A high-precision generalized dynamic image motion compensation method, device, equipment and storage medium

Abstract

The application provides a high-precision generalized dynamic image motion compensation method, device, equipment and storage medium, six homogeneous coordinate transformation matrices between a geographic coordinate system and an image plane coordinate system are sequentially multiplied to form a total transformation matrix, and a multi-stage series kinematics mapping of a ground point to an image plane point is completely expressed in a single mathematical framework. A homogeneous position vector of the image plane point obtained by the total transformation matrix is directly derived with respect to time, so that each motion source and its cross-coupling effect are naturally included in an image motion velocity vector expression in an analytical form, and a truncation error caused by component separation and superposition is avoided. Accurate image motion velocity vector values are obtained by substituting real-time parameters into the expression, and then a deflection angle correction amount and a TDI charge transfer matching rate are solved to perform compensation.

Description

Technical Field

[0001] This invention relates to the field of aerospace remote sensing imaging technology, and in particular to a high-precision generalized dynamic image shift compensation method, device, equipment, and storage medium. Background Technology

[0002] During the imaging process of a spaceborne pushbroom camera, the projected position of a ground target point on the focal plane changes continuously over time, resulting in image shift. The essence of image shift is a multi-source coupled kinematic problem: satellite orbital motion, Earth's rotation, satellite three-axis attitude changes, perspective effects from side-swing imaging, and terrain undulations, each with different magnitudes, directions, and time-varying characteristics, jointly affect the instantaneous motion state of the same image point.

[0003] Existing image motion compensation schemes separate the aforementioned multi-source motions—first, they independently estimate the image motion contribution of each major component, then they superimpose the results to obtain a composite image motion vector, and finally, they determine the yaw angle and TDI line transfer rate based on this vector. This "component separation-linear superposition" processing framework implicitly assumes a fundamental assumption: the contributions of each motion component to the image motion are independent of each other and can be linearly added.

[0004] However, the above assumptions are not strictly valid in physics. The mapping from the ground target point to the image plane point is essentially a multi-stage kinematic chain. Each stage of coordinate transformation (e.g., from the Earth-fixed frame to the inertial frame, from the orbital frame to the local frame) is a nonlinear function of the output of the previous stage, and there is an inherent coupling relationship between each stage of transformation. When the satellite performs large-angle side-swings, multi-axis attitude maneuvers, or passes through high-latitude regions (where the Earth's rotation component is significant), the magnitude of the cross-coupling terms between the motion sources increases, and the inherent truncation error of the "component separation-linear superposition" framework increases significantly, causing the residual image shift to exceed the acceptable range.

[0005] In view of the above, this application is hereby submitted. Summary of the Invention

[0006] This invention discloses a high-precision generalized dynamic image shift compensation method, device, equipment, and storage medium, aiming to solve the problem that existing image shift compensation methods, which model each motion source independently and then linearly superimpose them, cannot accurately reflect the cross-coupling effect between multiple motion sources, resulting in excessive residual image shift.

[0007] The first embodiment of the present invention provides a high-precision generalized dynamic image shift compensation method, comprising: Homogeneous coordinate transformation matrices were established between the geographic coordinate system and the Earth-fixed coordinate system, between the Earth-fixed coordinate system and the geocentric inertial coordinate system, between the geocentric inertial coordinate system and the satellite orbit coordinate system, between the satellite orbit coordinate system and the satellite body coordinate system, between the satellite body coordinate system and the camera coordinate system, and between the camera coordinate system and the image plane coordinate system, resulting in six homogeneous coordinate transformation matrices. The six homogeneous coordinate transformation matrices are multiplied sequentially in the order of transformation to obtain the total transformation matrix. The homogeneous position vectors of the ground points are then transformed to the image plane coordinate system using the total transformation matrix to obtain the homogeneous position vectors of the image plane points. Differentiating the homogeneous position vector of the image point with respect to time yields an analytical expression for the image velocity vector, which includes the effects of each motion source and their cross-coupling. Substitute the real-time acquired satellite orbit parameters, attitude parameters, time information, and digital elevation model data into the analytical expression to calculate the image motion velocity vector value at the current imaging moment; The deflection angle correction and the charge transfer matching rate of the TDI sensor are calculated based on the image velocity vector value and sent to the satellite attitude control system and camera electronics system respectively to perform compensation.

[0008] Preferably, the analytical expression for the image velocity vector is: V_image = d(T_total) / dt · P_ground + T_total · d(P_ground) / dt Where T_total is the total transformation matrix, and P_ground is the homogeneous position vector of the ground point in the geographic coordinate system.

[0009] Preferably, the homogeneous coordinate transformation matrix between the Earth-fixed coordinate system and the geocentric inertial coordinate system includes the time-varying parameter of the Earth's rotation angular velocity; the homogeneous coordinate transformation matrix between the satellite orbit coordinate system and the satellite body coordinate system is determined by the satellite roll angle, pitch angle, and yaw angle; and the homogeneous coordinate transformation matrix between the satellite body coordinate system and the camera coordinate system is determined by the camera mounting angle.

[0010] Preferably, the digital elevation model data is used to correct the radial component of the homogeneous position vector of the ground point to reflect the actual altitude of the ground point.

