CN114693783A - Satellite autonomous pose determination method, system and storage medium - Google Patents

Abstract

The invention provides a method for judging the autonomous pose of a satellite, which comprises the steps of transmitting a remote sensing image shot by an optical camera and a template image retrieved from a knowledge graph spectrum to a deep convolution network to obtain matching characteristic information; and according to the matching characteristic information, substituting the matching characteristic information into the determined optical camera imaging model, the determined motion equation and the determined observation equation to solve so as to obtain the satellite attitude state. The method can effectively improve the efficiency and accuracy of satellite attitude and orbit estimation, thereby realizing the real-time attitude judgment of the satellite off-line state and improving the stability and reliability of the satellite operation control system.

Description

Satellite autonomous pose determination method, system and storage medium

technical field

The invention relates to the field of satellite control, in particular to a method, a system and a storage medium for determining an autonomous positioning and attitude of a satellite based on a visual SLAM (Simultaneous Localization and Mapping) technology.

Background technique

For a long time, space has been a holy place for human exploration and yearning. Since the beginning of the 21st century, mankind has made remarkable achievements in science and technology, political society, and financial economy, which has provided a solid foundation for mankind to explore the universe and move towards space. At the same time, the advancement of high-tech, represented by aerospace technology, provides strong support for a country's economic construction and prosperous development. Therefore, all countries in the world have placed the development of aerospace industry technology in an extremely important position.

Satellite navigation is a key link in the operation of spacecraft. With the continuous development of satellite navigation technology, a lot of research has been carried out on the problem of autonomous orbit determination of satellites. The autonomous orbit determination method, with the help of global, multi-observation data and low-cost on-board GPS measurements, processes on-orbit GPS observation data in real time, obtains high-precision satellite orbit parameters, and realizes autonomous real-time orbit determination of low-orbit satellites; The autonomous orbit determination method of the star sensor uses the star sensor to retrieve the background star data, and calculates the position and attitude of the satellite with the help of related algorithms, so as to realize the autonomous orbit determination; the autonomous orbit determination method based on the magnetometer uses the satellite Sensors, earth sensors and magnetometers are used as measurement sensors, and relevant optimization algorithms are used for state estimation to realize autonomous orbit determination of satellites. With the rapid development of science and technology in the space field and the rapid increase in the number of space vehicles, the above methods are gradually unable to meet the performance requirements of accuracy, intelligence, autonomy, stability and reliability.

SUMMARY OF THE INVENTION

The invention provides a determination method for satellite autonomous positioning and attitude determination. The method is applied to satellite autonomous navigation, and can effectively improve the efficiency and accuracy of satellite attitude and orbit estimation, thereby realizing real-time attitude determination in satellite offline state. , and improve the stability and reliability of the satellite operation control system.

The invention provides a method for judging the autonomous pose of a satellite, comprising: acquiring remote sensing images captured by an optical camera; searching in a knowledge map based on the remote sensing images to obtain template images; transmitting the remote sensing images and the template images to depth convolution network to obtain matching feature information; determine the optical camera imaging model, motion equation and observation equation; and solve the motion equation and observation equation according to the matching feature information to obtain the satellite attitude state.

Further, the step of transmitting the remote sensing image and the template image to a deep convolutional network to obtain matching feature information includes: transmitting the remote sensing image and the template image to the preprocessing layer of the deep convolutional network; extracting multiple sets of features through the preprocessing layer. information; build a feature space converter; perform normalization operations through the feature space converter according to the obtained sets of feature information; transmit the obtained normalized feature information to a deep convolutional network; information to perform a matching operation to obtain matching feature information.

Further, in the step of extracting multiple sets of feature information through the preprocessing layer, it also includes: extracting corresponding feature information according to different convolution kernels in the preprocessing layer; to obtain multiple sets of feature information.

Further, the step of determining the imaging model, the motion equation and the observation equation of the optical camera includes: acquiring the light beam through the optical camera, and projecting the image on the camera plane; according to the motion change of the satellite, the feature information extracted by the optical camera, and the location of the satellite in the camera; The noise obtained during the motion process is used to obtain the motion equation corresponding to the motion of the satellite; according to the observation data of the satellite and the noise obtained by the satellite during the observation process, the observation equation corresponding to the motion equation is obtained.

