Acquisition method and device for satellite data geometric positioning

Abstract

The invention provides an acquisition method and device for geometric positioning of satellite data, and the method comprises the steps: obtaining grid point pairs in an image pixel plane and a target space, and constructing a target rational polynomial model according to the grid point pairs; performing coefficient fitting of the target rational polynomial model by using a least square method to obtain a coefficient matrix, and if the coefficient matrix meets a preset condition, obtaining a target coefficient; if the coefficient matrix does not meet the preset condition, regularization is carried out on the coefficient matrix to obtain a target coefficient, and the target coefficient maps the ground coordinates to the pixel coordinates. A target rational polynomial model is constructed according to the grid point pairs, and the constructed target rational polynomial model can accurately export the three-dimensional-two-dimensional point pair relation in geometric positioning into a terrain-independent algorithm of the RPC model. The performance of the method is evaluated by changing the point pair relation and the size of the coverage area of the point pair relation based on the least square method, so that the point-surface position relation is accurately calculated according to limited information.

Description

Technical Field

[0001] This application relates to the field of satellite remote sensing technology, and particularly to a method and device for obtaining satellite data geometric positioning. Background Art

[0002] Satellite remote sensing applications rely on a series of tools, including geometric positioning. Geometric positioning includes the forward process of associating 3D world coordinates with 2D image coordinates, which is represented by the projection function as P:R 3 →R 2 , that is, the process of mapping 3D points to the image plane. Geometric positioning also includes the reverse process represented by the positioning function as P:R 3 →R 2 . When all physical components involved in the above geometric positioning process are known, geometric positioning is defined as a chained operation process that models these component factors, and the result of this series of process modeling is called a physical (rigorous) geometric model.

[0003] Traditional geometric modeling methods are complex. It is necessary to capture each row of optical images at different times, and then indirectly obtain the physical components in the geometric positioning process. Not only is the simulation of the geometric model cumbersome, but also the solution of the geometric model is complex and the accuracy is poor.

[0004] Therefore, a new acquisition scheme for satellite data geometric positioning is needed. Summary of the Invention

[0005] In view of this, embodiments of this specification provide a method and device for obtaining satellite data geometric positioning.

[0006] Embodiments of this specification provide the following technical solutions:

[0007] Embodiments of this specification provide a method for obtaining satellite data geometric positioning, including:

[0008] Obtaining pairs of grid points in the image pixel plane and the target space, and constructing a target rational polynomial model according to the pairs of grid points;

[0009] Using the least squares method to perform coefficient fitting on the target rational polynomial model to obtain a coefficient matrix. If the coefficient matrix meets the preset conditions, the target coefficients are obtained; wherein, the coefficient matrix is used to determine the mapping between the pairs of grid points and the size of the covered space region;

[0010] If the coefficient matrix does not meet the preset conditions, the coefficient matrix is regularized to obtain the target coefficients, and the target coefficients map ground coordinates to pixel coordinates.

[0011] Embodiments of this specification also provide a device for obtaining satellite data geometric positioning, including:

[0012] The first acquisition module is configured to acquire grid point pairs in the image pixel plane and the target space, and construct a target rational polynomial model according to the grid point pairs;

[0013] The second acquisition module is configured to perform coefficient fitting on the target rational polynomial model using the least squares method to obtain a coefficient matrix, and if the coefficient matrix meets a preset condition, obtain target coefficients; wherein, the coefficient matrix is used to determine the mapping between the grid point pairs and the size of the covered space region;

[0014] The third acquisition module is configured to, if the coefficient matrix does not meet the preset condition, perform regularization on the coefficient matrix to obtain target coefficients, and the target coefficients map the ground coordinates to pixel coordinates.

[0015] Compared with the prior art, the beneficial effects that can be achieved by at least one of the above technical solutions adopted in the embodiments of this specification at least include:

[0016] This application provides a new acquisition solution for satellite data geometric positioning. Grid point pairs in the image pixel plane and the target solution space are acquired, and a target rational polynomial model is constructed according to the grid point pairs. The constructed target rational polynomial model can accurately derive a terrain-independent algorithm of the RPC model from the three-dimensional to two-dimensional point pair relationship in geometric positioning. Based on the least squares method, the performance of this method is evaluated by changing the point pair relationship and the size of their covered regions, so as to ensure that the point-plane position relationship can be accurately calculated according to the limited information provided by the image provider without relying on the specific terrain, which has important significance in the field of remote sensing image applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of this application, the drawings required to be used in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of this application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic diagram of a method for solving rational polynomial coefficients for satellite geometric positioning provided by the embodiments of this specification;

[0019] Figure 2 It is a schematic diagram of satellite geometric positioning provided by the embodiments of this specification;

[0020] Figure 3 It is a flowchart of a method for solving rational polynomial coefficients for satellite geometric positioning provided by the embodiments of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The embodiments of this application will be described in detail below with reference to the drawings.