[0011] Preferably, the deflection angle correction is the arctangent of the ratio of the cross-track component to the along-track component of the image displacement velocity vector value, and the charge transfer matching rate is the ratio of the magnitude of the along-track component to the row spacing of the TDI sensor.

[0012] Preferably, the method further includes: acquiring on-orbit image data and extracting image edge sharpness indicators, and correcting key parameters of the homogeneous coordinate transformation matrix online based on the indicators, wherein the key parameters include the camera's effective focal length and camera mounting angle.

[0013] Preferably, the online correction aims to maximize the image edge sharpness index and uses an iterative optimization method to update the key parameters.

[0014] The second embodiment of the present invention provides a high-precision generalized dynamic image shift compensation device, comprising: The coordinate transformation matrix establishment module is used to establish homogeneous coordinate transformation matrices between the geographic coordinate system and the Earth-fixed coordinate system, between the Earth-fixed coordinate system and the geocentric inertial coordinate system, between the geocentric inertial coordinate system and the satellite orbit coordinate system, between the satellite orbit coordinate system and the satellite body coordinate system, between the satellite body coordinate system and the camera coordinate system, and between the camera coordinate system and the image plane coordinate system, resulting in six homogeneous coordinate transformation matrices. The image plane point transformation module is used to multiply the six homogeneous coordinate transformation matrices in the order of transformation to obtain the total transformation matrix, and use the total transformation matrix to transform the homogeneous position vector of the ground point to the image plane coordinate system to obtain the homogeneous position vector of the image plane point. The image velocity calculation module is used to differentiate the homogeneous position vector of the image point with respect to time to obtain an analytical expression for the image velocity vector that includes each motion source and its cross-coupling effect; The real-time calculation module is used to substitute the real-time acquired satellite orbit parameters, attitude parameters, time information and digital elevation model data into the analytical expression to calculate the image motion velocity vector value at the current imaging moment; The compensation execution module is used to calculate the deflection angle correction and the charge transfer matching rate of the TDI sensor based on the image displacement velocity vector value, and send them to the satellite attitude control system and camera electronics system respectively to perform compensation.

[0015] The third embodiment of the present invention provides a high-precision generalized dynamic image shift compensation device, including a memory and a processor. The memory stores a computer program, which can be executed by the processor to implement a high-precision generalized dynamic image shift compensation method as described in any of the above embodiments.

[0016] The fourth embodiment of the present invention provides a computer-readable storage medium, characterized in that it stores a computer program, which can be executed by the processor of the device in which the computer-readable storage medium is located, to implement a high-precision generalized dynamic image shift compensation method as described in any of the above claims.

[0017] Based on the high-precision generalized dynamic image shift compensation method, device, equipment, and storage medium provided by this invention, six homogeneous coordinate transformation matrices between the geographic coordinate system and the image plane coordinate system are multiplied sequentially to form a total transformation matrix, which completely expresses the multi-level serial kinematic mapping from ground points to image plane points within a single mathematical framework. The homogeneous position vector of the image plane points obtained by the total transformation matrix is ​​directly differentiated with respect to time, so that each motion source and its cross-coupling effect are naturally included analytically in the image shift velocity vector expression, avoiding the truncation error caused by component separation and superposition. Substituting real-time parameters into this expression yields accurate image shift velocity vector values, which are then used to calculate the yaw angle correction and TDI charge transfer matching rate to perform compensation. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a high-precision generalized dynamic image shift compensation method provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of a high-precision generalized dynamic image shift compensation device provided in the second embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To better understand the technical solution of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0021] This invention discloses a high-precision generalized dynamic image shift compensation method, device, equipment, and storage medium, aiming to solve the problem that existing image shift compensation methods, which model each motion source independently and then linearly superimpose them, cannot accurately reflect the cross-coupling effect between multiple motion sources, resulting in excessive residual image shift.

[0022] Please see Figure 1 The first embodiment of the present invention provides a high-precision generalized dynamic image shift compensation method, which can be executed by a high-precision generalized dynamic image shift compensation device (hereinafter referred to as the compensation device and system), and in particular, executed by one or more processors within the compensation device, to at least implement the following steps: S101, establish homogeneous coordinate transformation matrices between the geographic coordinate system and the ground-fixed coordinate system, between the ground-fixed coordinate system and the geocentric inertial coordinate system, between the geocentric inertial coordinate system and the satellite orbit coordinate system, between the satellite orbit coordinate system and the satellite body coordinate system, between the satellite body coordinate system and the camera coordinate system, and between the camera coordinate system and the image plane coordinate system, to obtain six homogeneous coordinate transformation matrices. In this embodiment, the compensation device can be a terminal with data processing capabilities, such as a spaceborne embedded processor or an on-board computing platform like an FPGA. The compensation device can be equipped with a corresponding operating system and application software, and the functions required in this embodiment can be realized through the combination of the operating system and application software.

[0023] In this embodiment, in order to construct a complete kinematic chain from the ground target point to the camera image plane, it is necessary to establish six homogeneous coordinate transformation matrices between seven coordinate systems in sequence.

[0024] First, a homogeneous coordinate transformation matrix is ​​established between the geographic coordinate system and the geodetic coordinate system. The geographic coordinate system has the ground target point as its origin, the local normal direction as one axis, and the local east and north directions as the other two axes. The geodetic coordinate system has the Earth's center of mass as its origin and rotates with the Earth's rotation. The transformation relationship between the two is determined by the geodetic longitude and latitude of the ground target point. Based on this, a rotation matrix containing trigonometric functions of longitude and latitude is constructed, and the geocentric distance of the target point is embedded as a translation component into the translation column of the homogeneous matrix, thus obtaining the first homogeneous coordinate transformation matrix.