Further, the step of solving the motion equation and the observation equation to obtain the satellite attitude state according to the matching feature information includes: performing a parameterization operation on the motion equation and the observation equation; The probability equation of the motion state; by using the maximum a posteriori probability method, the probability equation is solved to obtain the position and attitude state information of the satellite.

The invention also provides a satellite autonomous pose determination system, comprising: a remote sensing image acquisition module for acquiring remote sensing images captured by an optical camera; a template image acquisition module for retrieving in a knowledge map based on the remote sensing images, to obtain a template image; a feature information obtaining module for transmitting the remote sensing image and the template image to a deep convolutional network to obtain matching feature information; a model equation determination module for determining the optical camera imaging model, motion equation and observation equation; and The attitude state obtaining module is used for solving the motion equation and the observation equation according to the matching feature information to obtain the satellite attitude state.

Further, the feature information obtaining module includes: an image transmission unit for transmitting remote sensing images and template images to a preprocessing layer of a deep convolutional network; a feature extraction unit for extracting multiple sets of feature information through the preprocessing layer; converting The normalization operation unit is used to perform the normalization operation through the feature space converter according to the obtained multiple sets of feature information; the matching operation unit is used for the normalization operation. Perform a matching operation on the transformed feature information to obtain matching feature information.

Further, the model equation determination module includes: a beam acquisition unit, used for acquiring the beam through an optical camera, and projecting imaging on the camera plane; a motion equation obtaining unit, used for according to the motion change of the satellite and the feature extracted by the optical camera The information and the noise obtained by the satellite during the movement process can obtain the motion equation corresponding to the satellite movement; the observation equation obtaining unit is used to obtain the corresponding motion equation according to the satellite observation data and the noise obtained by the satellite during the observation process. observation equation.

Further, the attitude state obtaining module includes: a parameterized operation unit for performing parameterized operations on the motion equation and the observation equation; a probability equation obtaining unit for obtaining the probability of the motion state of the satellite according to the conditional probability information, etc. formula; the position and attitude state obtaining unit is used to solve the probability equation by using the maximum a posteriori probability method to obtain the position and attitude state information of the satellite.

The present invention also provides a storage medium, wherein a plurality of instructions are stored in the storage medium, and the instructions are suitable for being loaded by a processor to execute the aforementioned method for determining the autonomous position and attitude of a satellite.

The embodiment of the present invention provides a method for determining the autonomous pose of a satellite. By transmitting a remote sensing image captured by an optical camera and a template image retrieved from a knowledge map to a feature matching deep convolutional network, matching feature information is obtained; The feature information is brought into the determined optical camera imaging model, motion equation and observation equation to solve, so as to obtain the satellite attitude state. The method of the invention can effectively improve the efficiency and accuracy of satellite attitude and orbit estimation, so as to realize the real-time attitude determination of the satellite offline state, and improve the stability and reliability of the satellite operation control system.

Description of drawings

The technical solutions and other beneficial effects of the present invention will be apparent through the detailed description of the specific embodiments of the present invention with reference to the accompanying drawings.

FIG. 1 is a flowchart of a method for determining an autonomous pose of a satellite according to an embodiment of the present invention.

FIG. 2 is a schematic diagram of an imaging model of an optical camera provided by an embodiment of the present invention.

FIG. 3 is a schematic diagram of the architecture of a multi-scale deep convolutional network provided by an embodiment of the present invention.

FIG. 4 is a flowchart of sub-steps of step S3 according to an embodiment of the present invention.

FIG. 5 is a flowchart of sub-steps of step S302 provided by an embodiment of the present invention.

FIG. 6 is a flowchart of sub-steps of step S4 provided by an embodiment of the present invention.

FIG. 7 is a flowchart of sub-steps of step S5 according to an embodiment of the present invention.

FIG. 8 is a functional block diagram of a system for determining an autonomous position and attitude of a satellite according to an embodiment of the present invention.

FIG. 9 is a functional block diagram of the feature information obtaining module provided in an embodiment of the present invention.

FIG. 10 is a functional block diagram of the feature extraction unit block provided in an embodiment of the present invention.

