointing deviation correction coefficients based on the model, and correcting the pointing deviation by updating the star sensor installation matrix on-orbit. This invention can ensure the imaging geometric quality of high-resolution remote sensing satellites under various operating conditions.

Description

Technical Field

[0001] This invention relates to the field of satellite attitude control technology, and in particular to a method and related equipment for correcting pointing deviation in optical remote sensing satellite imaging. Background Technology

[0002] After years of development, spaceborne remote sensing technology has made significant progress in spectral resolution, spatial resolution, and temporal resolution, achieving high-spectral, high-spatial-resolution, all-weather, and near-real-time Earth observation capabilities. However, the imaging pointing accuracy of high-resolution optical remote sensing satellites is affected by various complex factors during their on-orbit operation. During launch, satellites experience extreme mechanical environments, and after entering orbit, they undergo space thermal cycling, leading to differences between the satellite platform and payload mechanical structures and thermoelastic deformations that vary over time. These deformations cause the geometric relationship between onboard attitude measurement sensors (such as star sensors) and the optical camera payload to deviate from ground calibration values, introducing pointing deviations during Earth imaging. This deviation can cause target positions to shift within the camera's field of view, severely impacting the geometric quality of high-resolution images and potentially causing imaging mission failures in specific areas.

[0003] For satellite attitude determination and pointing control, existing technologies generally employ high-precision inertial attitude measurement units (IATUs) consisting of a star sensor and a gyroscope. The gyroscope provides continuous angular velocity measurements, while the star sensor provides discrete absolute attitude references. However, existing star sensor + gyroscope attitude determination schemes primarily address the attitude measurement of the satellite itself relative to the inertial space or orbital coordinate system. They do not specifically model and correct systematic pointing deviations caused by on-orbit dynamic deformation and related to external imaging geometry (such as Beta angle and geographical location). Research and mature engineering implementation methods for systematically integrating onboard attitude measurement data, actual on-orbit imaging results, simulation models, and high-precision maps to establish an accurate mathematical model of pointing deviation, and further achieving real-time, on-orbit closed-loop correction of pointing deviation through onboard software updates, remain insufficient.

[0004] Therefore, in order to effectively overcome the imaging pointing deviation caused by the on-orbit environment and improve the imaging geometric accuracy and reliability of high-resolution optical remote sensing satellites, a systematic method that can comprehensively utilize multi-source information and realize on-orbit calibration and correction is needed. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to address the shortcomings of the prior art, and specifically provides a method and related equipment for correcting pointing deviation in optical remote sensing satellite imaging, as detailed below: 1) In a first aspect, the present invention provides a method for correcting pointing deviation in optical remote sensing satellite imaging, comprising: selecting one of at least three star sensors configured on the satellite as a measurement reference star sensor according to a selection principle, wherein the satellite is equipped with a gyroscope as an inertial attitude measurement device; compensating for time errors in the measurement data of each star sensor using the measurement data of the gyroscope to obtain time-uniformed measurement data; updating the transformation relationship between the measurement coordinate systems used by each star sensor based on the time-uniformed measurement data; and unifying the reference coordinate system used to determine the satellite attitude with the measurement coordinate system used by the measurement reference star sensor based on the updated transformation relationship to establish a unified attitude measurement system. A unified attitude reference coordinate system is established. Based on this system, the satellite is controlled to perform imaging tasks at different Beta angles, latitudes, and longitudes to acquire actual imaging data from the satellite's optical cameras. The actual imaging data is compared with typical feature regions of the map loaded in the simulation model to determine the error angle between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical cameras. A mathematical model of the error angle is established based on the error angles acquired at different Beta angles, latitudes, and longitudes. Based on this mathematical model, pointing deviation correction coefficients are generated, and the installation matrix of each star sensor in the onboard software is updated via on-orbit annotation to correct the satellite's imaging pointing deviation. The beneficial effects are: by determining the reference star sensor based on selection principles and merging the reference coordinate system of the entire attitude determination system with the measurement coordinate system of that reference star sensor, a consistent and stable high-precision attitude measurement foundation is first constructed. Time synchronization compensation is performed on multi-star sensor measurements using gyroscope data, resulting in time-uniform measurement data. This allows for dynamic updates of the transformation relationships between the measurement coordinate systems of each star sensor, significantly improving the real-time on-orbit accuracy of the star sensor installation matrix. Multi-dimensional parameter imaging experiments were conducted on the satellite using a unified attitude reference coordinate system. The acquired actual imaging data was compared with simulation models and high-precision maps to accurately separate and quantify the systematic error angles between the unified attitude reference coordinate system and the optical camera's line-of-sight coordinate system. By mathematically modeling the error angle data obtained under different Beta angles, latitudes, and longitudes, a mathematical model accurately describing the variation law of pointing deviation can be established. Finally, pointing deviation correction coefficients were generated based on this model and updated in the on-board installation matrix, achieving systematic, on-orbit closed-loop correction of imaging pointing deviations, thereby ensuring the imaging geometric quality of the high-resolution remote sensing satellite under various operating conditions.

[0006] 2) In a second aspect, the present invention also provides an optical remote sensing satellite imaging pointing deviation correction system, comprising: a measurement reference star sensor selection module, a time error compensation module, a transformation relationship determination module, a coordinate system module, an actual imaging data acquisition module, an error angle acquisition module, a mathematical model establishment module, and a correction module; the measurement reference star sensor selection module is used to: select one of at least three star sensors configured on the satellite as the measurement reference star sensor according to the reference star sensor selection principle, wherein the satellite is equipped with a gyroscope as an inertial attitude measurement device; the time error compensation module is used to: compensate for the time error of the measurement data of each star sensor by using the measurement data of the gyroscope to obtain time-uniform measurement data; the transformation relationship determination module is used to: update the transformation relationship between the measurement coordinate systems used by each star sensor according to the time-uniform measurement data; the coordinate system module is used to: determine the satellite based on the updated transformation relationship. The attitude reference coordinate system is unified with the measurement coordinate system used by the reference star sensor to establish a unified attitude reference coordinate system. The actual imaging data acquisition module is used to: control the satellite to perform imaging tasks in different Beta angles, latitudes, and longitudes based on the unified attitude reference coordinate system, and acquire the actual imaging data of the optical camera equipped on the satellite. The error angle acquisition module is used to: compare the actual imaging data with the map loaded in the simulation model to determine the error angle between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera. The mathematical model establishment module is used to: establish a mathematical model of the error angle based on the error angles acquired in different Beta angles, latitudes, and longitudes. The correction module is used to: generate pointing deviation correction coefficients based on the mathematical model of the error angles, and update the installation matrix of each star sensor in the on-board software through on-orbit annotation to correct the satellite's imaging pointing deviation.

[0007] 3) In a third aspect, the present invention also provides an electronic device, the electronic device including a processor coupled to a memory, the memory storing at least one computer program, the at least one computer program being loaded and executed by the processor, so that the electronic device implements any of the above-mentioned optical remote sensing satellite imaging pointing deviation correction methods.

[0008] 4) In a fourth aspect, the present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements any of the above-described optical remote sensing satellite imaging pointing deviation correction methods.

[0009] It should be noted that the beneficial effects of the technical solutions of the second to fourth aspects of the present invention and their corresponding possible implementations can be found in the above description of the technical effects of the first aspect and its corresponding possible implementations, and will not be repeated here. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating a method for correcting pointing deviation in optical remote sensing satellite imaging according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of an optical remote sensing satellite imaging pointing deviation correction system according to an embodiment of the present invention. Detailed Implementation

[0011] like Figure 1 As shown, an embodiment of the present invention provides a method for correcting pointing deviation in optical remote sensing satellite imaging, comprising the following steps: S1. Based on the selection principles for the reference star sensor, one of the at least three star sensors configured on the satellite is selected as the measurement reference star sensor. The satellite is equipped with a gyroscope as an inertial attitude measurement device. The selection principles for the reference star sensor are a set of engineering criteria used to select the measurement reference star sensor from multiple star sensors. These criteria are based on satellite structural layout, mechanical stability, thermal deformation analysis, and optical environment interference assessment. The aim is to select the star sensor with the smallest relative change between its position and the line-of-sight coordinate system of the optical camera payload, and the weakest susceptibility to external interference, as the measurement reference. Specifically, the criteria include: the star sensor mounting bracket is directly connected to the camera load-bearing frame; the star sensor is installed at the closest distance to the optical camera payload; structural mechanics analysis shows the minimum structural deformation between the star sensor and the optical camera payload line-of-sight coordinate system throughout the satellite's entire mission lifespan; and simulation analysis shows the minimum cumulative time the star sensor is affected by Earth's atmospheric light, sunlight, and moonlight throughout the satellite's mission lifespan. These criteria comprehensively ensure that the measurement reference star sensor provides a highly stable and reliable attitude measurement reference. Therefore, S1 specifically includes: S10. The satellite is equipped with at least three star sensors, labeled Star Sensor A, Star Sensor B, and Star Sensor C. Simultaneously, the satellite is equipped with gyroscopes as inertial attitude measurement devices. The initial installation matrices of the star sensors and gyroscopes were obtained from the satellite structural design drawings and precise measurement results. The installation matrix of Star Sensor A is denoted as... The installation matrix of star sensor B is denoted as The mounting matrix of star sensor C is denoted as The mounting matrix of the gyroscope is denoted as The installation matrix describes the orientation of these devices relative to the satellite's body coordinate system.

[0012] S11. Evaluate each star sensor using the reference star sensor selection principle, specifically: ① Criterion 1: Check whether the star sensor mounting bracket is directly connected to the camera load-bearing frame. Confirm the connection method between each star sensor mounting bracket and the load-bearing frame of the optical camera payload by reviewing the satellite's mechanical structure assembly drawings. Direct connection is considered compliant, indirect connection is considered compliant. ② Criterion 2: Calculate the installation distance between the star sensor and the optical camera payload. The distance calculation formula is: ,in, This represents the Euclidean distance between the star sensor and the optical camera payload. , , These are the coordinates of the star sensor in the satellite's body coordinate system. , , These are the coordinates of the optical camera payload in the satellite's body coordinate system. The coordinate values ​​are obtained from the satellite's structural precision measurement data. A smaller distance indicates a weaker influence of thermal and mechanical deformation on the relative position. ③ Criterion three: Evaluate the structural deformation between the star sensor and the optical camera payload's line-of-sight coordinate system through structural mechanics analysis. Use finite element analysis software to simulate the satellite's structural response under launch, on-orbit thermal cycling, and other environments. Calculate the relative deformation angle between each star sensor and the optical camera payload's line-of-sight coordinate system. Assume the initial installation matrix of the star sensor is... The modified installation matrix is Then the transformation matrix is: From the deformation matrix The Euler angles are extracted as the deformation amount. The magnitude of the deformation angle is denoted as... The calculation formula is: ,in, , , These are the roll, pitch, and yaw Euler angles calculated from the deformation matrix. Smaller deformation indicates higher structural stability. ④ Criterion four: Analyze the cumulative time the star sensor is affected by terrestrial atmospheric light, sunlight, and moonlight through simulation. Using satellite orbit simulation tools, import the satellite orbit parameters and the star sensor's field of view model to simulate the time each star sensor's field of view is exposed to terrestrial atmospheric light, sunlight, and moonlight throughout the entire mission cycle. The cumulative exposure time is denoted as... The shorter the time, the lower the risk of stray light interference to the star sensor measurement.