[0025] Next, a homogeneous coordinate transformation matrix is ​​established between the Earth-fixed coordinate system and the geocentric inertial coordinate system. The geocentric inertial coordinate system has its origin at the Earth's center of mass, and its coordinate axes remain fixed in inertial space. The difference between the two arises from the Earth's rotation. Therefore, the rotation component of this homogeneous coordinate transformation matrix consists of the rotation angle determined by the product of the Earth's rotational angular velocity and the current time, with zero translation component, thus yielding the second homogeneous coordinate transformation matrix. Since the product of the Earth's rotational angular velocity and time continuously changes over time, this matrix is ​​time-varying. This characteristic allows the contribution of the Earth's rotation to image displacement to be naturally incorporated when subsequently calculating the time derivative of the total transformation matrix.

[0026] A homogeneous coordinate transformation matrix is ​​then established between the geocentric inertial coordinate system and the satellite orbital coordinate system. The satellite orbital coordinate system has its origin at the satellite's center of mass, and its three axes are along the line connecting the satellite to the geocentric coordinate system, the direction in the orbital plane aligned with the velocity direction, and the direction normal to the orbital plane. The rotational part of this homogeneous coordinate transformation matrix is ​​determined by normalizing and cross-producting the satellite's position and velocity vectors in inertial space, while the translational part is the position vector of the satellite's center of mass in the geocentric inertial coordinate system. Together, they constitute a third homogeneous coordinate transformation matrix. The satellite's orbital root numbers or ephemeris data provide real-time input to this matrix.

[0027] Then, a homogeneous coordinate transformation matrix is ​​established between the satellite orbital coordinate system and the satellite body coordinate system. The satellite body coordinate system is fixed to the satellite structure, and its three axes are aligned with the satellite's roll, pitch, and yaw axes. When the satellite's attitude is perfectly aligned with the orbital coordinate system, the two systems coincide; however, in actual flight, there is a three-axis attitude deviation between them. The rotation part of this homogeneous coordinate transformation matrix is ​​obtained by multiplying three basic rotation matrices constructed sequentially according to a predetermined rotation order for the roll, pitch, and yaw angles. The translation part is zero, thus yielding the fourth homogeneous coordinate transformation matrix. Real-time measurements from the satellite's star sensor and gyroscope provide the data source for the three attitude angles in this matrix.

[0028] Next, a homogeneous coordinate transformation matrix is ​​established between the satellite body coordinate system and the camera coordinate system. The camera coordinate system has its origin at the principal point of the camera's optical system, with the optical axis as one axis, and the other two axes aligned with the row and column directions of the focal plane. The camera is fixed to the satellite platform via a mounting structure, and the relationship between the two coordinate systems is determined by the camera's mounting angle. The rotation part of this homogeneous coordinate transformation matrix is ​​represented by a rotation matrix constructed from the three mounting angles, while the translation part is the offset vector of the camera's principal point relative to the satellite's center of mass in the body coordinate system, thus yielding the fifth homogeneous coordinate transformation matrix. The mounting angles and offset vectors are typically obtained during ground assembly and adjustment and can be corrected during subsequent on-orbit calibration.

[0029] Finally, a homogeneous coordinate transformation matrix is ​​established between the camera coordinate system and the image plane coordinate system. The image plane coordinate system is defined on the focal plane, with the reference point of the pixel array as the origin and the row and column directions as the two axes. Three-dimensional points in the camera coordinate system are mapped onto the two-dimensional image plane through perspective projection; this projection relationship is determined by the camera's effective focal length. The perspective projection relationship is expressed in homogeneous matrix form, with the focal length appearing as a scaling factor in the matrix, thus obtaining the sixth homogeneous coordinate transformation matrix. The effective focal length is also initially obtained during ground calibration and can be corrected through on-orbit image quality feedback.

[0030] S102, the six homogeneous coordinate transformation matrices are multiplied sequentially in the order of transformation to obtain the total transformation matrix, and the homogeneous position vector of the ground point is transformed to the image plane coordinate system using the total transformation matrix to obtain the homogeneous position vector of the image plane point. After obtaining the six homogeneous coordinate transformation matrices, they are multiplied sequentially according to the transformation order from the geographic coordinate system to the image plane coordinate system. That is, following the order of the coordinate transformation chain, the six homogeneous coordinate transformation matrices are multiplied sequentially on the left. This ensures that the homogeneous coordinate transformation matrix between the geographic coordinate system and the Earth-fixed coordinate system acts first on the homogeneous position vector of the ground point, and the homogeneous coordinate transformation matrix between the camera coordinate system and the image plane coordinate system acts last. The result of multiplying the six matrices is the total transformation matrix. This total transformation matrix, in the form of a single matrix, completely encapsulates all the geometric and kinematic relationships of the ground target point mapped step-by-step from the seven coordinate systems to the image plane. This includes the spatial position of the ground point, the time-varying rotation of the Earth's rotation, the translation and rotation caused by the satellite's orbital motion, the satellite's three-axis attitude deviations, the camera's installation offset and rotation, and all physical effects such as optical perspective projection scaling. Because homogeneous coordinates are used, the three types of transformation operations—rotation, translation, and projection scaling—are unified in matrix multiplication. The nonlinear coupling relationships between each level of transformation are naturally preserved during matrix multiplication, rather than being artificially separated or approximately truncated.