FIG. 11 is a functional block diagram of the model equation determination module provided in an embodiment of the present invention.

FIG. 12 is a functional block diagram of an attitude state obtaining module provided in an embodiment of the present invention.

Detailed ways

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

The following disclosure provides many different embodiments or examples for implementing different structures of the present invention. In order to simplify the disclosure of the present invention, the components and arrangements of specific examples are described below. Of course, they are only examples and are not intended to limit the invention. Furthermore, the present disclosure may repeat reference numerals and/or reference letters in different instances for the purpose of simplicity and clarity and not in itself indicative of a relationship between the various embodiments and/or arrangements discussed. In addition, the present disclosure provides examples of various specific processes and materials, but one of ordinary skill in the art will recognize the application of other processes and/or the use of other materials.

With the increasing development of Synchronous Positioning and Mapping (SLAM) technology and the continuous improvement of the ground resolution of optical loads, the method of autonomous orbit determination and attitude determination of satellites based on visual SLAM technology has brought a brand-new method to the autonomous navigation of satellites. Direction of development. Compared with traditional satellite navigation methods, the method of autonomous orbit determination and attitude determination of satellites based on visual SLAM technology has obvious advantages: on the one hand, the optical image itself contains rich information, and the field of image vision based on visual SLAM method has made great progress. , enabling the realization of satellite autonomous navigation based on visual SLAM technology; on the other hand, using artificial intelligence methods based on deep learning, the satellite has high autonomous performance, effectively reducing the dependence on ground stations and saving human and material resources.

As shown in FIG. 1 , an embodiment of the present invention provides a SLAM-based satellite autonomous pose determination method, which includes the following steps S1 to S9.

With reference to Fig. 2, in step S1, a remote sensing image captured by an optical camera is acquired.

In step S2, based on the remote sensing image, the knowledge map is retrieved to obtain a template image.

Step S3, transmitting the remote sensing image and the template image to a deep convolutional network (as shown in Figure 3, convolutional neural network (CNN), dense feature (dense feature), sparse feature (derived feature)) to obtain matching feature information. As shown in FIG. 4 , step S3 specifically includes the following steps: S301 to S306.

Step S301, transmitting the remote sensing image and the template image to the preprocessing layer of the deep convolutional network.

Step S302, extracting multiple sets of feature information through the preprocessing layer.

Further, as shown in Figure 5, step S302 specifically includes the following steps:

Step S3021, extract corresponding feature information according to different convolution kernels in the preprocessing layer.

Step S3022, performing a superposition operation on the corresponding feature information extracted by different convolution kernels to obtain multiple sets of feature information. Considering the difference in scale and position between the target feature on the remote sensing image captured by the optical camera and the target feature on the retrieved template image, the captured remote sensing image and the retrieved template image are firstly transferred to the deep convolutional network. The preprocessing layer performs a multi-scale feature extraction layer, which has feature convolution kernels of multiple scales, and the remote sensing images will be transmitted in multiple copies to synchronize feature extraction operations on convolution kernels of different scales. At the same time, the convolution kernels of different scales are not independent. Different features are superimposed on multi-scale features through upsampling, and finally multiple sets of feature information with wide applicability are obtained.

Step S303, build a feature space converter.

Step S304, according to the obtained sets of feature information, normalization operation is performed through the feature space converter. After the preprocessing layer, considering the acquisition of the multiple sets of feature information, although each set of feature information has wide applicability, there is a problem of different feature spaces between different sets of features. Therefore, in the first step of the deep convolutional network In the second stage, a feature space normalization converter is introduced to perform feature space normalization operations on multiple sets of features.

Step S305, transmitting the obtained normalized feature information to the deep convolutional network.

Step S306, performing a matching operation on the normalized feature information to obtain matching feature information.

Specifically, after obtaining the normalized feature information, considering the sparseness and density of the normalized feature information, the feature information is subjected to sparse and dense synchronous matching convolution operations, and different feature matching algorithms are introduced for different operations. Then, different matching operations are superimposed to filter out the matching features with strong robustness.

In step S4, the imaging model of the optical camera, the motion equation and the observation equation are determined. As shown in FIG. 6 , step S4 specifically includes the following steps S401 to S404.