[0022] The following describes the implementation manners of the present application through specific specific examples. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The present application can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope of protection of the present application.

[0023] It should be noted that the following describes various aspects of the embodiments within the scope of the appended claims. It should be obvious that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is illustrative only. Based on the present application, those skilled in the art should understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects described herein can be used to implement the device and / or practice the method. In addition, this device can be implemented and this method can be practiced using other structures and / or functions in addition to one or more of the aspects described herein.

[0024] It should also be noted that the drawings provided in the following embodiments only illustrate the basic concept of the present application schematically. Only the components related to the present application are shown in the drawings, rather than being drawn according to the number, shape, and size of the components in actual implementation. The types, quantities, and proportions of the components in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0025] In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the examples can be practiced without these specific details.

[0026] Satellite remote sensing applications require a series of tools, including geometric positioning. Geometric positioning includes a forward process and a reverse process. When all the physical components involved in the above geometric positioning process are known, geometric positioning can be defined as a chained operation process of modeling these component factors. The result of modeling this series of processes is called a physical (rigorous) geometric model.

[0027] However, the construction and solution process of traditional physical geometric models are complex. For example, in a common optical image acquisition system, the "push-broom imaging scanner", it generally includes a single-row pixel sensor mounted on a satellite platform to capture each row of optical images at different times. Therefore, the exterior orientation elements of the camera, such as the perspective center and attitude angles, need to be changed with each row. While the internal parameters of the camera (including pixel size, focal length, lens distortion, etc.) remain fixed during the imaging process, physical-related designs suitable for the camera sensor are required, increasing the complexity.

[0028] Another example is in synthetic aperture radar (SAR) imaging, where the commonly used physical geometric model is the Range-Doppler model. The SAR satellite emits an electromagnetic wave, which then reflects off the ground and is finally received by the SAR sensor. Similar to the push-broom imaging mode of optical satellites, SAR imaging is also an electromagnetic wave beam imaging row by row, increasing the complexity. Additionally, the position of the ground block in the SAR image is related to the distance from the sensor, that is, the "range" quantity. The Range-Doppler model is constructed based on ephemeris data (time, position, and velocity along the orbit) and the acquired time information, also increasing the complexity.

[0029] To avoid the above complex processes such as the construction of physical geometric models, the inventors found that data providers offer a general Rational Polynomial Camera (RPC) model to downstream users to fit the cumbersome physical geometric models, especially to accurately derive the three-dimensional to two-dimensional point pair relationship and obtain an RPC model independent of the specific terrain.

[0030] Based on this, the embodiments of this specification propose a new solution for solving the rational polynomial system for satellite data geometric positioning: for the rational polynomial camera model, a terrain-independent solution algorithm is proposed. Without relying on the specific terrain conditions, it can accurately calculate the point-plane position relationship based on the limited information provided by the image provider, realizing fast and robust remote sensing geometric positioning for optical push-broom imaging and SAR imaging, which is of great significance in the field of remote sensing image applications.

[0031] The following will describe the technical solutions provided by the embodiments of this application with reference to the accompanying drawings.

[0032] This application proposes a method for solving the rational polynomial coefficients for satellite data geometric positioning. For optical push-broom imaging satellites and SAR satellites, a method for constructing a rational polynomial camera model independent of terrain is developed, which is evaluated by changing the point correspondence relationship and the size of their covered areas, and provides assistance in tasks such as geometric positioning, orthorectification, and image mosaicing.

[0033] Obtain the grid point pairs in the image pixel plane and the target solution space, and construct a target rational polynomial model based on these grid point pairs. The constructed target rational polynomial model can accurately derive the terrain-independent algorithm of the RPC model from the three-dimensional to two-dimensional point pair relationship in geometric positioning. Based on the least squares method, evaluate the performance of this method by changing the point pair relationship and the size of their coverage area, so as to ensure that the point-plane position relationship can be accurately calculated according to the limited information provided by the image provider without relying on the specific terrain.