[0031] After obtaining the total transformation matrix, the three-dimensional coordinates of the ground target point in the geographic coordinate system are represented in homogeneous form as a four-dimensional column vector. This involves adding a homogeneous component after the three spatial coordinate components to form the homogeneous position vector of the ground point. The spatial coordinate components in this homogeneous position vector are determined by the geodetic longitude, geodetic latitude, and the geocentric distance corrected using digital elevation model data, thus reflecting the true three-dimensional position of the target point on the Earth's surface. Multiplying the total transformation matrix by the homogeneous position vector of the ground point, and performing a matrix-vector multiplication operation, yields the homogeneous representation of the ground target point in the image plane coordinate system, i.e., the image plane point homogeneous position vector. Homogeneous normalization is then performed on this image plane point homogeneous position vector, i.e., dividing the first two components by the third component, to obtain the two-dimensional image plane coordinates of the ground target point on the focal plane. These coordinates, in pixels, represent the precise projection position of the ground target point onto the sensor array.

[0032] S103, Differentiate the homogeneous position vector of the image point with respect to time to obtain an analytical expression for the image velocity vector that includes each motion source and its cross-coupling effect; After obtaining the homogeneous position vector of the image plane point, it is necessary to perform a derivative operation with respect to time to obtain the instantaneous velocity of the image point on the image plane, i.e., the image velocity vector. As shown in the preceding steps, the homogeneous position vector of the image plane point is equal to the total transformation matrix T_total multiplied by the homogeneous position vector P_ground. According to the derivative rule of matrix multiplication, the derivative of this product with respect to time can be decomposed into the sum of two terms: the time derivative of the total transformation matrix multiplied by the homogeneous position vector, plus the time derivative of the total transformation matrix multiplied by the homogeneous position vector. This yields the analytical expression for the image velocity vector: V_image = d(T_total) / dt · P_ground + T_total · d(P_ground) / dt.

[0033] The first term in the above expression, d(T_total) / dt·P_ground, reflects the contribution of the coordinate transformation chain itself to the image shift over time. Since the total transformation matrix is ​​a continuous product of six homogeneous coordinate transformation matrices, when calculating its time derivative, according to the matrix multiplication rule, each submatrix containing time-varying parameters contributes a term during the derivative calculation, while the remaining submatrices retain their original values. Specifically, let the six homogeneous coordinate transformation matrices be denoted as M1 to M6 in order of action. Then, the time derivative of the total transformation matrix, according to the matrix multiplication rule, expands to a sum of six terms, where the k-th term is M6·…·M(k+1)·dMk / dt·M(k-1)·…·M1, meaning only the k-th matrix takes its time derivative, while the remaining matrices retain their original values ​​during the multiplication. In actual calculations, only the time derivatives of the submatrices containing time-varying parameters (i.e., M2, M3, and M4) are non-zero. Therefore, the time derivative of the total transformation matrix actually consists of three non-zero terms.

[0034] The homogeneous coordinate transformation matrix between the Earth-fixed coordinate system and the geocentric inertial coordinate system makes a non-zero contribution to the time derivative due to the inclusion of the time-varying parameter of the Earth's rotation angular velocity. This contribution, after being multiplied with the other five matrices, manifests as the image displacement component caused by the Earth's rotation. The homogeneous coordinate transformation matrix between the geocentric inertial coordinate system and the satellite orbital coordinate system contributes to the image displacement component caused by orbital motion due to the inclusion of the time-varying parameters of the satellite's position and velocity vectors. The homogeneous coordinate transformation matrix between the satellite orbital coordinate system and the satellite body coordinate system contributes to the image displacement component caused by attitude changes due to the inclusion of the time-varying three-axis attitude angles. Since the above contributions are differentiated within the framework of matrix multiplication, the cross-coupling effects between the motion sources are naturally preserved during the multiplication process. For example, the image displacement component caused by satellite attitude changes is modulated by the current orbital position and the Earth's rotation state, and the image displacement component caused by orbital motion during side-swing imaging is affected by the perspective projection scaling relationship. These coupling effects are implicitly contained in the mathematical structure of matrix multiplication differentiation without additional modeling.

[0035] The second term in the above expression, T_total·d(P_ground) / dt, reflects the contribution of the ground point's own position change over time to image motion. During pushbroom imaging, as the satellite flies along its orbit, the sensor's field of view sequentially sweeps across different ground regions. This is equivalent to the imaged ground target point continuously moving along the scanning direction in the geographic coordinate system. Therefore, the derivative of the homogeneous position vector P_ground with respect to time is not zero. This derivative value is related to the rate of change of the ground target point's latitude and longitude, as well as the terrain gradient given by the digital elevation model. After being mapped to the image plane by the total transformation matrix, it manifests as the image motion component caused by ground scanning and terrain undulations.