In step S401, a light beam is acquired by an optical camera, and an image is projected on the camera plane.

In step S402, a motion equation corresponding to the motion of the satellite is obtained according to the motion change of the satellite, the feature information extracted by the optical camera, and the noise obtained by the satellite during the motion process.

The satellite is running in orbit, and its motion equation is extracted into a mathematical equation: x _k =f(x _k-1 ,u _k ,w _k ), where x _k is the k-1th to kth satellites Motion change, _uk is the input of the sensor carried by the satellite (here is the feature extracted by the on-board optical camera), and w _k is the incoming noise during this motion.

In step S403, an observation equation corresponding to the motion equation is obtained according to the observation data of the satellite and the noise obtained by the satellite during the observation process.

该处的参数x、y和z表示坐标系三个坐标轴的位移量，参数θ表示转角，参数x下标k为第k-1至第k次的卫星运动变化，而输入的参数是两个时间间隔和转角的变化量，此时运动方程表示为：

最后联立上述的运动方程和观测方程，即

此时，w_k和v_k,_j为运动方程和观测方程的噪声，Z下标(k,j)表示x下标k运动过程中看到的第j个观察路标点y下标j，所产生的观察方程。Specifically, the satellite is running in orbit, and its pose is characterized by displacement and rotation angle, namely

The parameters x, y and z here represent the displacement of the three coordinate axes of the coordinate system, the parameter θ represents the rotation angle, the subscript k of the parameter x is the k-1 to k-th satellite motion change, and the input parameters are two The change of time interval and rotation angle, the equation of motion is expressed as:

Finally, the above equations of motion and observation equations are combined, namely

At this time, w _k and v _k , _j are the noise of the motion equation and the observation equation, and the Z subscript (k, j) represents the j-th observation landmark point y subscript j seen during the movement of the x subscript k, so The resulting observation equation.

In step S404, the coordinate information of an observation point and the pose information of the optical camera are acquired, and a constraint equation system of the observation point is constructed.

此处u下标1和v下标1分别表示像素平面横轴和纵轴坐标，X、Y和Z表征空间坐标系下的三轴的齐次分量。Specifically, consider a certain observation point P in the space, which is also the following feature point, the homogeneous coordinate of this point is P=(X, Y, Z, 1) ^T , and on the pixel plane, the corresponding projection point coordinate is P' =(u ₁ , v ₁ , 1) ^T ; at this time, the pose of the onboard optical camera is represented by R, t, where R is a rotation matrix and t is a translation vector. According to the coordinates of the observation point and the camera pose information, construct the equation:

Here, the u subscript 1 and the v subscript 1 represent the coordinates of the horizontal axis and the vertical axis of the pixel plane, respectively, and X, Y, and Z represent the homogeneous components of the three axes in the space coordinate system.

进一步地，为了简化书写，定义T的行向量为：

由上述可知，每对特征点提供两个约束等式，那么N对特征点的约束方程组为：

N表示多个，例如1对特征点，2对特征点，3对特征点……N对特征点(常见使用8对)，P为空间中某一个观测点。Using the last row of the augmented matrix consisting of the rotation matrix and the translation vector, eliminating s, we get two constraint equations:

Further, to simplify writing, define the row vector of T as:

It can be seen from the above that each pair of feature points provides two constraint equations, then the constraint equations of N pairs of feature points are:

N means multiple, such as 1 pair of feature points, 2 pairs of feature points, 3 pairs of feature points...N pairs of feature points (8 pairs are commonly used), and P is an observation point in the space.

Continuing to refer to FIG. 1 , step S5 , according to the matching feature information, the motion equation and the observation equation are solved to obtain the satellite attitude state.

As shown in FIG. 7 , step S5 specifically includes the following steps S501 to S503.

Step S501, parameterize the motion equation and the observation equation to obtain sz _k,j =K(R _k y _j +t _k ), where K is the internal parameter of the optical camera, and s is the distance of the pixel point, assuming this realization The noise of the mode satisfies the Gaussian distribution with zero mean, that is, w _k ～N(0,R _k ),v _k ～N(0,Q _k,j ), and the pose x is inferred from the noisy data z and u information, which constitutes a state estimation problem.