[0034] The input of this method is a regular grid of 3D points (China Navigation Positioning System, abbreviated as CNP, a preset navigation positioning system) with multiple height layers, and the 2D positions of points on the image plane. It mainly includes the following steps: constructing the grid point pairs between the image plane and the three-dimensional space, constructing the mathematical representation of the rational polynomial camera observation model, constructing the error observation equation obtained under N point pairs, iteratively optimizing the design matrix and the coefficient matrix, etc. It can ensure that the point-plane position relationship can be accurately calculated according to the limited information provided by the image provider without relying on the specific terrain, which is of great significance in the field of remote sensing image applications.

[0035] As Figure 1 and Figure 3 shown, a method for obtaining geometric positioning for satellite data provided in an embodiment of this specification includes steps S301 - S303. Specifically, in step S301, obtain the grid point pairs in the image pixel plane and the target space, and construct a target rational polynomial model based on the grid point pairs. In step S302, use the least squares method to perform coefficient fitting on the target rational polynomial model to obtain a coefficient matrix. If the coefficient matrix meets the preset conditions, obtain the target coefficients; wherein, the coefficient matrix is used to determine the mapping between the grid point pairs and the size of their coverage space region. In step S303, if the coefficient matrix does not meet the preset conditions, use regularization on the coefficient matrix to obtain the target coefficients, and the target coefficients map the ground coordinates to the pixel coordinates.

[0036] When solving the rational polynomial camera model in geometric positioning, the algorithm requires a set of grid point pairs in the image pixel plane and the target space as input to realize the fitting solution of the rational polynomial coefficients.

[0037] Specifically, obtain the grid point pairs in the image pixel plane and the target space, where determine the image pixel grid: the image grid contains m×n image pixel points, and these points are evenly distributed throughout the image. Optionally, both the row m and the column n are set to 10.

[0038] Among them, a three-dimensional target grid on the ground space is determined: The size of this three-dimensional grid is based on the entire range of the image and the range of the estimated terrain undulation, that is, the size of the grid covers the range of the three-dimensional terrain surface. Such a three-dimensional grid generally contains multiple different height levels. To effectively avoid the problem of the coefficient matrix being ill-conditioned during solution, the number of height levels of the three-dimensional grid is generally greater than 3 levels.

[0039] Construct a target rational polynomial model according to the grid point pairs. Combining the above-obtained grid point pairs, represent the two-dimensional image coordinates, that is, the pixel coordinates (r, c) in the image plane, as the proportional relationship of the polynomial constructed by the three-dimensional space coordinates (i.e., the ground coordinates (X, Y, Z)). The mathematical formula of the RPC model is expressed as:

[0040]

[0041] Among them, r n and c n are the normalized representations of the row and column coefficients in the pixel plane respectively, and X n , Y n and Z n are the normalized representations of the coordinate values of the target points in the ground space.

[0042] Among them, in order to minimize the introduced error during the calculation process and improve the numerical stability of the equation, it is necessary to perform a normalization operation on the pixel coordinate values and the space coordinate values. The formula is expressed as:

[0043]

[0044] Among them, r o and c o are the offset values of two pixel coordinate terms, and r s and c s are the scale values of two pixel coordinate terms. Similarly, X o , Y o and Z o are the offset values of three ground coordinate terms, and X s , Y s and Z s are the scale values of three ground coordinate terms.

[0045] Based on the above-constructed target rational polynomial model, according to the actual satellite application development experience, the number of coefficients of the polynomial p in the numerator and denominator in Formulas 1.1 and 1.2 is set to 20, and the specific representation is as follows:

[0046] p(X, Y, Z) = p 0 +p 1 Z + p 2 Y + p 3 X + p 4 ZY + p5 ZX + p 6 YX + p 7 X 2 + p 8 Y 2 + p 9 Z 2 + p 10 ZYX + p 11 Z 2 Y + p 12 Z 2 X + p 13 Y 2 Z + p 14 Y 2 X + p 15 ZX 2 + p 16 YX 2 + p 17 Z 3 + p 18 Y 3 + p 19 X 3 (1.5)

[0047] Among them, p i is the coefficient of polynomial p. In particular, for the first coefficient p 0 of the denominator polynomial in Formulas 1.1 and 1.2, it is set to 1. Therefore, a total of 78 coefficients need to be solved.

[0048] Among them, the number of polynomial coefficients of the numerator and denominator is set to 20, that is, each polynomial (numerator and denominator) has 20 coefficients. In the RPC model, the first coefficient (constant term) of the denominator polynomial is usually set to 1 to ensure the uniqueness and stability of the model. Therefore, the denominator polynomial actually has only 19 independent coefficients to be solved. Therefore, the total number of coefficients is 20 + 19 = 39 coefficients.