[0013] S12. Quantify and score the evaluation results of each star sensor under the above criteria. Each criterion is assigned a weight, which is determined based on engineering experience. For example, the weight for direct connection to the mounting bracket is... The installation distance weight is The structural deformation weight is The cumulative time weight of light influence is The overall scoring formula is: ,in, This represents the total score. It is a binary variable; if the star sensor mounting bracket is directly connected to the camera's load-bearing frame, then... ,otherwise ; It is the installation distance; It is the amplitude of the structural deformation angle; This refers to the cumulative time of light influence. A higher score indicates that the star sensor is more suitable as a reference star sensor for measurement. The star sensor with the highest total score is selected as the reference star sensor for measurement. In engineering practice, if criterion one is a mandatory condition, star sensors directly connected to the mounting bracket are selected first, and then other criteria are applied for scoring.

[0014] S13. Label the selected star sensor as Star Sensor A, and the remaining star sensors as Star Sensor B and Star Sensor C. Record the installation matrix of Star Sensor A. and the mounting matrix of the gyroscope. These installation matrices will be used for subsequent time error compensation and installation matrix update processes.

[0015] S2. Time error compensation is performed on the measurement data of each star sensor using the gyroscope's measurement data to obtain time-uniformed measurement data. Specifically, using the angular velocity measurement value output by the gyroscope, the attitude change quaternion corresponding to the time difference between the measurement time of each non-reference star sensor and the measurement reference star sensor is calculated. This attitude change quaternion is then applied to the measurement data of the non-reference star sensors to unify the measurement data of each non-reference star sensor to the measurement time of the measurement reference star sensor, thus obtaining time-uniformed measurement data. The specific implementation process is as follows: S20. Acquire the raw measurement data and time information of the reference star sensor and non-reference star sensors. The reference star sensor is denoted as star sensor A. The non-reference star sensors are denoted as star sensor B and star sensor C. In each data processing cycle, read the following data from the onboard software: the attitude quaternion from the inertial frame measured by star sensor A to the coordinate system measured by star sensor A is denoted as... The corresponding data sampling time is denoted as Data validity identifier is denoted as The attitude quaternion from the inertial frame measured by star sensor B to the coordinate system measured by star sensor B is denoted as... The corresponding data sampling time is denoted as Data validity identifier is denoted as The attitude quaternion from the inertial frame measured by star sensor C to the coordinate system measured by star sensor C is denoted as... The corresponding data sampling time is denoted as Data validity identifier is denoted as A value of 1 indicates valid data, while a value of 0 or other values ​​indicates invalid data. Simultaneously, the angular velocity measurement value output by the gyroscope is read. Its expression is: ,in, This represents the angular velocity component along the X-axis of the satellite's coordinate system as measured by the gyroscope. Represents the angular velocity component about the Y-axis. This represents the angular velocity component around the Z-axis, with units of radians per second (rad / s). Additionally, the mounting matrix of the gyroscope relative to the satellite's body coordinate system... As a known parameter.

[0016] S21. Calculate the measurement time difference between the non-reference star sensor and the reference star sensor. For star sensor B, the formula for calculating the time difference with star sensor A is: ,in, This represents the time difference between the measurement time of star sensor B and the reference measurement time of star sensor A, which is either ahead (positive value) or behind (negative value), and is usually expressed in seconds. For star sensor C, the formula for calculating the time difference with star sensor A is: ,in, This indicates the time difference between the measurement time of star sensor C and the measurement reference time of star sensor A.

[0017] S22. Convert the angular velocity measured by the gyroscope to the satellite's coordinate system. The original angular velocity measurement value output by the gyroscope. It is relative to the gyroscope's own measurement coordinate system and requires the use of the gyroscope's mounting matrix. Transform it to a unified satellite coordinate system. The transformation formula is: ,in, This represents the equivalent angular velocity vector in the transformed coordinate system of the star's body; it is a three-dimensional column vector. This ensures that subsequent attitude change calculations are based on a unified reference coordinate system.

[0018] S23. Calculate the satellite attitude angle increment caused by the time difference. Assuming the satellite's angular velocity is constant within a short time difference, the magnitude of the attitude angle increment is equal to the magnitude of the angular velocity vector multiplied by the absolute value of the time difference. For star sensor B, the angle increment... The calculation formula is: ,in, , , Representing vectors respectively The first, second, and third components correspond to the angular velocity components along the X, Y, and Z axes of the star's coordinate system. Absolute value calculations ensure that the angular increment is non-negative, representing the magnitude of the rotation angle. For star sensor C, the angular increment... The calculation formula is: .

[0019] S24. The quaternion for attitude change corresponding to the construction time difference. Attitude change can be represented as motion rotating by an angle around an instantaneous rotation axis. The direction of this instantaneous rotation axis is determined by the direction of the normalized angular velocity vector. First, calculate the normalized angular velocity direction vector. : For star sensor B, its time compensation quaternion The calculation formula is: For the star sensor C, its time compensation quaternion The calculation formula is: The quaternion constructed here represents the rotation required to change the attitude from a "future" or "past" moment to the attitude at the "current" measurement reference moment. It's important to note that the sign of the time difference determines the direction of rotation; in practical calculations, the sign of the time difference is usually implied in the angle increment calculation or the order of quaternion application. One approach is to retain the sign of the time difference when calculating the angle increment, i.e. Thus, when the time difference is negative, the angular increment is negative, and the corresponding quaternion represents a reverse rotation.

[0020] S25. Apply the attitude change quaternion to the raw measurement data of the non-reference star sensor. Convert the calculated time compensation quaternion into the corresponding attitude rotation matrix. Then, apply the compensation quaternion to star sensor B. Convert to a rotation matrix, denoted as The compensation quaternion of the star sensor C. Convert to a rotation matrix, denoted as At the same time, the original measurement quaternions of star sensor B will be... Convert to an attitude matrix, denoted as The original measurement quaternion of star sensor C. Convert to an attitude matrix, denoted as The time compensation process, mathematically speaking, involves left-multiplying the original measurement matrix by the compensation rotation matrix, thereby "rotating" the original measurement value to the measurement reference time. The measurement matrix of star sensor B after compensation. for: The measurement matrix of the compensated star sensor C for: .

[0021] If the measurement data from star sensor B or star sensor C is invalid, that is or If the value is not equal to 1, no compensation is applied to the data, and the corresponding compensation matrix is ​​the identity matrix. .

[0022] S26. After the above compensation calculation, the measurement matrices of star sensor B and star sensor C are... and The attitude it represents is theoretically related to the measurement matrix of star sensor A. The attitudes they represent are at the same moment, that is, the sampling moment of the reference star sensor A. These matrices together constitute time-uniform measurement data, which can serve as input for the next step of updating the star sensor mounting matrix.

[0023] S3. Based on the time-unified measurement data, update the transformation relationship between the measurement coordinate systems used by each star sensor. The specific implementation process is as follows: S30. Acquire time-uniformed measurement data and related status information. Input data includes: the raw attitude quaternion of the measurement reference star sensor A. and the attitude matrix obtained by its transformation This data represents the time at which... The transformation relationship from the inertial frame to the measurement coordinate system of star sensor A. The input data also includes the attitude matrices of the non-reference star sensors B and C after time error compensation. and These represent the measurement data of star sensor B and star sensor C being unified to the measurement reference time, respectively. The transformation relationship from the inertial frame to their respective measurement coordinate systems needs to be determined. Simultaneously, the validity indicators of the measurement data from star sensor B and star sensor C need to be input. and And the current installation matrix of the reference star sensor A. .here, This may be the latest installation matrix after processing with the deviation correction function. If the deviation correction function is not enabled or has not been corrected, then... Equal to the initial installation matrix .

[0024] S31. Extract the reference measurement matrix required for calculation. Obtain the original attitude quaternion of the reference star sensor A. Convert to the corresponding attitude direction cosine matrix This matrix is ​​a key benchmark in subsequent calculations, describing the time at a common measurement reference point. The instantaneous attitude relationship between the inertial reference coordinate system and the measurement coordinate system of star sensor A. Matrix transpose matrix This will be used to convert the measurement data of other star sensors from an expression relative to an inertial frame to an expression relative to the measurement coordinate system of star sensor A.

[0025] S32. Determine the validity of the non-reference star sensor measurement data to decide whether to update it. Determine the validity flag of the star sensor B's measurement data. Is it equal to 1? Similarly, determine the validity of the measurement data from star sensor C. Is it equal to 1? Only star sensors with valid data will have their mounting matrix updated using their time-consistent measurement data. For star sensors with invalid data, their mounting matrix remains unchanged, and no update operation is performed.

[0026] S33. For non-reference star sensors that are valid for the data, calculate their updated installation matrix. The core idea of ​​the calculation is to utilize the data at exactly the same time ( By analyzing the observations of each star sensor in the same inertial space, the relative relationships between their measurement coordinate systems are derived. Then, combining this with the known mounting matrix of the reference star sensor A relative to the star's coordinate system, the mounting matrices of the non-reference star sensors relative to the star's coordinate system are deduced. For star sensor B, its updated mounting matrix... The calculation formula is: , where the symbol This represents the measurement matrix of star sensor B after time unification. This represents the transpose of the measurement matrix of the reference star sensor A. This represents the current installation matrix of the reference star sensor A relative to the satellite's body coordinate system. (Product) The physical meaning is to calculate the rotation matrix from the measurement coordinate system of star sensor A to the measurement coordinate system of star sensor B. Then, this relative rotation matrix is ​​compared with the installation matrix of star sensor A relative to the star's body coordinate system. Multiplying these matrices yields the new mounting matrix of star sensor B relative to the star's coordinate system. Similarly, for star sensor C, its updated installation matrix... The calculation formula is: , This represents the measurement matrix of star sensor C after time unification. The meanings of other symbols are the same as in the formula above.

[0027] S34. Perform the installation matrix update operation. For star sensors with valid measurement data, overwrite their original installation matrix values ​​with the calculated new installation matrix. Specifically: update the calculated... The value is assigned to the installation matrix storage variable of star sensor B. The update is now complete. , calculate The value is assigned to the installation matrix storage variable of star sensor C. The update is now complete. If the data from star sensor B is invalid, its update calculation is skipped, and the process is maintained. Unchanged. If the data from star sensor C is invalid, its update calculation is skipped, and the status remains unchanged. constant.