In step S502, a probability equation of the motion state of the satellite is obtained according to the conditional probability information.

Knowing the input data u and observation data z, solve the state of the pose x and the landmark point y, and introduce the Bayesian rule, that is, get

Step S503 , by adopting the maximum a posteriori probability method, the probability equation is solved to obtain the position and attitude state information of the satellite.

该处的arg max指最大后验概率，下标MAP也是最大后验概率所对应的值。基于上述高斯分布的假设下，对应某一次观测：z_k,j＝h(y_j,x_k)+v_k,j，同时噪声满足v_k～N(0,Q_k,j)，所以此次观测的条件概率为：P(z_j,k|x_k,y_j)＝N(h(y_i,x_k),Q_k,j)。考察将高维高斯分布x～N(μ,Σ)展开，同时对其取负对数，即

考察上述负对数等式，因为对数单调递增，对原函数求最大相当于对负对数函数求最小，在最小化上述等式时，考虑到x值与右边第一项无关，因此只考虑右侧二次型项，带入观测方程，即

进一步假设各个时刻的运动和观测是独立的，因此可以对联合分布进行因式分解：

重新定义每次实际运动和观测与模型之间的误差：e_u,k＝x_k-f(x_k-1,u_k),e_z,j,k＝z_k,j-h(x_k,y_j)。将上述的最小化等式转化为联乘形式，即

这样就转化为一个最小二乘问题。最后由上述的最小二乘等式，可以求解出星载光学相机的位姿状态，进而可以计算出卫星的轨道等信息。其中，上述参数中，μ和Σ分别是高斯分布的均值和方差；det()用于求解一个方阵的行列式；Q及下标(k,j)，表示正太分布中噪声方差由参数(k,j)所决定，(k,j)见上述解释。E及下标(u,k)，表示每次实际运动和观察与模型之间的误差。Specifically, the optimization method is used to solve the maximization of the posterior probability of a state, namely

The arg max here refers to the maximum posterior probability, and the subscript MAP is also the value corresponding to the maximum posterior probability. Based on the assumption of the above Gaussian distribution, corresponding to a certain observation: z _k,j =h(y _j ,x _k )+v _k,j , and the noise satisfies v _k ～N(0,Q _k,j ), so this The conditional probability of the second observation is: P(z _j,k |x _k ,y _j )=N(h(y _i ,x _k ),Q _k,j ). The investigation expands the high-dimensional Gaussian distribution x～N(μ,Σ), and takes the negative logarithm of it at the same time, namely

Considering the above negative logarithmic equation, because the logarithm increases monotonically, maximizing the original function is equivalent to minimizing the negative logarithmic function. When minimizing the above equation, considering that the x value has nothing to do with the first item on the right, so only Consider the quadratic term on the right side and bring it into the observation equation, that is

Further assuming that the movements and observations at each moment are independent, the joint distribution can be factored:

Redefine each actual movement and the error between the observation and the model: e _u,k =x _k -f(x _k-1 ,u _k ),e _z,j,k =z _k,j -h(x _k , y _j ). Convert the above minimization equation into a multiplication form, that is

This translates into a least squares problem. Finally, from the above least squares equation, the pose state of the onboard optical camera can be solved, and then the orbit and other information of the satellite can be calculated. Among the above parameters, μ and Σ are the mean and variance of the Gaussian distribution, respectively; det() is used to solve the determinant of a square matrix; Q and the subscript (k, j) indicate that the noise variance in the normal distribution is determined by the parameter ( k,j), and (k,j) are explained above. E and the subscripts (u, k) represent each actual movement and the error between the observation and the model.

As shown in FIG. 8 , an embodiment of the present invention further provides a satellite autonomous pose determination system 200, including: a remote sensing image acquisition module 201, a template image acquisition module 202, a feature information acquisition module 203, and a model equation determination module 204 And attitude state obtaining module 205 .

The remote sensing image acquisition module 201 is used to acquire remote sensing images captured by an optical camera.

The template image acquisition module 202 is used for searching in the knowledge graph based on the remote sensing image to obtain the template image.