[0049] In the RPC model, the pixel coordinates have two components: row coordinates and column coordinates. Each component requires an independent RPC model to describe. Therefore, for the row coordinates and column coordinates, each requires 39 coefficients. Therefore, the total number of coefficients is: 39 × 2 = 78 coefficients.

[0050] Furthermore, for the mathematical representation of the constructed RPC model (i.e., the target rational polynomial model), the least squares solution method is used to achieve the coefficient fitting of the RPC model.

[0051] Specifically, using N control points, Formulas 1.1 and 1.2 can be rewritten as an equation system represented in matrix form:

[0052] WTI - WG = 0 (1.6)

[0053] Wherein:

[0054]

[0055] p 2 (V i ) = p 2 (X i , Y i , Z i ) (1.7)

[0056] T = block_diag[M r , M c , M r , M c ∈R N×39 (1.8)

[0057]

[0058] G = [r 0 , …, r N , c 0 , …, c N T (1.10)

[0059] Furthermore, the coefficient matrix I is directly solved from Equation 1.6 by the least squares method, and the mathematical representation is:

[0060] T T W 2 TI - T T W 2 G = 0 (1.11)

[0061] Wherein, the control points are a set of known three - dimensional ground coordinates (X, Y, Z) and the corresponding two - dimensional image pixel coordinates (r, c). These points are used to establish the coefficients of the RPC model. It should be noted that the number N of control points usually needs to be large enough to ensure the accuracy and stability of the model.

[0062] Specifically, by constructing the observation error equations and combining these equations into an error system in matrix form, the solution of the RPC model coefficients is finally realized.

[0063] The observation error equations are constructed as follows:

[0064] According to Equations 1.1 and 1.2, the processes of solving the coefficients of row r and column c are the same. Here, we take the construction of the observation error equation for row r as an example.

[0065] First, Equations 1.1 and 1.5 are adjusted, and the mathematical representation is as follows:

[0066] ​

[0067] Therefore, the observation error equation for Formula 1.25 can be expressed as:

[0068]

[0069] Where:

[0070]

[0071] Inputting the set of grid point pairs determined in the first step into the above observation error equation 1.25 will result in a system of observation error equations.

[0072] If the obtained coefficient matrix is positive definite, the target coefficients are obtained. The target coefficients map the three-dimensional space coordinates (i.e., ground coordinates) to two-dimensional image coordinates (pixel coordinates). Among them, the coefficient matrix is used to determine the mapping between the grid point pairs and the size of the covered spatial region. In the embodiments of this specification, the preset conditions include determining whether the coefficient matrix is positive definite.

[0073] If the coefficient matrix is not positive definite, regularization is performed on the coefficient matrix to obtain the target coefficients, and the target coefficients map the ground coordinates to the pixel coordinates.

[0074] In the process of solving the coefficients of the RPC model, the matrix to be solved is in the form of ATA. Where: A is the design matrix, which is composed of the coordinate values of the control points. ATA is a square matrix used for solving the coefficients by the least squares method. If the matrix ATA is full rank, that is, the rank is equal to the dimension of the matrix. The number of control points is large enough and evenly distributed in the three-dimensional space. The polynomial order used in the RPC model is relatively low (such as 2nd order or lower). Then the preset conditions are satisfied. If the above conditions are not met, it is not positive definite, and regularization operations need to be performed to enhance numerical stability.

[0075] Since a high-order polynomial (greater than 2nd order) is used in the RPC model, the coefficient solution matrix in the above steps is not positive definite. Therefore, the present invention further uses a regularization method for the coefficient solution matrix. Specifically, in order to increase the numerical stability of the coefficient I solution, a biased estimation regression method (such as ridge estimation) is usually used for regularization operations, and the mathematical expression is:

[0076] (T T W 2 T+h 2 E)I-T T W 2 G = 0 (1.12)

[0077] Where, E is the identity matrix, and h is the regularization control scalar.

[0078] In order to solve the final coefficient, an iterative solution method can usually be used. In this way, it is necessary to select the initial values ​​of W and h, and calculate the optimal coefficient result according to the set tolerance error threshold.

[0079] The specific detailed iterative solution process is as follows:

[0080] For formula 1.12, we set the initial value of W to the unit matrix E, that is, W (0) = E. Then use ridge estimation to select the optimal h value and the corresponding I (0) initialization.