[0028] S35. To maintain consistency within the attitude measurement system, when the mounting matrix of the reference star sensor changes due to the deviation correction function, the gyroscope's mounting matrix also needs to be updated accordingly. Assume the rotation matrix used to correct the reference star sensor's mounting matrix, calculated from the deviation correction model, is... , , So, what is the installation matrix after the gyroscope update? The calculation formula is: ,in, This is the initial installation matrix for the gyroscope. If the deviation correction function is not enabled, then... , , The rotation matrices are all identity matrices, at this time This means that the gyroscope mounting matrix is ​​not updated.

[0029] The transformation relationship between the measurement coordinate systems used by each star sensor is mathematically represented by the direction cosine matrix or attitude transformation matrix between the measurement coordinate system of each star sensor and the satellite's body coordinate system, usually referred to as the installation matrix of that star sensor. This installation matrix is ​​a 3x3 orthogonal matrix that defines the rotational relationship from the satellite's body coordinate system to the star sensor's measurement coordinate system. Due to factors such as thermal deformation and stress release during satellite operation, the actual physical connection between the star sensor and the satellite platform may change slightly relative to the design state, leading to deviations in the initial ground-calibrated installation matrix. Therefore, it is necessary to periodically calculate and update these installation matrices using onboard software with real-time on-orbit measurement data to ensure that the transformation relationship reflects the current true geometric configuration. This is a prerequisite for achieving high-precision attitude determination and pointing correction.

[0030] S4. Based on the updated transformation relationship, the reference coordinate system used to determine the satellite attitude is unified with the measurement coordinate system used by the reference star sensor to establish a unified attitude reference coordinate system. The specific implementation process is as follows: S40. Confirm the identity of the reference star sensor and obtain its latest installation matrix. Based on the reference star sensor selection criteria, star sensor A is determined as the reference star sensor. Read the latest installation matrix of star sensor A from the onboard software's storage unit, denoted as... This matrix may be the initial installation matrix. It could also be the matrix after being updated by the on-orbit deviation correction function. It precisely describes the rotational relationship between the satellite's body coordinate system and the measurement coordinate system of star sensor A. Simultaneously, it reads the installation matrices of the non-reference star sensors B and C, which have been updated according to time-unified measurement data, and denoted as follows: and These updated transformation relationships form the basis for performing coordinate system-one operations.

[0031] S41. Acquire the valid measurement data of each star sensor at the current moment. In each attitude determination cycle, collect the raw attitude quaternions output by the reference star sensor A, star sensor B, and star sensor C in real time. The quaternion from the inertial frame to its own measurement coordinate system measured by star sensor A is denoted as... The quaternion from the inertial frame measured by star sensor B to its own measurement coordinate system is denoted as... The quaternion from the inertial frame measured by star sensor C to its own measurement coordinate system is denoted as... And record the validity identifier corresponding to each data point. , , and timestamp , , These data are the raw inputs for attitude determination.

[0032] S42. Design an attitude determination algorithm so that all internal calculations are based on a unified attitude reference coordinate system. Traditional satellite attitude determination algorithms, such as Kalman filters based on "star sensors + gyroscopes," typically define their state vectors (such as attitude quaternions or error quaternions) in the satellite's body coordinate system relative to an inertial coordinate system. To achieve a unified reference, the algorithm needs to be reformulated. The satellite attitude to be estimated by the attitude determination filter is explicitly defined as the satellite's body coordinate system relative to an inertial coordinate system, but this attitude is expressed through the measurement coordinate system of the reference star sensor A. In other words, the filter's state variables... Conceptually, this represents a rotation from the inertial frame (i) to the star-based frame (b), but all vector and matrix operations in this rotational relationship are implicitly performed within or ultimately transformed into the measurement coordinate system of star sensor A (denoted as s_A). An equivalent and more easily implemented engineering approach is to pre-transform all measurements from all sensors into the measurement coordinate system of star sensor A before inputting them to the filter.

[0033] S43. Perform data conversion to unify all input measurements to the measurement coordinate system of the reference star sensor A. For the measurements of the reference star sensor A itself... It is itself the attitude from the inertial frame to the s_A frame, and can be used directly. For the measurement values ​​of the non-reference star sensor B... This expresses the attitude of the inertial frame relative to the star sensor B's measurement coordinate system (s_B frame). It requires the installation matrix of star sensor B relative to the star's home frame. Installation matrix of star sensor A relative to the star system Convert it to the s_A series. The specific conversion process is as follows: First, convert it to the s_A series. Convert to direction cosine matrix Then, the rotation matrix from the s_B system to the star-base system b is calculated, which is determined by the installation matrix. The transpose of is given, that is Next, the rotation matrix from the s_A system to the s_A system is calculated, i.e., the installation matrix. Therefore, the measurements of star sensor B are transformed into a composite rotation matrix in the s_A system. for: , This represents the mounting matrix from the star body coordinate system (b) to the measurement coordinate system (s_A) of the measurement reference star sensor A. The transpose of the matrix represents the mounting matrix from the star body coordinate system (b) to the star sensor B measurement coordinate system (s_B). This indicates a rotation from the s_B system to the b system. This represents the measurement matrix from the inertial coordinate system (i) to the s_B system. The product result... This is the equivalent attitude matrix from the inertial coordinate system (i) to the s_A system, which unifies the measurement reference of star sensor B. Similarly, for star sensor C, the matrix transforming its measurements to the s_A system... for: , This represents the mounting matrix from the star body coordinate system (b) to the star sensor C measurement coordinate system (s_C). This represents the measurement matrix from the inertial coordinate system (i) to the s_C system.

[0034] S44. Convert the gyroscope's angular velocity measurements to a unified attitude reference coordinate system. The original angular velocity output by the gyroscope. This is relative to its own installation coordinate system. First, the installation matrix of the gyroscope is used. (or the updated version) Transform it to the satellite's body coordinate system to obtain Then, the angular velocity in the star's coordinate system needs to be calculated. The coordinates are transformed to the measurement coordinate system s_A of the reference star sensor A. This transformation is determined by the mounting matrix of the reference star sensor A. This is achieved by transposing, because This means rotating a vector from frame b to frame s_A, and its transpose rotates the vector from frame s_A back to frame b. Conversely, for vectors like angular velocities, the transformation from frame b to frame s_A should use... Since the matrix is ​​orthogonal, its inverse is equal to itself, so the transformation matrix is... Therefore, the angular velocity in the unified reference s_A system for: , This represents the satellite angular velocity vector expressed in a unified attitude reference coordinate system (i.e., the s_A system).

[0035] S45. Perform attitude determination calculations in a unified attitude reference coordinate system. Input the sensor data, after the above transformation and expressed in the s_A coordinate system, into the attitude determination filter. This data includes: direct measurements from star sensor A. (or the corresponding quaternion), the converted measurement from star sensor B. Measurement after conversion from star sensor C and angular velocity under a unified standard Attitude determination filters (such as extended Kalman filters) will use this data to estimate and output the satellite's attitude in its body coordinate system relative to its inertial coordinate system. The key point is that all state variables, measurement models, and dynamic models within the filter are derived and calculated based on the s_A frame; therefore, the final output is the optimal estimated attitude quaternion. Although it describes the relationship between the stellar system and the inertial frame, it is defined and solved in the mathematical space of the s_A frame. This means that the coordinate reference for the entire attitude determination process, from input to output, is always the s_A frame, thus establishing a unified attitude reference coordinate system.

[0036] S46. Attitude determination filter output attitude This can be directly used in the satellite attitude control system. Since the control law is usually designed in the satellite's own coordinate system, this output can be used without additional transformation. For other modules that need to use the coordinate system information of the measurement reference star sensor A, since they coincide with the unified reference coordinate system, the relevant data can be directly referenced. Through the continuous cyclic execution of the above steps, the satellite attitude determination system firmly establishes and maintains a unified attitude reference coordinate system based on the measurement coordinate system of the measurement reference star sensor, ensuring the consistency and accuracy of the attitude information of the entire system.

[0037] The unified attitude reference coordinate system refers to a unique and fixed three-dimensional Cartesian coordinate system used by all algorithm modules in the entire satellite attitude determination and control system. It serves as a common reference for describing satellite attitude, performing sensor data fusion, and generating control commands. In this invention, this coordinate system is set to completely coincide with the measurement coordinate system used by the reference star sensor. By transforming and unifying the angular velocity integral of the gyroscope, the measured values ​​of other star sensors, and the state of the attitude determination filter to this coordinate system, it is ensured that all data sources and state variables in the attitude determination process have a consistent geometric definition, fundamentally avoiding errors introduced by multi-coordinate system transformations and improving the overall accuracy and internal consistency of attitude determination.

[0038] S5. Based on a unified attitude reference coordinate system, the satellite is controlled to perform imaging tasks in different Beta angles, latitudes, and longitudes to acquire actual imaging data from the satellite's optical cameras. The specific implementation process is as follows: S50. Receive and parse the imaging mission command injected from the ground. The ground control station sends the imaging mission command to the satellite via the space-to-ground communication link. This command contains the target parameters for this imaging mission, specifically including: the geographical longitude of the center point of the target area. Geographical latitude And the solar beta angle expected during imaging. The onboard mission management unit receives and parses these parameters, using them as input for this attitude maneuver and imaging control. The imaging mission command also specifies the operating mode of the optical camera, such as integration time and gain settings.

[0039] S51. To ensure the optical camera's line of sight is precisely pointed to the target ground point, the required attitude of the satellite's body coordinate system must be calculated. This target attitude calculation is performed within a unified attitude reference coordinate system. The calculation requires several input data points: the current orbital position and velocity vector of the satellite, provided by the onboard navigation system; and the target's geographical latitude and longitude. The nominal installation direction of the optical camera's line of sight in the satellite's body coordinate system is a fixed unit vector, denoted as . ; and the latest installation matrix of the measurement reference star sensor relative to the star body coordinate system. The core of target attitude calculation is solving for a target attitude quaternion. This ensures that when the satellite is in this attitude, the optical camera's line-of-sight, expressed in a unified attitude reference coordinate system, points precisely to the target ground point. The calculation process involves converting the target ground point from geographic coordinates to a direction vector in a geocentric inertial frame, and then, through a series of coordinate system transformations, solving for the satellite's attitude relative to the inertial coordinate system, which satisfies the pointing requirement. Since all calculations are performed using the same coordinate system as the reference star sensor, calculation errors caused by different references are avoided.