The feature information obtaining module 203 is used to transmit the remote sensing image and the template image to the deep convolutional network to obtain matching feature information.

As shown in FIG. 9 , the feature information obtaining module 203 includes: an image transmission unit 2031 , a feature extraction unit 2032 , a converter construction unit 2033 , a normalization operation unit 2034 and a matching operation unit 2035 .

The image transmission unit 2031 is used to transmit the remote sensing image and the template image to the preprocessing layer of the deep convolutional network.

The feature extraction unit 2032 is used to extract multiple sets of feature information through the preprocessing layer; as shown in FIG. 10 , the feature extraction unit 2032 includes a feature information extraction subunit 20321 and a feature information superposition subunit 20322 . The feature information extraction subunit 20321 is used to extract corresponding feature information according to different convolution kernels in the preprocessing layer; the feature information superimposition subunit 20322 is used to superimpose the corresponding feature information extracted by different convolution kernels. , to obtain multiple sets of feature information.

The converter building unit 2033 is used to build a feature space converter.

The normalization operation unit 2034 is configured to perform a normalization operation through a feature space converter according to the obtained sets of feature information.

The matching operation unit 2035 is configured to perform a matching operation on the normalized feature information to obtain matching feature information.

The model equation determination module 204 is used to determine the imaging model of the optical camera, the motion equation and the observation equation. As shown in FIG. 11 , the model equation determination module 204 includes: a beam obtaining unit 2041 , a motion equation obtaining unit 2042 and an observation equation obtaining unit 2043 .

The light beam acquisition unit 2041 is used for acquiring the light beam through an optical camera, and projecting the image on the camera plane.

The motion equation obtaining unit 2042 is configured to obtain the motion equation corresponding to the satellite motion according to the motion change of the satellite, the feature information extracted by the optical camera, and the noise obtained by the satellite during the motion process.

The observation equation obtaining unit 2043 obtains the observation equation corresponding to the motion equation according to the observation data of the satellite and the noise obtained by the satellite during the observation process.

The attitude state obtaining module 205 is used for solving the motion equation and the observation equation according to the matching feature information, so as to obtain the satellite attitude state.

As shown in FIG. 12 , the posture state obtaining module 205 includes: a parameterized operation unit 2051 , a probability equation obtaining unit 2052 , and a posture state obtaining unit 2053 .

The parameterization operation unit 2051 is used to perform parameterization operations on the motion equation and the observation equation.

The probability equation obtaining unit 2052 is configured to obtain the probability equation of the motion state of the satellite according to the conditional probability information.

The pose state obtaining unit 2053 is configured to solve the probability equation by using the maximum a posteriori probability method to obtain the pose state information of the satellite.

The present invention also provides a storage medium, wherein a plurality of instructions are stored in the storage medium, and the instructions are suitable for being loaded by a processor to execute the method for determining the autonomous position and attitude of a satellite.

The invention provides a method for determining the autonomous pose of a satellite. By transmitting the remote sensing image captured by an optical camera and the template image retrieved from the knowledge map to a feature matching deep convolution network, matching feature information is obtained; according to the matching feature information , and bring it into the determined optical camera imaging model, motion equation and observation equation to solve, so as to obtain the satellite attitude state. The method of the invention can effectively improve the efficiency and accuracy of satellite attitude and orbit estimation, so as to realize the real-time attitude determination of the satellite offline state, and improve the stability and reliability of the satellite operation control system.

The present invention has been described in detail above, and the principles and implementations of the present invention are described in this paper by using specific examples. The skilled person should understand that it is still possible to modify the technical solutions recorded in the foregoing embodiments, or to perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the implementation of the present invention. Examples of the scope of technical solutions.

Claims (10)

1. A method for judging the autonomous pose of a satellite is characterized by comprising the following steps:

acquiring a remote sensing image shot by an optical camera;

retrieving in a knowledge graph based on the remote sensing image to obtain a template image;

transmitting the remote sensing image and the template image to a depth convolution network to obtain matching characteristic information;

determining an imaging model, a motion equation and an observation equation of an optical camera; and

and solving the motion equation and the observation equation according to the matching characteristic information to obtain the satellite attitude state.

