A space target visual measurement method combined with acceleration drift estimation

Abstract

The present application relates to the field of space in-orbit close-range relative visual navigation technology, and particularly relates to a space target visual measurement method combined with acceleration drift estimation, a tracking spacecraft is provided with an acceleration sensor and a camera, S1: obtaining initial relative position information and relative velocity information through a measurement system, S2: obtaining acceleration values and space target images in real time; S3: establishing a relative motion equation, substituting the obtained acceleration values and the relative position and relative velocity information at the previous moment into the relative motion equation to calculate position information; S4: calculating an error amount, and estimating the errors of the relative position information, the relative velocity information and the acceleration values in the dynamic equation through a filtering algorithm according to the error amount to correct the calculation results obtained in the dynamic equation, so that accurate positioning of a target spacecraft can be realized.

Description

Technical Field

[0001] This invention relates to the field of on-orbit close-range relative visual navigation technology, specifically to a visual measurement method for space targets that incorporates acceleration drift estimation. Background Technology

[0002] Almost all on-orbit operational missions involve guiding a target spacecraft using visual sensors to perform operations at a suitable location. However, ordinary CCD cameras have limited measurement range, making it difficult to "see" a target clearly at distances exceeding 100 meters. Within this distance range, ground-based tracking and measurement systems are limited by tracking conditions and cannot accurately measure the target's relative position to the tracked spacecraft. To address the relative measurement problem in the transition from ground-based to visual measurement, current solutions involve using relatively expensive equipment such as microwave radar or long-focal-length cameras.

[0003] Deep learning-based target recognition technology is increasingly widely used in ground-based scenarios, and it can now identify small targets in remotely sensed images. Applying this technology to space target recognition, by training the network with ground-based simulated data and pseudo-realistic images, and then loading the trained network onto the onboard computing unit, it can quickly and accurately identify spacecraft targets of different scales. However, when the distance between the target and the camera is large, the target's projection on the imaging plane occupies very few pixels, making it difficult to separate the target spacecraft from other stationary targets in the space background. Since tracking the spacecraft is in a maneuvering state, traditional trajectory association methods are no longer applicable. In such cases, some prior information is needed to filter potential targets. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the present invention provides a spatial target visual measurement method that combines acceleration drift estimation to achieve accurate positioning of target spacecraft.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0006] A visual measurement method for space targets combining acceleration drift estimation, tracking spacecraft equipped with acceleration sensors and cameras.

[0007] Includes the following steps:

[0008] S1: Acquire initial relative position and relative velocity information through a measurement system, wherein the relative position and relative velocity information are the position and velocity information of the space target relative to the tracking spacecraft;

[0009] S2: Real-time acquisition of acceleration values ​​and spatial target images; wherein, the acceleration values ​​are parameters measured by the acceleration sensor; and the spatial target images are acquired by the camera.

[0010] S3: Establish the relative motion equation, substitute the current acceleration value and the relative position and relative velocity information of the previous calculation cycle into the relative motion equation, and calculate the position information, which is the relative position of the space target relative to the tracking spacecraft at the current moment;

[0011] S4: Calculate the error amount. Based on the error amount, use a filtering algorithm to estimate the error of the relative position information, relative velocity information and acceleration value in the dynamic equation to correct the calculation results obtained in the dynamic equation. The error amount is the error between the position information and the image position information. The image position information is the position of the spatial target obtained from the spatial target image.

[0012] Furthermore, in a spatial target visual measurement method combining acceleration drift estimation in this application, in step S3, the relative motion coordinate system of the tracking spacecraft and the space target is defined using the spacecraft orbital coordinate system, and the relative motion equations established under the relative motion coordinate system are as follows:

[0013]

[0014] Where n is the orbital angular velocity of the target spacecraft, f = [f x f y f z [ represents the acceleration value;]

[0015] The results obtained from the equations of relative motion are as follows:

[0016]

[0017] Where T is the calculation period, [x k y k , z k [Δx] represents the position information at the current time k. k-1 Δy k-1 Δz k-1 [The image shows] the relative position error estimated by the filtering algorithm. Let be the relative velocity of the spatial target at the current moment. The relative velocity error estimated by the filtering algorithm is Δf = [Δf x , Δf y , Δf z ] represents the accelerometer error estimated by the filtering algorithm, and k-1 represents the time of the previous calculation cycle before the current time k.