[0036] Combining the above two items yields a complete analytical expression for the image motion velocity vector. Within a unified homogeneous coordinate transformation framework, this expression, in a closed analytical form, simultaneously incorporates the contributions of all motion sources to image motion, including satellite orbital motion, Earth's rotation, three-axis attitude angular velocity variations, camera mounting relationships, optical perspective projection, and terrain undulations. Furthermore, the cross-coupling effects between these motion sources are precisely preserved as a natural result of matrix multiplication and differentiation, without requiring any component separation or linear superposition approximations.

[0037] S104, Substitute the real-time acquired satellite orbit parameters, attitude parameters, time information and digital elevation model data into the analytical expression to calculate the image motion velocity vector value at the current imaging moment; After obtaining the analytical expression for the image velocity vector, the real-time parameters at the current imaging moment need to be substituted into this expression to calculate the specific value of the image velocity vector. The parameters to be substituted include four categories: satellite orbit parameters, attitude parameters, time information, and digital elevation model data.

[0038] Satellite orbital parameters are provided by the onboard GPS receiver or the ground control system, including the position and velocity vectors of the satellite's center of mass in the geocentric inertial coordinate system at the current moment, or equivalently given in the form of orbital six-element numbers. These parameters are used to determine the values ​​of the homogeneous coordinate transformation matrix and its time derivative between the geocentric inertial coordinate system and the satellite orbital coordinate system. The position and velocity vectors determine the directions of the three basis vectors and their rates of change in the rotation part of this matrix, while the position vector also determines the value of the translation part.

[0039] Attitude parameters are obtained in real time by onboard star sensors and gyroscopes, including the satellite's roll angle, pitch angle, yaw angle, and corresponding three-axis attitude angular velocities relative to the orbital coordinate system. The three attitude angles are used to determine the values ​​of the rotating part of the homogeneous coordinate transformation matrix between the satellite's orbital coordinate system and its body coordinate system, while the three attitude angular velocities are used to calculate the time derivative of this matrix. During side-slip imaging, the roll angle can reach a large value. At this point, the influence of the attitude angles on image motion, after being modulated by the product of the remaining submatrices in the overall transformation matrix, will produce a significant coupling contribution. Accurate substitution of real-time attitude parameters ensures that these coupling effects are correctly reflected in the numerical calculations.

[0040] Time information is provided by a high-precision onboard clock or GPS timing system, accurately marking the current imaging moment. This time information is used for two purposes: first, to calculate the Earth's rotation angle in the homogeneous coordinate transformation matrix between the Earth-fixed coordinate system and the geocentric inertial coordinate system, i.e., the product of the Earth's rotation angular velocity and the current moment, as well as the Earth's rotation angular velocity component in the time derivative of this matrix; second, to precisely align the orbital and attitude parameters with the imaging moment, ensuring that all parameters substituted into the analytical expression are strictly synchronized in time.

[0041] Digital elevation model (DEM) data originates from global or regional DEM databases pre-stored in onboard memory. The elevation values ​​of corresponding ground target points are retrieved based on the geographic latitude and longitude of the current imaging area. When constructing the homogeneous position vector of the ground point, this elevation value is used to correct the radial component of the ground point. Specifically, the geocentric distance of the ground point is corrected from the value on the reference ellipsoid to the sum of the reference ellipsoid height and the actual elevation, ensuring that the radial component of the homogeneous position vector reflects the true three-dimensional spatial position of the ground target point rather than an approximate position on the ellipsoid. This correction is particularly important in mountainous or plateau regions with significant topographic relief, as the elevation difference between the actual elevation of the ground point and the reference ellipsoid alters the slant distance and line-of-sight direction between the target point and the satellite, thus having a non-negligible impact on the image velocity vector. Without elevation correction, additional image motion calculation errors will be introduced in areas with significant topographic relief. Furthermore, the second term in the analytical expression of the image velocity vector requires determining the value of the derivative of the homogeneous position vector of the ground point with respect to time, d(P_ground) / dt. During pushbroom imaging, the sensor's field of view continuously scans the ground as the satellite flies, which is equivalent to the imaged ground target point continuously moving along the scanning direction in the geographic coordinate system. Therefore, d(P_ground) / dt is not zero. The value of this derivative is determined by the latitude and longitude variation rate based on the projection component of the satellite's orbital velocity onto the ground, and by combining it with the local terrain gradient of the digital elevation model to determine the radial component variation rate. Together, these three factors constitute the complete value of the time derivative of the homogeneous position vector of the ground point.

[0042] After aligning the four types of parameters at the same imaging moment, the corresponding variable positions in the total transformation matrix and its time derivative, as well as the corresponding components in the homogeneous position vector of the ground point and its time derivative, are substituted into the matrix and vector multiplication and addition operations. This allows for the calculation of the specific values ​​of the two components of the image velocity vector at the current imaging moment in the image plane coordinate system: the along-track direction and the cross-track direction. These values ​​accurately reflect the combined influence of all motion sources and their cross-coupling effects on the motion state of the imaging point on the focal plane at the current moment, providing a direct numerical basis for subsequent calculations of the yaw angle correction and the charge transfer matching rate of the TDI sensor. During continuous satellite imaging, the above calculations are updated in real-time, frame-by-frame or line-by-line, as the imaging time progresses, to adapt to the continuous evolution of dynamic conditions such as orbital motion, attitude changes, and terrain changes.

[0043] S105, calculate the deflection angle correction and the TDI sensor charge transfer matching rate based on the image motion velocity vector value, and send them to the satellite attitude control system and camera electronics system respectively to perform compensation.