2. The method for determining the autonomous pose of a satellite according to claim 1,

the step of transmitting the remote sensing image and the template image to the depth convolution network to obtain the matching characteristic information comprises the following steps:

transmitting the remote sensing image and the template image to a preprocessing layer of a depth convolution network;

extracting a plurality of groups of characteristic information through a pretreatment layer;

building a feature space converter;

carrying out normalization operation through a feature space converter according to the obtained multiple groups of feature information;

transmitting the obtained normalized feature information to a deep convolutional network; and

and performing matching operation on the normalized characteristic information to obtain matched characteristic information.

3. The method for determining the autonomous pose of a satellite according to claim 2, wherein the step of extracting a plurality of sets of feature information through a preprocessing layer further comprises:

extracting corresponding characteristic information according to different convolution kernels in the pretreatment layer;

and performing superposition operation on the corresponding characteristic information extracted by the different convolution kernels to obtain multiple groups of characteristic information.

4. The method for determining the autonomous pose of a satellite according to claim 1,

the step of determining the imaging model, the motion equation and the observation equation of the optical camera comprises the following steps:

acquiring a light beam through an optical camera, and projecting and imaging on a camera plane;

obtaining a motion equation corresponding to the motion of the satellite according to the motion change of the satellite, the characteristic information extracted by the optical camera and the noise obtained by the satellite in the motion process;

and obtaining an observation equation corresponding to the motion equation according to the observation data of the satellite and the noise obtained by the satellite in the observation process.

5. The method for determining the autonomous pose of a satellite according to claim 1, wherein the step of solving a motion equation and an observation equation to obtain the attitude state of the satellite according to the matching feature information comprises:

carrying out parameterization operation on the motion equation and the observation equation;

obtaining a probability equation of the motion state of the satellite according to the conditional probability information;

and solving the probability equation by adopting a maximum posterior probability method to obtain the pose state information of the satellite.

6. A system for determining an autonomous pose of a satellite, comprising:

the remote sensing image acquisition module is used for acquiring a remote sensing image shot by the optical camera;

the template image acquisition module is used for retrieving in the knowledge graph based on the remote sensing image to obtain a template image;

the characteristic information obtaining module is used for transmitting the remote sensing image and the template image to the depth convolution network so as to obtain matched characteristic information;

the model equation determining module is used for determining an optical camera imaging model, a motion equation and an observation equation; and

and the attitude state obtaining module is used for solving the motion equation and the observation equation according to the matching characteristic information so as to obtain the satellite attitude state.

7. The satellite autonomous pose determination system according to claim 6, wherein the feature information obtaining module comprises:

the image transmission unit is used for transmitting the remote sensing image and the template image to a preprocessing layer of the depth convolution network; the characteristic extraction unit is used for extracting a plurality of groups of characteristic information through the pretreatment layer;

the converter building unit is used for building a feature space converter;

the normalization operation unit is used for performing normalization operation through the feature space converter according to the obtained multiple groups of feature information;

and the matching operation unit is used for executing matching operation on the normalized characteristic information to obtain matched characteristic information.

8. The satellite autonomous pose determination system of claim 6, wherein the model equation determination module comprises:

the light beam acquisition unit is used for acquiring a light beam through an optical camera and projecting and imaging on a camera plane; the motion equation obtaining unit is used for obtaining a motion equation corresponding to the motion of the satellite according to the motion change of the satellite, the feature information extracted by the optical camera and the noise obtained by the satellite in the motion process; and the observation equation obtaining unit is used for obtaining an observation equation corresponding to the motion equation according to the observation data of the satellite and the noise obtained by the satellite in the observation process.

9. The satellite autonomous pose determination system of claim 6, wherein the attitude state derivation module comprises:

the parameterization operation unit is used for carrying out parameterization operation on the motion equation and the observation equation;

a probability equation obtaining unit, configured to obtain a probability equation of a motion state of the satellite according to the conditional probability information;

and the pose state obtaining unit is used for solving the probability equation by adopting a maximum posterior probability method so as to obtain the pose state information of the satellite.

10. A storage medium having stored therein a plurality of instructions adapted to be loaded by a processor to perform the method of determining an autonomous pose of a satellite of any of claims 1 to 5.

Images