[0040] S52, The attitude control software calculates the target attitude quaternion. The attitude error is generated by comparing the current estimated satellite attitude quaternion fed back by the attitude determination system. Based on this error, the required control torque is calculated for the attitude control law (e.g., proportional-derivative control law). The formula for calculating the control torque can be expressed as: ,in, It is the calculated three-axis control torque command vector. and It is the matrix of proportional and differential coefficients of the controller. It is the vector part extracted from the attitude error quaternion, representing the angle and axis of the attitude error. This is the satellite's angular velocity error vector. All these calculations are performed in a unified attitude reference coordinate system, ensuring that the reference for control commands is completely consistent with the reference for attitude sensing. The calculated control torque commands are sent to the satellite's actuators, typically reaction wheels or control torque gyroscopes. The actuators generate corresponding control torques, driving the satellite to rotate around its center of mass, gradually eliminating attitude errors and ultimately stabilizing the satellite in the target attitude. superior.

[0041] S53. When the error between the actual satellite attitude and the target attitude is less than a preset threshold, and the satellite angular velocity is also lower than the stability threshold required for imaging, the attitude control system issues an "attitude stable" flag. Upon receiving this flag, the onboard mission management unit sends an imaging trigger command to the optical camera payload. The optical camera opens its shutter according to preset operating parameters and begins Earth exposure imaging. During the exposure, the attitude control system continues to operate, maintaining the high stability of the satellite attitude and ensuring that the imaging quality is not affected by satellite jitter.

[0042] S54. At the precise moment of exposure and imaging by the optical camera, the onboard data recording system needs to simultaneously record a set of auxiliary data that strictly matches the imaging data. This auxiliary data is crucial for subsequent image positioning and pointing deviation analysis, and mainly includes: the satellite orbital position vector at the center of the imaging exposure. and velocity vector The precise attitude quaternions of the satellite in a unified attitude reference coordinate system at the center of the imaging exposure are provided by the onboard GNSS receiver or orbit determination filter. The attitude determination system provides the imaging mission parameters, including target latitude and longitude. and the corresponding Beta angle These auxiliary data, along with the image data generated by the optical camera, will be packaged together to form a complete data frame.

[0043] After the S55 optical camera completes exposure, it converts the light signal into a digital electrical signal, generating raw image data. After preliminary processing (such as dark current correction and flat-field correction), the image data is correlated and encapsulated with recorded auxiliary data packets to form the final actual imaging data product. This product is temporarily stored on-board. When the satellite flies over a ground receiving station, the onboard communication system transmits the stored actual imaging data product downlink to the ground station according to plan. The ground station receives and demodulates the data to obtain usable satellite images and precisely matched auxiliary information such as orbit, attitude, and time.

[0044] S6. Compare the actual imaging data with the map loaded in the simulation model to determine the typical feature regions of the images, and determine the error angle between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera. Specifically, this includes: S60. Based on satellite orbit data, satellite attitude data, latitude and longitude of the imaging target point, and the field of view parameters of the optical camera, a simulation model is established and a map is loaded into the simulation model. The simulation model is equipped with a sensor whose field of view parameters are consistent with those of the optical camera. Specifically, ground data processing personnel extract the precise time stamp of the imaging exposure center time from the transmitted actual imaging data packets. and the corresponding satellite orbit data at that moment. and satellite attitude quaternions Create a new scene in the satellite simulation software. Insert a satellite object into the scene and use the extracted orbital data. and precise time Define the satellite's state at the time of imaging. Associate a sensor object with this satellite object. When configuring the sensor object's properties, on its "Define" page, set the "Sensor Type" to "Rectangular Field of View". On the "Basic" page, based on the precise calibration values ​​provided by the satellite platform, set the optical camera's field of view parameters: set the "Horizontal Half Angle" to half of the optical camera's total field of view angle in the vertical direction, denoted as... The "vertical half-angle" is set to half of the total field of view of the optical camera along the track direction, denoted as... These parameters ensure that the geometry of the virtual sensor perfectly matches that of a real optical camera. In the sensor's "Pointing" properties page, select the "From Vector" pointing method and enter the initial pointing vector. The initial pointing is typically set to align the sensor's line of sight with the latitude and longitude of the imaging target point. The corresponding geocentric direction, and using satellite attitude data at the time of imaging. The coordinates are calculated through a transformation. After completing the sensor configuration, a high-resolution, high-precision digital map or satellite base map with geographic coordinates is loaded into the scene's map layer as a comparison benchmark.

[0045] Specifically, the satellite simulation software can be the CSTK software platform, VVP space system analysis toolbox, OrbitLab software, or SpaceSim software, or other satellite simulation software can be selected according to the actual situation.

[0046] The latitude and longitude of the imaging target point refers to the geographical coordinates of the point on the Earth's surface theoretically pointed to by the center of the satellite's line of sight at the moment of imaging exposure, represented by both longitude and latitude values. This parameter is usually specified by the imaging mission instructions and serves as the initial input for defining the sensor's target orientation when building the simulation model, used to determine the approximate location of the simulated strip on the map. The field-of-view parameters of the optical camera are a set of technical parameters describing the geometry of the camera's imaging range, typically including instantaneous field of view angle, total field of view angle, vertical-orbit direction field of view angle, along-orbit direction field of view angle, number of pixels, and pixel size. These parameters determine the size and shape of the ground area that the optical camera can cover in a single imaging session. They are attributes that must be precisely set when configuring the virtual sensor model in the satellite simulation software to ensure that the geometric dimensions of the simulated imaging strip are consistent with the actual image. The simulation model refers to a high-fidelity virtual imaging scene constructed in the satellite simulation software based on the satellite's actual orbital data, attitude data, target point latitude and longitude, and the specific field-of-view parameters of the optical camera at the moment of imaging. This model is configured with a virtual sensor whose field-of-view parameters are completely identical to those of the real camera and uses a high-precision digital map as a reference. By running this model, simulated strips of Earth imaging by a satellite under ideal, unbiased conditions can be generated, thus providing an accurate theoretical reference for geometric comparison with actual imaging results.

[0047] S61. Select typical areas as the left and right boundary matching areas and the boundary matching area at the start of imaging, respectively. Specifically, ground data processing personnel open the actually acquired satellite image. In the image display software, carefully browse the entire image to find ground targets that meet the characteristics of the "typical area". The selection process follows the principles that: the features are clearly distinguishable in the image and have distinct outlines; the features are also clearly present in the high-precision map data and are easy to locate; the feature area is stable in position within the imaging strip and is not easily obscured by clouds or shadows. Near the left edge of the image, select a typical area and mark it as the "left boundary matching area". Near the right edge of the image, select a typical area and mark it as the "right boundary matching area". At the top edge of the image, corresponding to the start of the imaging scan, select a typical area and mark it as the "imaging start time boundary matching area". Record the approximate range or feature point coordinates of these areas in the actual image pixel coordinate system.

[0048] The typical region refers to one or more local image blocks with significant, stable, and easily identifiable features, selected manually or automatically from the actual image data acquired by the satellite. These features include, but are not limited to, coastline corners, lake outlines, river confluences, specially shaped man-made structures, and ridgelines. The purpose of selecting the typical region is to provide a clear and unambiguous reference for subsequent image comparison. The left and right boundary matching regions refer to the typical regions selected from the actual imaging data, located near the left and right edges of the image. These two regions are used to determine the coverage of the imaging strip in the vertical direction (perpendicular to the satellite's flight direction). By matching the left and right boundaries of the simulated strip with these two actual regions, pointing deviation around the satellite's pitch axis can be corrected. The imaging start-time boundary matching region refers to the typical region selected from the actual imaging data, located near the upper edge of the image (i.e., the ground area corresponding to the start of the imaging scan). This region is used to determine the starting position of the imaging strip in the orbital direction (parallel to the satellite's flight direction). By matching the starting boundary of the simulated strip with this actual region, pointing deviation around the satellite's roll axis can be corrected.

[0049] S62. By adjusting the Euler angles of the sensor in the simulation model, the left and right boundary coverage areas of the simulated imaging strip on the map are made consistent with the left and right boundary matching areas in the actual imaging data, thus obtaining the first Euler angle. Specifically, in the satellite simulation software, the display of the sensor's ground coverage area is enabled, and a rectangular strip enclosed by four boundaries will appear on the map. The operator simultaneously observes the simulated strip in the satellite simulation software and the actual satellite image. The focus is on comparing the ground areas covered by the left and right boundary lines of the strip with the ground positions corresponding to the selected "left boundary matching area" and "right boundary matching area" in the actual image. Enter the "offset angle" or "Euler angle" setting interface in the sensor's "pointing" attribute. Select to define the additional pointing deviation of the sensor according to a specific rotation sequence (e.g., ZYX, i.e., yaw-pitch-roll). By interactively adjusting the corresponding Euler angle value around the orbital axis (usually the roll axis, denoted as EulerA), observe the left and right movement and rotation of the simulated strip on the map. The goal of the adjustment is to ensure that the left boundary line of the simulated strip crosses or closely follows the map position corresponding to the left boundary matching area of ​​the actual image, and simultaneously ensures that the right boundary line of the simulated strip crosses or closely follows the map position corresponding to the right boundary matching area of ​​the actual image. Adjustment is stopped when the left and right boundaries achieve optimal matching visually or through coordinate measurements. At this point, the Euler angle component in the sensor's "pointing" attribute that primarily controls the vertical track direction offset, such as the "roll" angle, is recorded and designated as the first Euler angle, denoted as . In some implementations, It may be a vector containing multiple components, but its core is to reflect the deviation of the vertical direction.

[0050] The simulated imaging strip's left and right boundary coverage areas refer to the projected coverage areas on a map of the two boundary lines located on the left and right sides of the strip, within the virtual sensor coverage area calculated and rendered by the satellite simulation software based on the input satellite orbit, attitude, and sensor parameters. Adjusting the sensor's pointing angle can change the positions of these two boundary lines on the map.

[0051] S63. By adjusting the Euler angles of the sensor in the simulation model, the area covered by the starting time boundary of the simulated imaging strip on the map is made consistent with the area matching the starting time boundary of the actual imaging data, thus obtaining the second Euler angle. Specifically, the Euler angles used for matching the left and right boundaries, which were adjusted in the previous step, remain unchanged. Now, turn your attention to the upper boundary of the simulation strip, i.e., the starting time boundary. Compare the ground area covered by this boundary line with the ground position corresponding to the selected "starting time boundary matching area" in the actual image. Return to the Euler angle setting interface of the sensor's "pointing" attribute and adjust the corresponding Euler angle value around the vertical axis (usually the pitch axis, denoted as EulerB). Observe the vertical movement of the simulation strip on the map. The goal of the adjustment is to make the upper boundary line of the simulation strip (starting time boundary) pass through or closely adhere to the map position corresponding to the starting time boundary matching area of ​​the actual image. Stop adjusting when the starting time boundary reaches the best match. At this point, record the Euler angle components in the sensor's "pointing" attribute that are primarily responsible for controlling the offset along the track direction, such as the "pitch" angle. Record this as the second Euler angle, denoted as... .