[0044] After obtaining the value of the image displacement velocity vector in the image plane coordinate system at the current imaging moment, it is necessary to calculate two compensation parameters, namely the deflection angle correction and the charge transfer matching rate of the TDI sensor, and send them to the satellite attitude control system and the camera electronics system respectively to perform compensation.

[0045] The image velocity vector in the image plane coordinate system has along-track and cross-track components. The along-track component aligns with the charge transfer direction of the TDI sensor, while the cross-track component is orthogonal to it. Ideally, the image velocity vector should be strictly aligned with the charge transfer direction of the TDI sensor, ensuring complete alignment between the motion direction of the ground image point and the charge integration transfer direction. However, due to Earth's rotation, satellite attitude deviations, perspective effects from side-swing imaging, and cross-coupling between various motion sources, there is a deviation angle between the actual image velocity vector direction and the TDI charge transfer direction, known as the drift angle. Without correction for this drift angle, the ground image point will continuously shift in the cross-track direction during charge integration by the TDI sensor along the track direction, resulting in image blurring in this direction. The drift angle correction is calculated by taking the arctangent of the ratio of the cross-track component to the along-track component of the image velocity vector. This arctangent is the angle by which the image velocity vector direction deviates from the TDI charge transfer direction, which is the drift angle that needs correction.

[0046] While determining the deflection angle correction, it is also necessary to calculate the charge transfer matching rate of the TDI sensor. The TDI sensor achieves long-term integration of moving targets by sequentially transferring and accumulating signal charges between multiple rows of pixels in a fixed time sequence. Its row transfer frequency determines the equivalent movement rate of the charge along the transfer direction on the focal plane. To ensure precise synchronization between the charge transfer rate and the actual movement rate of the ground image point along the track direction on the focal plane, the row transfer frequency needs to be determined based on the track component of the image velocity vector. Specifically, the magnitude of the track component of the image velocity vector is divided by the row spacing of the TDI sensor; the quotient is the row transfer frequency matching the current image velocity, which is the charge transfer matching rate. If there is a deviation between this matching rate and the actual row transfer frequency, the signal will diffuse along the track direction during charge accumulation, causing blurring of the image along the track direction.

[0047] After calculating the yaw angle correction and charge transfer matching rate, the yaw angle correction is sent to the satellite attitude control system as a command. Upon receiving this command, the satellite attitude control system adjusts the yaw attitude angle of the satellite around the optical axis to align the charge transfer direction of the TDI sensor on the camera's focal plane with the actual image motion velocity vector direction, thereby eliminating the image motion component in the trans-track direction. Simultaneously, the charge transfer matching rate is sent to the camera electronics system as a command. The camera electronics system adjusts the line transfer clock frequency of the TDI sensor accordingly to precisely match the magnitude of the charge transfer rate with the magnitude of the image motion velocity vector's along-track component, thereby eliminating image motion in the along-track direction. The yaw angle correction and rate matching work together to eliminate the components of the image motion velocity vector in both the trans-track and along-track directions, respectively, achieving complete compensation for image motion.

[0048] During continuous satellite imaging, as the orbital position, attitude state, and imaging area change continuously, the magnitude and direction of the image displacement velocity vector also evolve. Therefore, the calculation of the yaw angle correction and the charge transfer matching rate, as well as the command transmission, are updated in real time as the imaging progresses, ensuring that the compensation parameters always maintain a dynamic match with the actual image displacement state under the current operating conditions.

[0049] In one possible implementation of the present invention, after image shift compensation is completed and on-orbit imaging data is acquired, it is also necessary to use actual image quality information to perform feedback correction on key parameters in the image shift calculation model in order to eliminate the systematic compensation error introduced by the deviation between the ground calibration value and the on-orbit true value, and realize the online self-calibration of the compensation system.

[0050] On-orbit image data is acquired by the camera's focal plane TDI sensor during normal imaging after image shift compensation. Regions with distinct edge features, such as building outlines, road boundaries, and water-land boundaries with significant gray-level abrupt changes, are selected from the acquired image data as the analysis objects for image quality evaluation. An image edge sharpness index is extracted from the selected regions; this index quantitatively characterizes the sharpness of image edges. The specific calculation method for the image edge sharpness index involves analyzing the gray-level profile of the edge region and evaluating the gray-level gradient magnitude or equivalent width of the edge transition area. A larger gray-level gradient or narrower transition width indicates a sharper edge, meaning less residual image shift; conversely, a smaller gradient or narrower transition width indicates that there is still insufficiently compensated residual image shift causing edge blurring.

[0051] Image edge sharpness is closely related to the accuracy of several parameters in the image motion calculation model, among which the camera's effective focal length and camera mounting angle are the two most significant key parameters. The camera's effective focal length determines the projection scaling factor in the homogeneous coordinate transformation matrix between the camera coordinate system and the image plane coordinate system. Its value directly affects the mapping ratio of the image motion velocity vector from three-dimensional space to the two-dimensional image plane. Even a small deviation in focal length will cause a systematic deviation between the calculated charge transfer matching rate and the actual required value. The camera mounting angle determines the value of the rotated part in the homogeneous coordinate transformation matrix between the satellite body coordinate system and the camera coordinate system. A deviation in the mounting angle will cause the calculated value of the eccentricity angle correction to deviate from the true value, making it impossible to completely eliminate the image motion component in the transorbital direction. Although these two parameters are calibrated during the ground assembly and adjustment phase, factors such as vibration and shock during satellite launch, structural micro-deformation caused by the on-orbit thermal alternation environment, and material creep during long-term operation can all cause the actual values ​​of the effective focal length and mounting angle to deviate from the ground calibration values. Furthermore, this deviation has a gradual change characteristic and is difficult to permanently eliminate through a single calibration.