[0018] Furthermore, in a spatial target visual measurement method combining acceleration drift estimation in this application, in step S4, after converting the position information into image coordinate system position information, the error between the image coordinate system position information and the image position information is calculated; the image coordinate system position information corresponds to the coordinate system of the spatial target image information.

[0019] Furthermore, a spatial target visual measurement method combining acceleration drift estimation in this application converts position information into image coordinate system position information according to the following formula:

[0020]

[0021] in, Let Pc be the spatial target position in the camera coordinate system, which is obtained through position information transformation. f is the camera focal length, and (Cx, Cy) are the coordinates of the camera principal point on the camera plane.

[0022] Optionally, a spatial target visual measurement method combining acceleration drift estimation in this application uses the following formula to calculate the spatial target position information P in the orbital coordinate system. H Convert to the spatial target position Pc in the camera coordinate system:

[0023]

[0024] Assuming the camera's optical axis is parallel to the Y-direction of the orbital coordinate system, and the camera's x and y axes are aligned with the X and Z directions of the orbital coordinate system, then the attitude transformation matrix from the camera coordinate system to the orbital coordinate system is:

[0025] Optionally, a spatial target visual measurement method combining acceleration drift estimation in this application uses the following formula to convert the position information P... H Convert to the spatial target position Pc in the camera coordinate system:

[0026]

[0027] in, To determine the attitude matrix of the tracking spacecraft's body coordinate system relative to the orbital coordinate system, as measured by the attitude measurement system, the camera coordinate system is aligned with the tracking spacecraft's body coordinate system.

[0028] Furthermore, in a spatial target visual measurement method combining acceleration drift estimation in this application, the error MSE between the position information and the image position information is obtained by the following formula:

[0029]

[0030] Where (u,v) represents the image location information. This refers to the position information in the image coordinate system.

[0031] Furthermore, in this application, a spatial target visual measurement method combining acceleration drift estimation is proposed. The acquired spatial target image information is input into a preset target detection network to output target detection results. If multiple target detection results are output, only the target detection result closest to the predicted position is retained. The center position of the spatial target is calculated based on the target detection results as the image position information. As a preferred embodiment of this application, the target detection network can be implemented by adding small detection heads to a general target detection network, or a Transformer-type network can be used.

[0032] As can be seen from the above technical solution, the present invention has the following beneficial effects:

[0033] This invention provides a visual measurement method for space targets that combines acceleration drift estimation. It utilizes accelerometers to measure and track the maneuvering acceleration of a spacecraft, predicts the relative position of the target spacecraft using relative motion equations, and estimates acceleration drift through filtering. This allows the target spacecraft to be separated from other stationary targets in the space background when it is far from the camera. This method integrates acceleration drift estimation with deep learning technology, increasing the camera's measurement range and thus enabling precise positioning of the target spacecraft. Furthermore, filtering for acceleration error estimation improves measurement accuracy. Attached Figure Description

[0034] Figure 1 This is a diagram defining the orbital coordinate system of the space target and the tracking spacecraft in the embodiments of this application. Detailed Implementation

[0035] This embodiment provides a visual measurement method for space targets that combines acceleration drift estimation. The tracking spacecraft is equipped with an acceleration sensor and a camera.

[0036] Includes the following steps:

[0037] S1: Acquire initial relative position and relative velocity information through a measurement system. The relative position and relative velocity information are the position and velocity information of the space target relative to the tracking spacecraft. The specific measurement system can be GPS or other ground measurement systems.

[0038] S2: Real-time acquisition of acceleration values ​​and spatial target images; wherein, the acceleration values ​​are parameters measured by the acceleration sensor; and the spatial target images are acquired by the camera.