[0052] The simulated imaging strip's starting moment boundary coverage area refers to the projected coverage area on the map of the boundary line corresponding to the start of the imaging scan within the virtual sensor's Earth coverage area in satellite simulation software. Adjusting the sensor's pointing angle will also change the position of this starting boundary line on the map.

[0053] S64. The first Euler angle and the second Euler angle are determined as error angles characterizing the deviation between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera. Specifically, the first Euler angle obtained in the above adjustment process is... Second Euler angle These two angles have been formally defined as the error angles characterizing the deviation between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera in this imaging mission. This directly reflects the minute angular difference in the roll and pitch directions between the satellite attitude determined by the measurement reference star sensor (corresponding to a unified attitude reference coordinate system) and the actual Earth pointing direction achieved by the optical camera (corresponding to the line-of-sight coordinate system) at the instant of satellite imaging. Typically, the error angle around the yaw axis... The impact on strip matching is minimal and can be set to 0 or calibrated using other methods. Thus, through precise comparison between the simulation model and actual images, the key pointing deviation observations were successfully extracted from single-shot imaging data. This provides data samples for subsequent mathematical modeling.

[0054] S7. Based on the error angles obtained from different Beta angles, latitudes, and longitudes, establish a mathematical model for the error angles, specifically including: S70. Based on the satellite's orbit, determine the range of variation of the Beta angle and the coverage area of ​​the nadir latitude within the mission cycle. Specifically, using a satellite orbital dynamics model or orbital prediction tools, calculate the curve of the solar Beta angle changing over time over a complete mission cycle (usually one year). The solar Beta angle is defined as the angle between the solar direction vector and the satellite's orbital plane. Extract the maximum value of the Beta angle from the curve. and minimum value Thus, its range of variation is determined. Simultaneously, based on parameters such as the satellite's orbital inclination, the highest latitude in the Northern Hemisphere that the nadir point can reach within the same mission cycle is calculated. and the lowest latitude in the Southern Hemisphere (Or, depending on the task area), determine the latitude coverage area. The output of this step is two intervals, which are used to guide the selection of subsequent sampling points.

[0055] S71. Within the range of Beta angle variation, select multiple discrete Beta angle values ​​at a first preset interval; within the latitude coverage range, select multiple discrete latitude values ​​at a second preset interval; and combine each selected Beta angle value and each latitude value with the longitude of the nadir point corresponding to the imaging time to form multiple sets of different imaging experimental conditions, specifically: Within the range of Beta angle variation Within, at the first preset interval Perform uniform discrete sampling. First preset interval. This is a constant value determined based on the modeling accuracy requirements and imaging opportunities, for example, 5 degrees. The sampled discrete Beta angle value sequence is as follows: ,in The function represents rounding the calculation result. Let be the number of sampling points, such that No more than .symbol This represents the i-th discrete Beta angle sample value. Within the latitude coverage area... Within, at the second preset interval Perform uniform discrete sampling. Second preset interval. This is another constant value chosen based on the latitude zone characteristics, for example, 2 degrees. The resulting discrete latitude value sequence is: ,in, Let j represent the j-th discrete latitude sample value. This represents the number of sampling points per latitude.

[0056] Each discrete Beta angle value With each discrete latitude value Combine them. For each pair... Combined with the actual latitude the satellite flew over at the time of imaging... longitude of the sub-satellite point at that time Together they constitute a set of imaging test conditions longitude of the sub-satellite point It is determined by orbital prediction at specific times and latitudes. From this, a [database / structure] will be generated. Different imaging experimental conditions were used to form a sampling grid covering the two-dimensional parameter space of Beta angle-latitude.

[0057] S72. Under each set of imaging test conditions, the satellite is controlled to perform imaging based on a unified attitude reference coordinate system to obtain the corresponding actual imaging data, and to obtain multiple sets of error angles including the first Euler angle and the second Euler angle. Specifically, for each set of imaging test conditions generated in the previous step... The ground control center will send imaging mission commands to the satellite at an opportune time. Based on the established unified attitude reference coordinate system, the satellite will control its own attitude, ensuring the optical camera points towards the latitude and longitude coordinates. The target area is imaged, and actual imaging data is acquired. During this stage, the "deviation correction function" of the onboard software is turned off, meaning no correction is applied to obtain the original pointing deviation data. For each imaging session, the actual imaging data is processed at the ground station using the same method as steps S60 to S64: a simulation model is established, a typical area is selected, and the sensor Euler angles are adjusted to match the left and right boundaries and the starting time boundary of the actual image, thereby calculating the first Euler angle corresponding to that imaging session. Second Euler angle By traversing all M sets of imaging experiments, M sets of error angle observation data are finally obtained, with each set of data including imaging conditions. and the corresponding error angle observation value .

[0058] S73. Based on multiple sets of error angles, construct the first mathematical relationship between the first Euler angle and the Beta angle, latitude, and longitude during imaging, and the second mathematical relationship between the second Euler angle and the Beta angle, latitude, and longitude during imaging. Specifically: based on the analysis of the physical mechanisms of satellite thermal deformation and structural deformation, assume the first Euler angle... (Primarily reflecting pitch deviation) has a linear relationship with the Beta angle, longitude, and latitude during imaging. The first mathematical expression is: ,in, It is a constant deviation term. It is the error angle Relative to Beta angle The linear coefficient reflects the sensitivity of the deviation to changes in the angle of sunlight. It is the error angle Relative to longitude The linear coefficients. It is the error angle Relative to latitude The linear coefficients. , , These are all parameters from the imaging test conditions.

[0059] Similarly, assuming the second Euler angle (Mainly reflecting roll deviation) and imaging conditions also have a similar linear relationship, and a second mathematical relationship can be constructed as follows: ,in, It is a constant deviation term. It is the error angle Relative to Beta angle The linear coefficients. It is the error angle Relative to longitude The linear coefficients. It is the error angle Relative to latitude The linear coefficients.

[0060] S74. The least squares method is used to fit the unknown coefficients in the first and second mathematical relationships. The first and second mathematical relationships with the fitted coefficients are used together as the mathematical model for the error angle. Specifically, the M sets of observation data are obtained. Substituting the first mathematical relation, we form an overdetermined system of equations. We then use the least squares method to solve for the coefficients. This makes the predicted value calculated from the relational expression... Compared with measured values The sum of squared residuals is minimized. The fitting process can be accomplished by constructing a design matrix and normal equations. Assuming there are N valid observation samples (N≤M), the observation data is organized into matrix form. Define the parameter vector. Observation vector Design matrix The m-th line The least squares solution is: , This represents the vector of coefficients to be determined in the first mathematical relation. (Symbols) This represents the vector consisting of all N first Euler angle observations. (Symbol) This represents an N-row, 4-column design matrix consisting of the corresponding imaging condition parameters. (Superscript) Indicates matrix transpose, superscript " "" indicates the inversion of a matrix.

[0061] Similarly, using observational data The coefficients of the second mathematical relationship are fitted using the least squares method. For the error angle around the yaw axis It is usually assumed that it is unrelated to these conditions or has a negligible impact, and its coefficients in the mathematical model can be directly set to zero, that is: The coefficients obtained from the fitting , , Substituting the first and second mathematical relations (and the yaw axis relation) into the equations, the first and second mathematical relations, now complete with coefficient fitting, together constitute the mathematical model of the error angle. This model expresses the deviation angle between the unified attitude reference coordinate system and the optical camera's line-of-sight coordinate system in functional form. The solar beta angle of the satellite at the time of imaging Target longitude Target latitude The quantitative dependence between them can be used to predict pointing deviation under any imaging conditions.

[0062] The nadir latitude refers to the geographical latitude value corresponding to the vertical projection of the satellite's center of mass onto the Earth's surface at a given moment, i.e., the nadir point. In sun-synchronous orbit missions, the latitude of the satellite's nadir point changes periodically with orbital precession and Earth's rotation, forming a coverage zone. Determining the coverage range of the nadir latitude is crucial for planning imaging experiments to ensure that the sampling points represent the entire working latitude range of the satellite, thus giving the established error angle model sufficient spatial representativeness. Imaging experiment conditions refer to a pre-designed set of imaging mission parameters for systematically calibrating pointing deviations. Each set of imaging experiment conditions consists of three key parameters: a specific solar beta angle value, a specific target latitude value, and the longitude value of the nadir point corresponding to the imaging time. By designing multiple sets of different imaging experiment conditions and actually imaging under these conditions, discrete sample data of pointing deviations under various orbital and illumination geometries can be obtained, providing a sufficient data foundation for fitting mathematical models.

[0063] S8. Based on the mathematical model of the error angle, a pointing deviation correction coefficient is generated, and the installation matrix of each star sensor in the on-board software is updated by on-orbit annotation to correct the satellite's imaging pointing deviation. The specific implementation process is as follows: S80. At the ground data processing center, after establishing the mathematical model of the error angle and fitting the coefficients, a set of definite coefficient values ​​is obtained: , , These coefficients are stored in floating-point format. To ensure the consistency of the values ​​calculated by the onboard software, the ground software needs to encode and encapsulate these coefficients according to the data format specified by the onboard software (such as single-precision floating-point or fixed-point format) in preparation for generating the uploading data package.

[0064] S81. Ground control station operators assemble a dedicated "deviation correction coefficient update" command packet according to the satellite remote control command protocol. This data packet contains a command identifier header, data length information, and the generated, formatted correction coefficient data body. A cyclic redundancy check (CRC) code or other error control code is appended to the end of the data packet for on-board verification of data transmission integrity. During the predetermined satellite transit arc, the ground control antenna tracks the satellite and establishes uplink and downlink links. The ground remote control terminal modulates the assembled data packet onto a radio frequency carrier and transmits it to the satellite.

[0065] The remote demodulation module in the S82 satellite's integrated electronic unit receives radio frequency signals from the ground and demodulates the data packets. The onboard software performs frame synchronization, address identification, and cyclic redundancy check on the received data packets. If the check passes, confirming the instruction as a valid "deviation correction coefficient update" instruction, it extracts the encoded correction coefficient data body from the data packet. The onboard software decodes these coefficient values, converts them into floating-point format used for onboard calculations, and writes them into a dedicated storage area in the onboard non-volatile memory for "deviation correction model parameters," overwriting the original parameter values ​​in that area (initial values ​​are usually all zero). After successful storage, the onboard software sends a "parameter update successful" confirmation signal to the ground via the telemetry channel.

[0066] S83. In each subsequent attitude determination and control cycle, the onboard software reads the currently stored correction coefficients from the non-volatile memory when executing the attitude sensor data processing procedure. The software first determines whether the "deviation correction function" is set to "on". This function's on / off state can be set by a separate remote control command from the ground. If the function is "on", the software will perform the following calculations: Based on the current imaging mission plan or real-time attitude control status, obtain the current solar Beta angle. Longitude of the target area Latitude of the target area Using the correction factor read and the obtained The value is calculated in real time according to the mathematical model to correct the deviation angles of the three axes. : in, It is the calculated real-time correction angle. It refers to the solar beta angle, longitude, and latitude parameters at the time of imaging or the current time. These are the correction coefficients that have already been stored and added.