[0081] In the process of selecting the optimal regularization scalar h, we use the L curve indicator to select. Specifically, we calculate the functional relationship between the logarithm of I corresponding to the regularization scalar h and the logarithm of the error term. The former is expressed as log||I|| h , the latter is expressed as log||WTI-WG|| h . When such a functional relationship is plotted, it takes the shape of L, so it is called the L-function indicator.

[0082] Then, for the case where the number of iterations i ≥ 1, W (i) byI (i-1) OK, then W (i) Substitute into formula 1.12 and further solve I (i) , until the error threshold we set is met, then the iteration is stopped and the I value that meets the conditions is obtained. The calculation of the kth iteration of formula 1.12 can be expressed as:

[0083] (T T (W (k) ) 2 T+E)I (k) =T T (W (k) ) 2 G+I (k-1) (1.29)

[0084] The embodiment of this specification also provides a device for obtaining satellite data geometric positioning, the device comprising:

[0085] A first acquisition module, used for acquiring grid point pairs in an image pixel plane and a target space, and constructing a target rational polynomial model according to the grid point pairs;

[0086] A second acquisition module is used to use the least squares method to fit the coefficients of the target rational polynomial model to obtain a coefficient matrix, and if the coefficient matrix meets the preset conditions, the target coefficients are obtained; wherein the coefficient matrix is ​​used to determine the mapping between the grid point pairs and the size of the spatial area covered by them;

[0087] A third acquisition module, configured to, if the coefficient matrix does not meet a preset condition, perform regularization on the coefficient matrix to obtain a target coefficient, and the target coefficient maps ground coordinates to pixel coordinates.

[0088] Correspondingly, the apparatus can be used to execute Figure 3 the steps in the method embodiments shown. The implementation principles and technical effects are similar and will not be elaborated here.

[0089] An electronic device provided in an embodiment of this specification, the electronic device includes: a processor, a memory, and a computer program; where

[0090] The memory is used to store the computer program, and the memory can also be a flash memory. The computer program is, for example, an application program or a functional module that implements the above method.

[0091] The processor is configured to execute the computer program stored in the memory to implement each step executed by the device in the above method. For details, reference can be made to the relevant descriptions in the foregoing method embodiments.

[0092] Optionally, the memory can be either independent or integrated with the processor.

[0093] When the memory is a device independent of the processor, the device may further include:

[0094] A bus, configured to connect the memory and the processor.

[0095] This application also provides a readable storage medium, in which a computer program is stored, and the computer program, when executed by a processor, is used to implement the methods provided by the above various embodiments.

[0096] Among them, the readable storage medium can be a computer storage medium or a communication medium. The communication medium includes any medium that facilitates transmitting a computer program from one place to another. The computer storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. For example, the readable storage medium is coupled to the processor, so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an Application Specific Integrated Circuits (ASIC). Additionally, the ASIC can be located in a user device. Of course, the processor and the readable storage medium can also exist as discrete components in a communication device. The readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0097] For the same or similar parts among the various embodiments in this specification, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the product embodiments described later, since they correspond to the methods, the descriptions are relatively simple, and the relevant parts can be referred to the descriptions in the system embodiments.

[0098] As described above, the above are only specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for obtaining geometric positioning of satellite data, characterized in that: include: Obtaining grid point pairs in the image pixel plane and the target space, and constructing a target rational polynomial model based on the grid point pairs; The coefficients of the target rational polynomial model are fitted using the least squares method to obtain a coefficient matrix. If the coefficient matrix satisfies a preset condition, the target coefficients are obtained; wherein the coefficient matrix is ​​used to determine the mapping between the grid point pairs and the size of the spatial area covered by them; If the coefficient matrix does not meet the preset conditions, regularization is applied to the coefficient matrix to obtain target coefficients, where the target coefficients map the ground coordinates to the pixel coordinates.

2. The method for obtaining satellite data geometric positioning according to claim 1, characterized in that: Constructing a target rational polynomial model according to the grid point pairs, including: The pixel coordinates (r, c) in the image plane are expressed as the ground coordinates (X, Y, Z) to construct a polynomial proportional relationship, and the target rational polynomial model is obtained as follows: Among them, r n and c n are the normalized representations of row and column coefficients in the pixel plane, respectively, and X n , Y n and Z n It is the normalized representation of the coordinate value of the target point in the ground space.