[0052] To achieve online correction of the aforementioned key parameters, maximizing the image edge sharpness index is used as the optimization objective. An iterative optimization method is employed to update the values ​​of the camera's effective focal length and camera mounting angle. Specifically, a small parameter perturbation is applied to the current parameter values. The perturbed parameter values ​​are then re-substituted into the image motion velocity vector analytical expression to calculate the compensation amount. The changes in the edge sharpness index of the images acquired before and after the perturbation are compared to determine the direction and magnitude of the parameter adjustment. The parameter values ​​are gradually updated along the direction that increases the image edge sharpness index. After multiple iterations, the values ​​converge to the parameter values ​​that maximize the edge sharpness index. This value represents the optimal estimate of the camera's effective focal length and mounting angle under the current on-orbit conditions. The corrected parameter values ​​are updated to the corresponding homogeneous coordinate transformation matrix in the image shift calculation module, replacing the original ground calibration values. Subsequent image shift compensation calculations are then based on the corrected parameters, thereby eliminating systematic compensation errors introduced by parameter drift. The iterative optimization process is set to convergence termination conditions as follows: the change in the effective focal length of the camera between two adjacent iterations is less than a preset focal length threshold and the change in the camera mounting angle is less than a preset angle threshold, or the change in the image edge sharpness index between two adjacent iterations is less than a preset sharpness threshold. If either condition is met, the iteration terminates and the current parameter value is output as the correction result. The above online correction process can be periodically triggered after the satellite completes a preset number of imaging tasks, or it can be automatically triggered when the image edge sharpness index drops below a preset warning value. This ensures that the image shift compensation system maintains high-precision compensation performance throughout its on-orbit lifespan without relying on manual intervention or recalibration at the ground station.

[0053] Please see Figure 2 The second embodiment of the present invention provides a high-precision generalized dynamic image shift compensation device, comprising: The coordinate transformation matrix establishment module 201 is used to establish homogeneous coordinate transformation matrices between the geographic coordinate system and the ground-fixed coordinate system, between the ground-fixed coordinate system and the geocentric inertial coordinate system, between the geocentric inertial coordinate system and the satellite orbit coordinate system, between the satellite orbit coordinate system and the satellite body coordinate system, between the satellite body coordinate system and the camera coordinate system, and between the camera coordinate system and the image plane coordinate system, to obtain six homogeneous coordinate transformation matrices. Image plane point transformation module 202 is used to multiply the six homogeneous coordinate transformation matrices in the order of transformation to obtain the total transformation matrix, and use the total transformation matrix to transform the homogeneous position vector of the ground point to the image plane coordinate system to obtain the homogeneous position vector of the image plane point. Image velocity calculation module 203 is used to differentiate the homogeneous position vector of the image point with respect to time to obtain an analytical expression for the image velocity vector that includes each motion source and its cross-coupling effect; The real-time calculation module 204 is used to substitute the real-time acquired satellite orbit parameters, attitude parameters, time information and digital elevation model data into the analytical expression to calculate the image motion velocity vector value at the current imaging moment; The compensation execution module 205 is used to calculate the deflection angle correction and the TDI sensor charge transfer matching rate based on the image motion velocity vector value, and send them to the satellite attitude control system and camera electronics system respectively to perform compensation.

[0054] The third embodiment of the present invention provides a high-precision generalized dynamic image shift compensation device, including a memory and a processor. The memory stores a computer program, which can be executed by the processor to implement a high-precision generalized dynamic image shift compensation method as described in any of the above embodiments.

[0055] The fourth embodiment of the present invention provides a computer-readable storage medium, characterized in that it stores a computer program, which can be executed by the processor of the device in which the computer-readable storage medium is located, to implement a high-precision generalized dynamic image shift compensation method as described in any of the above claims.

[0056] Based on the high-precision generalized dynamic image shift compensation method, device, equipment, and storage medium provided by this invention, six homogeneous coordinate transformation matrices between the geographic coordinate system and the image plane coordinate system are multiplied sequentially to form a total transformation matrix, which completely expresses the multi-level serial kinematic mapping from ground points to image plane points within a single mathematical framework. The homogeneous position vector of the image plane points obtained by the total transformation matrix is ​​directly differentiated with respect to time, so that each motion source and its cross-coupling effect are naturally included analytically in the image shift velocity vector expression, avoiding the truncation error caused by component separation and superposition. Substituting real-time parameters into this expression yields accurate image shift velocity vector values, which are then used to calculate the yaw angle correction and TDI charge transfer matching rate to perform compensation.

[0057] Exemplary examples show that the computer program described in the third and fourth embodiments of the present invention can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in implementing a high-precision generalized dynamic image shift compensation device. For example, the apparatus described in the second embodiment of the present invention.

[0058] The processor referred to can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the high-precision generalized dynamic image shift compensation method, connecting various parts of the method through various interfaces and lines.