[0067] S84. Update the mounting matrix of the reference star sensor using the correction angle. Calculate the correction angle. Then, the onboard software constructs rotation matrices around the X, Y, and Z axes based on these angles. , , : Then, these rotation matrices are applied to the initial mounting matrix of the measurement reference star sensor A. The corrected reference star-sensor mounting matrix is ​​obtained. : , This represents the initial calibration and installation matrix of the measurement reference star sensor A. This represents the new mounting matrix used for attitude determination at the current moment, after applying bias corrections. This updated... It will be used immediately in subsequent processes such as unifying star-sensor data time, updating the installation matrix, and establishing the attitude reference coordinate system.

[0068] S85. To maintain the consistency of the coordinate system definition within the attitude measurement system, the gyroscope's mounting matrix needs to be updated synchronously while updating the measurement reference star sensor mounting matrix. The update formula is as follows: in, This is the initial mounting matrix for the gyroscope. This is the updated gyroscope mounting matrix. This ensures that the angular velocities measured by the gyroscope, when transformed to the star body coordinate system, are based on the updated star sensor reference.

[0069] Through the cyclical execution of S80 to S85 described above, once the model coefficients are uploaded to the ground and the correction function is activated, the onboard software can automatically calculate the precise pointing deviation correction amount based on real-time orbital and illumination geometry conditions in each imaging mission. This correction amount is then integrated into the attitude determination link at its source by dynamically updating the installation matrix of the measurement reference star sensor. In this way, the "unified attitude reference coordinate system" upon which the satellite attitude control system relies inherently includes compensation for optical camera line-of-sight deviations. This allows attitude control commands generated based on this reference coordinate system to directly drive the satellite, enabling the optical camera to achieve precise pointing without deviation, ultimately achieving the goal of correcting satellite imaging pointing deviations. The entire correction process is closed-loop, automatic, and real-time, significantly improving the geolocation accuracy of on-orbit imaging.

[0070] The technical solution of the present invention will be further described through another embodiment, including the following steps: S101. The satellite is equipped with three star sensors. Star sensor A is selected as the reference measurement star sensor according to the selection principle. The other two star sensors are designated as star sensor B and star sensor C, respectively. Specifically: 1) The selection principles for the reference measurement star sensor include: the mounting bracket of star sensor A is directly connected to the camera load-bearing frame; the installation distance between star sensor A and the optical camera load is the shortest; structural mechanics analysis shows that the structural deformation between the line-of-sight coordinate system of star sensor A and the optical camera load is minimized throughout the entire life cycle; and optical environment simulation analysis shows that the cumulative time of star sensor A being affected by terrestrial light, sunlight, and moonlight throughout the entire life cycle is minimized. Based on the overall satellite structure layout design and precision measurement results, the initial installation matrices of star sensor A, star sensor B, and star sensor C relative to the satellite body coordinate system are calculated and denoted as follows: , , It should be noted that in the satellite design, the satellite's body coordinate system coincides with the optical camera's line-of-sight coordinate system. The initial installation matrix of the gyroscope relative to the satellite's body coordinate system is denoted as... The angular velocity measurement value output by the gyroscope in real time is denoted as... The unit is radians per second. The attitude quaternions measured by star sensors A, B, and C are denoted as follows: , , The corresponding time information is , , Validity identifier is , , When the validity flag is equal to 1, it indicates that the star sensor measurement data is valid; other values ​​indicate that the data is invalid.

[0071] 2) Determine the validity of the reference star sensor data. The onboard software first determines whether the reference star sensor measurement data is valid at the current calculation time. The determination criterion is to check... Is it equal to 1? If it is equal to 1, the first judgment result is yes; otherwise, the first judgment result is no.

[0072] 3) If the first judgment result is negative, that is, the data of the measurement reference star sensor A is invalid at the current moment, then the installation matrix of star sensor B, star sensor C and gyroscope will not be updated during this calculation cycle, and the process will jump to the end of the cycle.

[0073] 4) If the first judgment result is yes, that is, the data from the reference star sensor A is valid, the process continues to determine whether the deviation correction function is enabled, and obtains the second judgment result. The on / off state of this function is controlled by ground commands.

[0074] 5) If the second judgment result is yes, that is, the deviation correction function has been enabled, the on-board software performs the following operations: read the solar Beta angle at the current moment. and the longitude of the target area for the current imaging mission. and latitude Subsequently, the deviation correction angles for the three axes are calculated based on the preset deviation angle correction mathematical model. , , They initially all default to zero and can be updated in orbit via ground-based injection. Then, the mounting matrix of the measurement reference star sensor A is updated using the calculated correction angle. The corrected mounting matrix... The calculation formula is: Subsequently, the reference star-sensor matrix currently used for attitude determination will be installed. Set as .

[0075] 6) If the second judgment result is negative, that is, the deviation correction function is turned off, then set the deviation correction angle. , , At this point, the installation matrix of the reference star sensor A is not updated with deviation correction; its initial installation matrix is ​​directly used in subsequent calculations. .

[0076] 7) Determine and process the validity of non-reference star sensor data. The onboard software continues to determine the validity of the measurement data from star sensors B and C. If the measurement data from both star sensors B and C are invalid, i.e. and If none of them are equal to 1, then the installation matrices of star sensor B and star sensor C will not be updated during this calculation period, and the process will jump to the next step.

[0077] 8) Time error compensation for valid non-reference star sensor data. If the measurement data from both star sensors B and C are valid, then the measurement time error between them and the reference star sensor A needs to be compensated using gyroscope measurement data to obtain time-consistent measurement data. For star sensor B, calculate the time difference: Convert the gyroscope angular velocity to the star's coordinate system: Calculate the magnitude of the attitude angle increment corresponding to the time difference. And calculate the normalized rotation axis vector. Construction time-compensated quaternions Quaternions Convert to rotation matrix At the same time, the original measurement quaternions of star sensor B will be... Convert to attitude matrix Then, the measurement matrix of star sensor B after time compensation is: For star sensor C, the same method as for star sensor B is used to calculate and obtain the time-compensated measurement matrix. If the measurement data of either star sensor B or star sensor C is invalid, then its corresponding time compensation matrix is ​​set to an identity matrix, i.e. or This means that no time compensation will be made.

[0078] 9) Update the mounting matrix of non-reference star sensors. Using the time-unified measurement data, update the mounting matrix of non-reference star sensors to align them with the reference defined by the reference star sensor A. First, adjust the measurement quaternions of the reference star sensor A. Convert to attitude matrix Then, calculate the new mounting matrix of star sensor B after unifying it to the reference star sensor A. : Similarly, calculate the new mounting matrix of star sensor C after unifying it to the reference star sensor A. : .

[0079] 10) Perform the assignment and update of the installation matrix. Assign the calculated new installation matrix to the corresponding star sensor installation matrix storage variable to complete the update operation: .

[0080] 11) Update the gyroscope mounting matrix. To maintain consistency within the attitude measurement system, the gyroscope mounting matrix also needs to be updated synchronously based on the current deviation correction angle. Updated gyroscope mounting matrix. The calculation formula is: .

[0081] S102. By comparing the actual acquired satellite imaging data with the high-precision map in the simulation model, typical feature areas of the image are determined to identify the error angle between the satellite measurement coordinate system and the camera line-of-sight coordinate system, including: 1) Ground processing personnel extract auxiliary data that strictly matches a particular imaging session from the transmitted satellite data packets, including: the center time of the imaging exposure. The position vector of the satellite in the geocentric inertial frame at that moment and velocity vector And the satellite's attitude quaternion relative to the inertial frame at that moment. Simultaneously, the latitude and longitude of the target point specified in this imaging mission were obtained. And the precise field-of-view parameters of the optical camera (such as vertical and along-track field of view angles). In the satellite simulation software, create a new scene and set the correct time system. Insert a satellite object and use the extracted orbit vectors. and precise time The satellite's state at the moment of imaging is defined by its "orbit vector." A sensor object is added to this satellite object. In the sensor properties, the sensor type is defined as "rectangular field of view," and its "horizontal half-angle" and "vertical half-angle" are accurately set according to the optical camera's field of view parameters to ensure that the geometry of the virtual sensor is completely consistent with the real camera. In the sensor's "pointing" property, its basic pointing is configured so that its initial state is aligned with the target point's latitude and longitude. Finally, high-resolution, high-precision georeferenced digital maps or satellite imagery base maps covering the imaging area are loaded into the scene as a benchmark for subsequent image comparison.

[0082] 2) Select typical feature regions for matching from actual satellite imagery. Operators open the actual optical remote sensing image acquired in this mission using specialized image processing software. Carefully review the entire image, identifying and delineating multiple candidate typical feature regions based on the principles of "clear, easily identifiable, easily distinguishable, and highly visible in the loaded high-precision map." These regions should be distributed in different locations, especially including features near the left, right, and top edges of the image (corresponding to the imaging start time). Finally, determine a pair of feature regions for left and right boundary matching, and one feature region for boundary matching at the imaging start time. Record the approximate pixel coordinate range of these feature regions in the actual image, and locate the corresponding actual geographic coordinates of these features in map software or satellite simulation software.

[0083] 3) Perform left and right boundary matching and record Euler angles. In the satellite simulation software, enable the display function of the sensor's ground coverage area. A rectangular coverage area representing the simulated imaging strip will appear in the map window. The operator simultaneously views the simulated strip in the satellite simulation software and the actual satellite image. Focus attention on the left and right boundaries of the simulated strip. By adjusting the Euler angle parameter (denoted here as EulerA) under the "jog" or "offset angle" settings in the sensor's "pointing" attribute, observe the left and right movement of the simulated strip on the map. The goal of the adjustment is to make the left boundary line of the simulated strip exactly pass through or be adjacent to the map position corresponding to the selected typical feature area of ​​the left boundary in the actual image; at the same time, make the right boundary line of the simulated strip exactly pass through or be adjacent to the map position corresponding to the selected typical feature area of ​​the right boundary in the actual image. By repeatedly fine-tuning the value of EulerA, until the left and right boundary coverage areas of the simulated strip and the feature areas of the left and right edges of the actual image achieve the best visual or coordinate measurement match. At this point, lock the sensor's pointing setting and accurately record the Euler angle value under the Pointing item in the current sensor attribute, thus obtaining the first Euler angle. This angle value directly represents the amount of angle correction required to compensate for deviations in the vertical rail direction.