[0039] S3: Establish the relative motion equation, substitute the current acceleration value and the relative position and relative velocity information of the previous calculation cycle into the relative motion equation, and calculate the position information, which is the relative position of the space target relative to the tracking spacecraft at the current moment;

[0040] S4: Calculate the error amount. Based on the error amount, use a filtering algorithm to estimate the error of the relative position information, relative velocity information and acceleration value in the dynamic equation to correct the calculation results obtained in the dynamic equation. The error amount is the error between the position information and the image position information. The image position information is the position of the spatial target obtained from the spatial target image.

[0041] In this embodiment, the space target is an in-orbit spacecraft.

[0042] In this embodiment, in step S2, the error of the measurement data is estimated by a filtering algorithm to correct the calculation results obtained in the dynamic equation.

[0043] In this embodiment, in step S2, the relative motion coordinate system between the tracking spacecraft and the space target is defined using the spacecraft orbit coordinate system, i.e., as shown below. Figure 1 As shown, the origin of the orbital coordinate system of the space target is fixed to the center of mass of the space target and moves along the orbit with it. The X-axis coincides with the geocentric vector of the space target. The Y-axis is perpendicular to the X-axis in the orbital plane and points in the direction of motion as positive. The Z-axis is determined by the right-hand rule. The relative motion equations established under the aforementioned relative motion coordinate system are as follows:

[0044]

[0045] Where n is the orbital angular velocity of the target spacecraft, f = [f x f y f z [ ] represents the acceleration value;

[0046] The results obtained from the equations of relative motion are as follows:

[0047]

[0048] Where T is the calculation period, [x k y k , z k ] represents the position information at the current time k, [Δx k-1 Δy k-1 Δz k -1 [The image shows] the relative position error estimated by the filtering algorithm. Let be the relative velocity of the spatial target at the current moment. The relative velocity error estimated by the filtering algorithm is Δf = [Δf x , Δf y , Δf z ] represents the accelerometer error estimated by the filtering algorithm, and k-1 represents the time of the previous calculation cycle before the current time k.

[0049] In this embodiment, in step S3, after converting the position information into image coordinate system position information, the error between the image coordinate system position information and the image position information is calculated; the image coordinate system position information corresponds to the coordinate system of the spatial target image information.

[0050] In this embodiment, the position information is converted into image coordinate system position information according to the following formula:

[0051]

[0052] in, Let Pc be the spatial target position in the camera coordinate system, which is obtained by transforming the position information. f is the camera focal length, and (Cx, Cy) are the coordinates of the camera principal point on the camera plane.

[0053] In one embodiment, the spatial target position information P in the orbital coordinate system is expressed by the following formula. H Convert to the spatial target position Pc in the camera coordinate system:

[0054]

[0055] Assuming the camera's optical axis is parallel to the Y-direction of the orbital coordinate system, and the camera's x-axis and y-axis are aligned with the X and Z directions of the orbital coordinate system, then the attitude transformation matrix from the camera coordinate system to the orbital coordinate system is:

[0056] In another embodiment, the location information P is expressed by the following formula: H Convert to the spatial target position Pc in the camera coordinate system:

[0057]

[0058] in, To determine the attitude matrix of the tracking spacecraft's body coordinate system relative to the orbital coordinate system, as measured by the attitude measurement system, the camera coordinate system is aligned with the tracking spacecraft's body coordinate system.

[0059] In this embodiment, the error MSE between the location information and the image location information is obtained by the following formula:

[0060]

[0061] Where (u,v) represents the image location information. This refers to the position information of the image coordinate system.

[0062] In this embodiment, the acquired spatial target image information is input into a preset target detection network to output target detection results. If multiple target detection results are output, only the target detection result closest to the predicted position is retained, and the center position of the spatial target is calculated based on the target detection results as the image position information. The target detection network can be implemented by adding a small detection head to a general target detection network, or a Transformer-type network can be used. In this embodiment, a YOLOv81 with a small detection head scheme is used. Specifically, refer to the GEO target detection method based on an improved YOLOv81 disclosed in Chinese Invention Patent Publication No. CN118918312A. The original YOLOv81 has three detection heads with target sizes of 32x32, 16x16, and 8x8 respectively. A small detection head of size 4x4 is added to this, and the modified network can be used for multi-scale target detection.