[0059] The memory can be used to store the computer program and / or modules. The processor implements various functions of a high-precision generalized dynamic image shift compensation method by running or executing the computer program and / or modules stored in the memory, and by calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, text conversion function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, text message data, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0060] If the implemented module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0061] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0062] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A high-precision generalized dynamic image shift compensation method, characterized in that, include: Homogeneous coordinate transformation matrices were established between the geographic coordinate system and the Earth-fixed coordinate system, between the Earth-fixed coordinate system and the geocentric inertial coordinate system, between the geocentric inertial coordinate system and the satellite orbit coordinate system, between the satellite orbit coordinate system and the satellite body coordinate system, between the satellite body coordinate system and the camera coordinate system, and between the camera coordinate system and the image plane coordinate system, resulting in six homogeneous coordinate transformation matrices. The six homogeneous coordinate transformation matrices are multiplied sequentially in the order of transformation to obtain the total transformation matrix. The homogeneous position vectors of the ground points are then transformed to the image plane coordinate system using the total transformation matrix to obtain the homogeneous position vectors of the image plane points. Differentiating the homogeneous position vector of the image point with respect to time yields an analytical expression for the image velocity vector, which includes the effects of each motion source and their cross-coupling. Substitute the real-time acquired satellite orbit parameters, attitude parameters, time information, and digital elevation model data into the analytical expression to calculate the image motion velocity vector value at the current imaging moment; The deflection angle correction and the charge transfer matching rate of the TDI sensor are calculated based on the image velocity vector value and sent to the satellite attitude control system and camera electronics system respectively to perform compensation.

2. The high-precision generalized dynamic image shift compensation method according to claim 1, characterized in that, The analytical expression for the image displacement velocity vector is: V_image = d(T_total) / dt · P_ground + T_total · d(P_ground) / dt Where T_total is the total transformation matrix, and P_ground is the homogeneous position vector of the ground point in the geographic coordinate system.

3. The high-precision generalized dynamic image shift compensation method according to claim 1, characterized in that, The homogeneous coordinate transformation matrix between the Earth-fixed coordinate system and the geocentric inertial coordinate system includes the time-varying parameter of the Earth's rotation angular velocity. The homogeneous coordinate transformation matrix between the satellite orbit coordinate system and the satellite body coordinate system is determined by the satellite roll angle, pitch angle, and yaw angle. The homogeneous coordinate transformation matrix between the satellite body coordinate system and the camera coordinate system is determined by the camera mounting angle.

4. The high-precision generalized dynamic image shift compensation method according to claim 1, characterized in that, The digital elevation model data is used to correct the radial component of the homogeneous position vector of ground points to reflect the actual altitude of the ground points.

5. The high-precision generalized dynamic image shift compensation method according to claim 1, characterized in that, The deflection angle correction is the arctangent of the ratio of the cross-track component to the along-track component of the image displacement velocity vector value, and the charge transfer matching rate is the ratio of the magnitude of the along-track component to the row spacing of the TDI sensor.

6. The high-precision generalized dynamic image shift compensation method according to claim 1, characterized in that, Also includes: Acquire on-orbit image data and extract image edge sharpness indicators. Based on the indicators, perform online correction on the key parameters of the homogeneous coordinate transformation matrix. The key parameters include the camera's effective focal length and camera mounting angle.

7. A high-precision generalized dynamic image shift compensation method according to claim 6, characterized in that, The online correction aims to maximize the image edge sharpness index and uses an iterative optimization method to update the key parameters.

8. A high-precision generalized dynamic image shift compensation device, characterized in that, include: The coordinate transformation matrix establishment module is used to establish homogeneous coordinate transformation matrices between the geographic coordinate system and the Earth-fixed coordinate system, between the Earth-fixed coordinate system and the geocentric inertial coordinate system, between the geocentric inertial coordinate system and the satellite orbit coordinate system, between the satellite orbit coordinate system and the satellite body coordinate system, between the satellite body coordinate system and the camera coordinate system, and between the camera coordinate system and the image plane coordinate system, resulting in six homogeneous coordinate transformation matrices. The image plane point transformation module is used to multiply the six homogeneous coordinate transformation matrices in the order of transformation to obtain the total transformation matrix, and use the total transformation matrix to transform the homogeneous position vector of the ground point to the image plane coordinate system to obtain the homogeneous position vector of the image plane point. The image velocity calculation module is used to differentiate the homogeneous position vector of the image point with respect to time to obtain an analytical expression for the image velocity vector that includes each motion source and its cross-coupling effect; The real-time calculation module is used to substitute the real-time acquired satellite orbit parameters, attitude parameters, time information and digital elevation model data into the analytical expression to calculate the image motion velocity vector value at the current imaging moment; The compensation execution module is used to calculate the deflection angle correction and the charge transfer matching rate of the TDI sensor based on the image displacement velocity vector value, and send them to the satellite attitude control system and camera electronics system respectively to perform compensation.

9. A high-precision generalized dynamic image shift compensation device, characterized in that, The system includes a memory and a processor. The memory stores a computer program that can be executed by the processor to implement a high-precision generalized dynamic image shift compensation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The device contains a computer program that can be executed by a processor of the device in which the computer-readable storage medium is located, to implement a high-precision generalized dynamic image shift compensation method as described in any one of claims 1 to 7.

Images