[0084] 4) Perform imaging start-time boundary matching and record the Euler angle EulerB. Keep the EulerA value adjusted in the previous step for left and right boundary matching unchanged. Now turn your attention to the upper boundary of the simulated strip, which corresponds to the start time of the imaging scan. Observe the vertical movement of the simulated strip on the map by adjusting another Euler angle parameter in the sensor's "Pointing" attribute (denoted as EulerB here). The goal of the adjustment is to make the start-time boundary (upper boundary) of the simulated strip exactly pass through or be adjacent to the map position corresponding to the selected typical feature area of ​​the imaging start time in the actual image. By repeatedly fine-tuning the value of EulerB, until the area covered by the upper boundary of the simulated strip achieves optimal matching with the feature area of ​​the upper edge of the actual image. At this point, record the corresponding Euler angle value under the Pointing item in the sensor attributes, which is the second Euler angle EulerB. This angle value directly represents the amount of angle correction required to compensate for deviations along the track direction.

[0085] 5) Handling the Euler angle EulerC around the viewing axis. For optical pushbroom cameras, the rotational deviation around the viewing axis (i.e., the optical axis) has a complex effect on the geometric deformation within a single image and is often difficult to accurately separate using simple boundary matching. In most engineering practices, it can be assumed that this deviation is small or has little coupling effect with other deviations. Therefore, in this comparative analysis, the third Euler angle EulerC under the Pointing item in the sensor properties is usually set to 0, or fixed to a priori value obtained through other calibration methods. Record this setting.

[0086] S103. Based on the error angle between the satellite measurement coordinate system and the camera line-of-sight coordinate system obtained from imaging data from regions with different Beta angles, latitudes, and longitudes, a mathematical model is performed on the error angle. Specifically: 1) Determine the range of Beta angle variation and the coverage area of ​​the nadir latitude within the mission cycle based on the satellite's orbit. Using a high-precision satellite orbital dynamics model or specialized orbital prediction software, input the satellite's orbital elements to calculate a continuous curve showing the change of the solar Beta angle over time within a complete mission cycle (usually one year). The solar Beta angle is defined as the angle between the solar direction vector and the satellite's orbital plane. From this curve, the maximum value of the Beta angle that can be reached throughout the entire cycle can be extracted and denoted as . And the minimum value, denoted as This allows us to determine the theoretical range of variation for the Beta angle. .symbol Represents the maximum value of the Beta angle, symbol This represents the minimum value of the Beta angle. Simultaneously, based on the satellite's orbital inclination, right ascension of the ascending node, and other parameters, combined with the Earth's rotation, the geographical latitude limit that the satellite's nadir point can reach within the same mission cycle is calculated. The maximum latitude range covered by the satellite's routine imaging mission is determined and denoted as . The minimum value is denoted as This determines the latitudinal coverage area. .symbol Indicates the maximum latitude coverage, symbol This represents the minimum value of latitude coverage.

[0087] 2) Design multiple sets of different deviation angle calibration modeling imaging test conditions. Within the determined range of Beta angle variation... Within, at the first preset interval Perform uniform discrete sampling. First preset interval. This is a constant integer, such as 5 degrees, determined based on modeling accuracy requirements and satellite imaging opportunities. The formula for calculating the discrete Beta angle value sequence obtained from sampling is: ,in, The function represents a rounding operation on the calculation result. Indicates the first A discrete Beta angle sample value. It is the number of sampling points, and its value makes No more than . Indicates the sampling interval.

[0088] Within the established latitude coverage area Within, at the second preset interval Perform uniform discrete sampling. Second preset interval. This is another constant value determined based on latitudinal zone characteristics and imaging planning, such as 2 degrees. The formula for calculating the discrete latitude value sequence obtained from sampling is: ,in, Indicates the first A discrete latitude sample value. It represents the number of latitude sampling points. This represents the latitude sampling interval. Each discrete Beta angle value... With each discrete latitude value Pair them up. For each pair... Its corresponding longitude value Taken as when the satellite flies over that latitude At that time, with the current Beta perspective The longitude of the nadir point at the corresponding imaging time. This longitude value is determined by orbit prediction software within a specific time window. Therefore, each group... This constitutes a specific set of deviation angle calibration modeling imaging test conditions. A total of [number] will be generated. Different experimental conditions were used.

[0089] 3) Set the onboard software status before performing imaging experiments. To obtain raw, uncompensated pointing deviation data, the "deviation correction function" in the onboard software must be set to "off" via ground command during this series of calibration modeling imaging experiments. This setting ensures that the onboard attitude determination system uses the initial installation matrix of the star sensor without introducing any model-based corrections, thus ensuring that the subsequently acquired error angles accurately reflect the satellite's inherent pointing deviation.

[0090] 4) Perform imaging experiments and acquire error angle data. The ground control center, according to the plan, sequentially applies each set of generated imaging experiment conditions. The imaging mission command is transmitted to the satellite. Based on a unified attitude reference coordinate system (without bias correction at this point), the satellite controls its optical camera to image targets at specified latitude and longitude, and transmits the actual imaging data and its auxiliary data. For each transmitted imaging data, ground processing personnel strictly follow the method detailed in S2: establishing a corresponding simulation model, selecting typical feature regions, and adjusting the sensor's Euler angles to achieve left-right boundary matching and imaging start-time boundary matching, thereby calculating the first Euler angle corresponding to that imaging session. Second Euler angle After traversing all M imaging experiments, a dataset containing M sets of observations is obtained, with each data sample including imaging conditions. and the observed error angle .

[0091] 5) Construct the mathematical relationship for the error angle. Based on engineering experience and analysis of satellite thermal deformation and structural deformation mechanisms, assume the first Euler angle. (Primarily corresponding to pitch deviation) has a linear dependence on the Beta angle, longitude, and latitude during imaging. Therefore, the first mathematical relationship can be constructed as follows: Similarly, construct the second Euler angle. The second mathematical relationship between (primarily corresponding to roll deviation) and imaging conditions: ,in, It is an observation. It is the constant deviation term to be determined. , , They are respectively the relative values ​​to be determined , , The linear coefficients.

[0092] 6) Calculate the coefficients in the first mathematical relation. , , , First, calculate the constant terms. Assume a total of [number] were obtained. One valid first Euler angle observation sample ,but This can be obtained by calculating the arithmetic mean of these observations: , This represents the total number of valid observation samples. This is the sample index. After calculation, for each observed sample, calculate its residual after removing constant bias. Utilizing the imaging conditions of all samples and residual Construct a new linear fitting problem: The coefficients can be solved by applying the least squares method to fit the problem. , , Least squares fitting is achieved by constructing a design matrix. (its first) Behavior ) and observation vector (Its elements are) Solve the normal equations Obtain the parameter vector .

[0093] 7) First, calculate the average value of the second Euler angle observations: Then calculate the residuals. Finally, utilize the conditions. and residual The coefficients were fitted using the least squares method. , , .

[0094] 8) Determine the error angle model coefficients around the line of sight. The impact of deviations around the optical camera's line of sight (yaw rate) is usually small and its relationship with the Beta angle and latitude / longitude is complex. In the simplified model, the coefficients in its mathematical model can be directly set to zero. That is, take: This completes the calculation of all coefficients in the mathematical model for the error angle. , , Solving for these coefficients and substituting them back into the first and second mathematical relations yields a complete mathematical model of the error angle that can be used for prediction.

[0095] S104. Based on the mathematical model of the error angle between the satellite measurement coordinate system and the camera line-of-sight coordinate system, the pointing deviation is corrected by updating the star-sensor installation matrix in the on-orbit software. See the specific implementation process in S8 for details.

[0096] In the above embodiments, although the steps are numbered S1, S2, etc., they are only specific embodiments given by the present invention. Those skilled in the art can adjust the execution order of S1, S2, etc. according to the actual situation. The scheme after adjusting the order is also within the protection scope of the present invention. It can be understood that in some embodiments, some or all of the above embodiments may be included.

[0097] like Figure 2 As shown in the figure, an optical remote sensing satellite imaging pointing deviation correction system according to an embodiment of the present invention includes: a measurement reference star sensor selection module, a time error compensation module, a transformation relationship determination module, a coordinate system module, an actual imaging data acquisition module, an error angle acquisition module, a mathematical model establishment module, and a correction module. The measurement reference star sensor selection module is used to: select one of at least three star sensors configured on the satellite as the measurement reference star sensor according to the selection principle, wherein the satellite is equipped with a gyroscope as an inertial attitude measurement device. The time error compensation module is used to: compensate for the time error of the measurement data of each star sensor using the measurement data of the gyroscope to obtain time-uniform measurement data. The transformation relationship determination module is used to: update the transformation relationship between the measurement coordinate systems used by each star sensor according to the time-uniform measurement data. The coordinate system module is used to: determine the satellite attitude based on the updated transformation relationship. The reference coordinate system used in the simulation is unified with the measurement coordinate system used by the reference star sensor to establish a unified attitude reference coordinate system. The actual imaging data acquisition module is used to control the satellite to perform imaging tasks in different Beta angles, latitudes and longitudes based on the unified attitude reference coordinate system, and to acquire the actual imaging data of the optical camera equipped on the satellite. The error angle acquisition module is used to compare the actual imaging data with the map loaded in the simulation model to determine the error angle between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera. The mathematical model establishment module is used to establish a mathematical model of the error angle based on the error angles acquired in different Beta angles, latitudes and longitudes. The correction module is used to generate pointing deviation correction coefficients based on the mathematical model of the error angles, and to update the installation matrix of each star sensor in the on-board software through on-orbit annotation to correct the satellite's imaging pointing deviation.

[0098] Optionally, in the above technical solution, the time error compensation module is specifically used to: use the angular velocity measurement value output by the gyroscope to calculate the attitude change quaternion corresponding to the time difference between the measurement time of each non-reference star sensor and the measurement reference star sensor, and apply the attitude change quaternion to the measurement data of the non-reference star sensor, so as to unify the measurement data of each non-reference star sensor to the measurement time of the measurement reference star sensor, and obtain time-uniformed measurement data.

[0099] Optionally, in the above technical solution, the error angle acquisition module is specifically used for: establishing a simulation model based on satellite orbit data, satellite attitude data, latitude and longitude of the imaging target point, and field-of-view parameters of the optical camera at the imaging time, and loading a map into the simulation model. The simulation model is equipped with a sensor whose field-of-view parameters are consistent with those of the optical camera; selecting typical areas as the left and right boundary matching areas and the imaging start time boundary matching areas, respectively; adjusting the Euler angles of the sensor in the simulation model to make the left and right boundary coverage areas of the simulated imaging strip on the map consistent with the left and right boundary matching areas in the actual imaging data, thus obtaining the first Euler angle; adjusting the Euler angles of the sensor in the simulation model to make the simulated imaging strip start time boundary coverage area on the map consistent with the imaging start time boundary matching area in the actual imaging data, thus obtaining the second Euler angle; and determining the first and second Euler angles as error angles characterizing the deviation between the unified attitude reference coordinate system and the line-of-view coordinate system used by the optical camera.