[0063] Based on the above method, simulation verification showed that when using a binocular camera to measure a target spacecraft with a diameter of one meter, the detection accuracy was 100% within a distance of 60 meters, 98% within 60-70 meters, 80% within 70-80 meters, and 60% within 80-100 meters. Using the method described in this embodiment, the detection accuracy in the 60-100 meter range was greater than 90%.

[0064] The technical principles of the present invention have been described above with reference to specific embodiments. These descriptions are merely for explaining the principles of the invention and should not be construed as limiting the scope of protection of the invention in any way. Based on this explanation, those skilled in the art can conceive of other specific embodiments of the invention without creative effort, and these embodiments will all fall within the scope of protection of the present invention.

Claims

1. A spatial target visual measurement method combining acceleration drift estimation, characterized in that: The tracking spacecraft is equipped with an acceleration sensor and a camera. Includes the following steps: S1: Acquire initial relative position and relative velocity information through a measurement system, wherein the relative position and relative velocity information are the position and velocity information of the space target relative to the tracking spacecraft; S2: Real-time acquisition of acceleration values ​​and spatial target images; wherein, the acceleration values ​​are parameters measured by the acceleration sensor; and the spatial target images are acquired by the camera. S3: Establish the relative motion equation, substitute the current acceleration value and the relative position and relative velocity information of the previous calculation cycle into the relative motion equation, and calculate the position information, which is the relative position of the space target relative to the tracking spacecraft at the current moment; S4: Calculate the error amount, and based on the error amount, use a filtering algorithm to estimate the error of the relative position information, relative velocity information and acceleration value in the dynamic equation to correct the calculation results obtained in the dynamic equation. The error amount is the error between the position information and the image position information. The image position information is the position of the spatial target obtained from the spatial target image. In step S4, after converting the position information into image coordinate system position information, the error between the image coordinate system position information and the image position information is calculated; the image coordinate system position information corresponds to the coordinate system of the spatial target image information. The location information is converted into image coordinate system location information according to the following formula: in, Let Pc be the spatial target position in the camera coordinate system, which is obtained by transforming the aforementioned position information. f is the camera focal length, and (Cx, Cy) are the coordinates of the camera principal point on the camera plane. The spatial target position information Pc in the orbital coordinate system is then transformed using the following formula. H Convert to the spatial target position Pc in the camera coordinate system: Assuming the camera's optical axis is parallel to the Y-direction of the orbital coordinate system, and the camera's x-axis and y-axis are aligned with the X and Z directions of the orbital coordinate system, then the attitude transformation matrix from the camera coordinate system to the orbital coordinate system is: The location information P is expressed by the following formula. H Convert to the spatial target position Pc in the camera coordinate system: in, To align the camera coordinate system with the tracking spacecraft's body coordinate system based on the attitude matrix of the tracking spacecraft's body coordinate system relative to the orbital coordinate system, as measured by the attitude measurement system; The error MSE between the location information and the image location information is obtained by the following formula: Where (u,v) represents the image location information. This refers to the position information of the image coordinate system; The acquired spatial target image information is input into a preset target detection network to output target detection results; if multiple target detection results are output, only the target detection result closest to the predicted position is retained, and the center position of the spatial target is calculated based on the target detection results as the image position information.

2. The spatial target visual measurement method combining acceleration drift estimation according to claim 1, characterized in that: In step S3, the relative motion coordinate system between the tracking spacecraft and the space target is defined using the spacecraft orbital coordinate system. The relative motion equations established under this coordinate system are as follows: Where n is the orbital angular velocity of the target spacecraft, f = [f x f y f z [ ] represents the acceleration value; The results obtained from the equations of relative motion are as follows: Where T is the calculation period, [x k ,y k ,z k [Δx] represents the position information at the current time k. k-1 ,Δy k-1 ,Δz k-1 [The image shows] the relative position error estimated by the filtering algorithm. Let k be the relative velocity of the spatial target at the current time k. The relative velocity error estimated by the filtering algorithm is Δf = [Δf x , Δf y , Δf z ] represents the acceleration error estimated by the filtering algorithm, and k-1 represents the time of the previous calculation cycle before the current time k.

Images