3. The method for obtaining satellite data geometric positioning according to claim 2, characterized in that: Also includes: Normalize the pixel coordinate values ​​and space coordinate values. The pixel coordinates and space coordinates are expressed as Among them, r o and c o is the offset value of the two pixel coordinate items, r s and c s is the scale value of the two pixel coordinate items; X o , Y o and Z o is the offset value of the three ground coordinate items, X s , Y s and Z s are the scale values ​​of the three ground coordinate items.

4. The method for obtaining satellite data geometric positioning according to claim 1, characterized in that: The coefficient matrix is ​​obtained by fitting the coefficients of the target rational polynomial model using the least squares method, including: Using N control points, equations 1.1 and 1.2 can be rewritten as a system of equations in matrix form: WTI-WG=0(1.6) in: p2(V i )=p2(X i ,Y i ,Z i )(1.7) T=block_diag[M r ,M c ],M r ,M c ∈R N×39 (1.8) G=[r0,…,r N ,c0,…,c N ] T (1.10) Among them, formula 1.6 is directly solved by the least squares method to obtain the coefficient matrix I, which is mathematically expressed as: T T W 2 TI-T T W 2 G=0(1.11)。 5. The method for obtaining satellite data geometric positioning according to claim 1, characterized in that: If the coefficient matrix does not meet the preset conditions, the coefficient matrix is ​​regularized to obtain the target coefficients, including: Use biased estimation regression method to perform regularization operation and obtain mathematical representation (T T W 2 T+h 2 E)I-T T W 2 G=0(1.12) Where E is the identity matrix and h is the regularized control scalar; The iterative solution method is adopted to select the initial values ​​of W and h, and the optimal coefficient result is calculated according to the set tolerance error threshold.

6. A device for acquiring geometric positioning of satellite data, characterized in that: include: A first acquisition module, used for acquiring grid point pairs in an image pixel plane and a target space, and constructing a target rational polynomial model according to the grid point pairs; A second acquisition module is used to use the least squares method to fit the coefficients of the target rational polynomial model to obtain a coefficient matrix, and if the coefficient matrix meets the preset conditions, the target coefficients are obtained; wherein the coefficient matrix is ​​used to determine the mapping between the grid point pairs and the size of the spatial area covered by them; The third acquisition module is used to obtain target coefficients by regularizing the coefficient matrix if the coefficient matrix does not meet the preset conditions, and the target coefficients map the ground coordinates to the pixel coordinates.

7. The device for acquiring satellite data geometric positioning according to claim 1, characterized in that: Constructing a target rational polynomial model according to the grid point pairs, including: The pixel coordinates (r, c) in the image plane are expressed as the ground coordinates (X, Y, Z) to construct a polynomial proportional relationship, and the target rational polynomial model is obtained as follows: Among them, r n and c n are the normalized representations of row and column coefficients in the pixel plane, respectively, and X n , Y n and Z n It is the normalized representation of the coordinate value of the target point in the ground space.

8. The device for acquiring satellite data geometric positioning according to claim 7, characterized in that: Also includes: Normalize the pixel coordinate values ​​and space coordinate values. The pixel coordinates and space coordinates are expressed as Among them, r o and c o is the offset value of the two pixel coordinate items, r s and c s is the scale value of the two pixel coordinate items; X o , Y o and Z o is the offset value of the three ground coordinate items, X s , Y s and Z s are the scale values ​​of the three ground coordinate items.

9. The device for acquiring satellite data geometric positioning according to claim 6, characterized in that: The coefficient matrix is ​​obtained by fitting the coefficients of the target rational polynomial model using the least squares method, including: Using N control points, equations 1.1 and 1.2 can be rewritten as a system of equations in matrix form: WTI-WG=0(1.6) in: p2(V i )=p2(X i ,Y i ,Z i )(1.7) T=block_diag[M r ,M c ],M r ,M c ∈R N×39 (1.8) G=[r0,…,r N ,c0,…,c N ] T (1.10) Among them, formula 1.6 is directly solved by the least squares method to obtain the coefficient matrix I, which is mathematically expressed as: T T W 2 TI-T T W 2 G=0(1.11)。 10. The device for acquiring satellite data geometric positioning according to claim 6, characterized in that: If the coefficient matrix does not meet the preset conditions, the coefficient matrix is ​​regularized to obtain the target coefficients, including: Use biased estimation regression method to perform regularization operation and obtain mathematical representation (T T W 2 T+h 2 E)I-T T W 2 G=0(1.12) Where E is the identity matrix and h is the regularized control scalar; The iterative solution method is adopted to select the initial values ​​of W and h, and the optimal coefficient result is calculated according to the set tolerance error threshold.

Images