[0100] Optionally, in the above technical solution, the mathematical model establishment module is specifically used for: determining the range of variation of the Beta angle and the coverage range of the nadir latitude within the mission cycle based on the satellite's orbit; selecting multiple discrete Beta angle values ​​at a first preset interval within the range of Beta angle variation, and selecting multiple discrete latitude values ​​at a second preset interval within the latitude coverage range, and combining each selected Beta angle value and each latitude value with the longitude of the nadir point corresponding to the imaging time to form multiple sets of different imaging test conditions; under each set of imaging test conditions, controlling the satellite to perform imaging based on a unified attitude reference coordinate system, acquiring the corresponding actual imaging data, and obtaining multiple sets of error angles including the first Euler angle and the second Euler angle; based on the multiple sets of error angles, constructing a first mathematical relationship between the first Euler angle and the Beta angle, latitude, and longitude at the time of imaging, and a second mathematical relationship between the second Euler angle and the Beta angle, latitude, and longitude at the time of imaging; using the least squares method to fit the unknown coefficients in the first and second mathematical relationships, and using the first and second mathematical relationships with the completed coefficient fitting as the mathematical model of the error angles.

[0101] It should be noted that the beneficial effects of the optical remote sensing satellite imaging pointing deviation correction system provided in the above embodiments are the same as those of the optical remote sensing satellite imaging pointing deviation correction method described above, and will not be repeated here. Furthermore, the system and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation processes are detailed in the method embodiments, and will not be repeated here.

[0102] An electronic device according to an embodiment of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the aforementioned optical remote sensing satellite imaging pointing deviation correction methods. A computer-readable storage medium according to an embodiment of the present invention stores a computer program, which, when executed by a processor, implements any of the aforementioned optical remote sensing satellite imaging pointing deviation correction methods.

[0103] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. An optical remote sensing satellite imaging pointing deviation correction method, characterized in that, include: According to the principle of selecting the reference star sensor, one of the at least three star sensors configured on the satellite is selected as the measurement reference star sensor, wherein the satellite is configured with a gyroscope as an inertial attitude measurement device; The measurement data of each star sensor is compensated for time error by using the measurement data of the gyroscope to obtain time-uniform measurement data; Based on the time-unified measurement data, update the transformation relationship between the measurement coordinate systems used by each star sensor; Based on the updated transformation relationship, the reference coordinate system used to determine the satellite attitude is unified with the measurement coordinate system used by the measurement reference star sensor to establish a unified attitude reference coordinate system. Based on the unified attitude reference coordinate system, the satellite is controlled to perform imaging tasks in different Beta angles, different latitudes and different longitudes to obtain the actual imaging data of the optical camera equipped on the satellite. The actual imaging data is compared with the map loaded in the simulation model to determine the error angle between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera by comparing the typical feature regions of the images. A mathematical model for the error angle is established based on the error angles obtained from different Beta angles, latitudes, and longitudes. Based on the mathematical model of the error angle, a pointing deviation correction coefficient is generated, and the installation matrix of each star sensor in the onboard software is updated by on-orbit injection to correct the satellite's imaging pointing deviation.

2. The method for correcting imaging pointing deviation of an optical remote sensing satellite according to claim 1, characterized in that, The measurement data of each star sensor is compensated for time errors using the measurement data from the gyroscope to obtain time-uniform measurement data, including: Using the angular velocity measurement value output by the gyroscope, the attitude change quaternion corresponding to the time difference between each non-reference star sensor and the measurement reference star sensor is calculated. This attitude change quaternion is then applied to the measurement data of the non-reference star sensor to unify the measurement data of each non-reference star sensor to the measurement time of the measurement reference star sensor, thereby obtaining time-uniformed measurement data.

3. The method for correcting imaging pointing deviation of an optical remote sensing satellite according to claim 2, characterized in that, The actual imaging data is compared with the map loaded in the simulation model to determine the typical feature regions of the images, and the error angle between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera is determined, including: Based on satellite orbit data, satellite attitude data, latitude and longitude of the imaging target point, and field of view parameters of the optical camera at the imaging time, a simulation model is established and a map is loaded into the simulation model. The simulation model is equipped with a sensor whose field of view parameters are consistent with those of the optical camera. Typical regions were selected as the left and right boundary matching regions and the boundary matching region at the imaging start time, respectively. By adjusting the Euler angles of the sensor in the simulation model, the left and right boundary coverage areas of the simulated imaging strip on the map by the sensor are made consistent with the left and right boundary matching areas in the actual imaging data, thus obtaining the first Euler angle. By adjusting the Euler angles of the sensor in the simulation model, the boundary coverage area of ​​the simulated imaging strip on the map is made consistent with the boundary matching area of ​​the imaging start time in the actual imaging data, thus obtaining the second Euler angle. The first Euler angle and the second Euler angle are defined as error angles characterizing the deviation between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera.

4. The method for correcting imaging pointing deviation of an optical remote sensing satellite according to claim 3, characterized in that, Based on the error angles obtained from different Beta angles, latitudes, and longitudes, a mathematical model for the error angles is established, including: The range of Beta angle variation and the coverage area of ​​the nadir latitude are determined based on the satellite's orbit during the mission cycle. Within the range of Beta angle variation, multiple discrete Beta angle values ​​are selected at a first preset interval. Within the range of latitude coverage, multiple discrete latitude values ​​are selected at a second preset interval. Each selected Beta angle value and each latitude value are combined with the longitude of the nadir point corresponding to the imaging time to form multiple sets of different imaging test conditions. Under each set of imaging test conditions, the satellite is controlled to perform imaging based on the unified attitude reference coordinate system to obtain the corresponding actual imaging data and obtain multiple sets of error angles including the first Euler angle and the second Euler angle. Based on the multiple sets of error angles, a first mathematical relationship is constructed between the first Euler angle and the Beta angle, latitude and longitude at the time of imaging, and a second mathematical relationship is constructed between the second Euler angle and the Beta angle, latitude and longitude at the time of imaging. The least squares method is used to fit the unknown coefficients in the first mathematical relationship and the second mathematical relationship, and the first mathematical relationship and the second mathematical relationship with the completed coefficient fitting are used as the mathematical model of the error angle.

5. A pointing deviation correction system for optical remote sensing satellite imaging, characterized in that, include: The system includes a measurement reference star sensor selection module, a time error compensation module, a transformation relationship determination module, a coordinate system module, an actual imaging data acquisition module, an error angle acquisition module, a mathematical model establishment module, and a correction module. The measurement reference star sensor selection module is used to: select one of the at least three star sensors configured on the satellite as the measurement reference star sensor according to the reference star sensor selection principle, wherein the satellite is configured with a gyroscope as an inertial attitude measurement device; The time error compensation module is used to: compensate for the time error of the measurement data of each star sensor using the measurement data of the gyroscope, so as to obtain time-uniform measurement data; The transformation relationship determination module is used to: update the transformation relationship between the measurement coordinate systems used by each star sensor based on the time-unified measurement data; The coordinate system module is used to: unify the reference coordinate system used to determine the satellite attitude with the measurement coordinate system used by the measurement reference star sensor based on the updated transformation relationship, so as to establish a unified attitude reference coordinate system. The actual imaging data acquisition module is used to: control the satellite to perform imaging tasks in different Beta angles, different latitudes and different longitudes based on the unified attitude reference coordinate system, and acquire the actual imaging data of the optical camera equipped on the satellite; The error angle acquisition module is used to: compare the actual imaging data with the map loaded in the simulation model to determine the error angle between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera; The mathematical model building module is used to: build a mathematical model of the error angle based on the error angle obtained in different Beta angle, latitude and longitude regions; The correction module is used to: generate pointing deviation correction coefficients based on the mathematical model of the error angle, and update the installation matrix of each star sensor in the on-board software by means of on-orbit annotation, so as to correct the imaging pointing deviation of the satellite.

6. The optical remote sensing satellite imaging pointing deviation correction system according to claim 5, characterized in that, The time error compensation module is specifically used to: calculate the attitude change quaternion corresponding to the time difference between the measurement time of each non-reference star sensor and the measurement reference star sensor using the angular velocity measurement value output by the gyroscope, and apply the attitude change quaternion to the measurement data of the non-reference star sensor to unify the measurement data of each non-reference star sensor to the measurement time of the measurement reference star sensor, thereby obtaining time-uniformed measurement data.

7. The optical remote sensing satellite imaging pointing deviation correction system according to claim 6, characterized in that, The error angle acquisition module is specifically used for: Based on satellite orbit data, satellite attitude data, latitude and longitude of the imaging target point, and field of view parameters of the optical camera at the imaging time, a simulation model is established and a map is loaded into the simulation model. The simulation model is equipped with a sensor whose field of view parameters are consistent with those of the optical camera. Typical regions were selected as the left and right boundary matching regions and the boundary matching region at the imaging start time, respectively. By adjusting the Euler angles of the sensor in the simulation model, the left and right boundary coverage areas of the simulated imaging strip on the map by the sensor are made consistent with the left and right boundary matching areas in the actual imaging data, thus obtaining the first Euler angle. By adjusting the Euler angles of the sensor in the simulation model, the boundary coverage area of ​​the simulated imaging strip on the map is made consistent with the boundary matching area of ​​the imaging start time in the actual imaging data, thus obtaining the second Euler angle. The first Euler angle and the second Euler angle are defined as error angles characterizing the deviation between the unified attitude reference coordinate system and the line-of-sight coordinate system used by the optical camera.

8. The optical remote sensing satellite imaging pointing deviation correction system according to claim 7, characterized in that, The mathematical model building module is specifically used for: The range of Beta angle variation and the coverage area of ​​the nadir latitude are determined based on the satellite's orbit during the mission cycle. Within the range of Beta angle variation, multiple discrete Beta angle values ​​are selected at a first preset interval. Within the range of latitude coverage, multiple discrete latitude values ​​are selected at a second preset interval. Each selected Beta angle value and each latitude value are combined with the longitude of the nadir point corresponding to the imaging time to form multiple sets of different imaging test conditions. Under each set of imaging test conditions, the satellite is controlled to perform imaging based on the unified attitude reference coordinate system to obtain the corresponding actual imaging data and obtain multiple sets of error angles including the first Euler angle and the second Euler angle. Based on the multiple sets of error angles, a first mathematical relationship is constructed between the first Euler angle and the Beta angle, latitude and longitude at the time of imaging, and a second mathematical relationship is constructed between the second Euler angle and the Beta angle, latitude and longitude at the time of imaging. The least squares method is used to fit the unknown coefficients in the first mathematical relationship and the second mathematical relationship, and the first mathematical relationship and the second mathematical relationship with the completed coefficient fitting are used as the mathematical model of the error angle.

9. An electronic device, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the optical remote sensing satellite imaging pointing deviation correction method according to any one of claims 1 to 4.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the optical remote sensing satellite imaging pointing deviation correction method according to any one of claims 1 to 4.

Images