Method and system for capturing and tracking moving target with high confidence by satellite

Abstract

The invention provides a high-confidence capturing and tracking method and system for a moving target by a satellite. The method comprises the following steps: scheduling a load to point to the target according to target initial guide information; taking a vector from the load to the target as a center, and superposing sine control quantity to control the target to perform circular motion to obtain a target track; target over-threshold points are extracted, target alternative points are formed, multi-frame association is carried out, and the confidence coefficient is calculated; continuous frame judgment is carried out, and abnormal elements and fixed star targets are identified; and eliminating abnormal elements and fixed star targets, and updating a target detection result in real time. An interference source is removed through a dynamic background difference method, the target tracking stability is improved, and the high-confidence capturing and tracking capability is achieved; an active circumferential control means is adopted, so that the target is in circular motion in a view field, and the method has more practical value in capturing and tracking application of a space moving target; a false alarm source is effectively distinguished by using target image surface motion characteristics and coordinate transformation, the influence of multiple false alarm sources in a target tracking process is reduced, and the tracking confidence is improved.

Description

Technical Field

[0001] The present invention belongs to the field of onboard processing and autonomous control technology, and specifically relates to a method and system for satellites to capture and track moving targets with high confidence, thereby reducing the impact of false alarm sources such as image plane anomalies, fixed bright elements, background fluctuations, stellar targets, and stray light on detection and tracking performance. Background Art

[0002] The ability to capture and track moving targets in space using space-based imaging payloads is fundamental to space applications such as space debris monitoring, asteroid exploration, and autonomous spacecraft navigation. In the context of deep space exploration, to ensure detection sensitivity, space-based payloads typically employ optical imaging cameras with long focal lengths and small fields of view, using a tracking mode to ensure long-term, continuous tracking of rapidly moving targets in space. However, factors such as changes in the starry sky in the image background, fluctuations in the ground's stray light, changes in external stray light, and anomalous pixels within the image plane can all interfere with target detection and tracking. Therefore, improving the reliability and continuity of target detection during the capture and tracking process is key to achieving high-confidence capture and tracking.

[0003] In the existing technology, the technical solution disclosed in the patent document "Method and System for Tracking and Pointing Satellite Attitude Control of Space-Based Motion Targets" (CN109911248A) can generate a control torque for the satellite to adjust the satellite's attitude, thereby achieving the purpose of tracking the target, but it cannot reduce the impact of multiple interference sources on the image plane, and the confidence level of the capture and tracking capability is low.

[0004] The technical solution disclosed in the patent document "A guided imaging method, a space-based imaging method and device for space targets" (CN104034314A) can perform threshold judgment based on the distance between the observation satellite and the target satellite, which serves as the basis for switching between the three space-based imaging modes of guided imaging, waiting imaging and tracking imaging. However, it is limited to using spatial distance as the basis for switching imaging modes, and does not consider the issue of on-board target detection confidence during the actual target capture and tracking process.

[0005] The technical solution disclosed in the patent document "Space Dynamic Multi-Target On-Sat Autonomous Tracking Method and System" (CN114489099A) calculates the trackable period based on satellite visibility constraints, and realizes attitude maneuvering and stable tracking through the satellite attitude control system. However, its focus is on task allocation during the target visibility period, and the tracking method is not clearly explained.

[0006] The technical solution disclosed in the patent document "A Space-Based Imaging Method and Apparatus for Space Targets" (CN103983250A) determines observation feasibility based on the target's sub-satellite location and schedules the space-based imaging system. Once the target enters the space-based imaging system's guidance field of view, tracking imaging is used to correct for deviations from ground-based planning. However, it fails to address the false alarm problem inherent in target detection and tracking, nor does it address specific tracking methods.

[0007] The patent document "A method for detecting and tracking space debris targets in space-based wide-area surveillance images" (CN117876736A) discloses a method for extracting faint space debris targets against a dense star map background. However, it does not consider false alarms such as image plane anomalies, and the image preprocessing and star map matching algorithms used require high computing resources.

[0008] The patent document "Method and System for Satellite Guidance and Tracking of Non-cooperative Moving Targets" (CN118859279A) discloses the use of satellites to perform high-spatial and high-spectral resolution imaging of high-speed dynamic targets, enabling low-orbit optical satellites to search, discover, guide and continuously track non-cooperative targets moving in the air, but also does not consider the problem of false alarms.

[0009] Therefore, there is an urgent need for a high-confidence capture and tracking method and system for satellites to track moving targets, which can eliminate interference from factors such as changes in the starry sky in the image background, fluctuations in the ground and atmospheric stray light, changes in external stray light, and abnormal pixels on the internal image plane. Summary of the Invention

[0010] In view of the defects in the prior art, the purpose of the present invention is to provide a method and system for satellite to capture and track moving targets with high confidence.

[0011] The method for capturing and tracking a moving target with high confidence by a satellite provided by the present invention comprises:

[0012] Step S1: Schedule the payload to point to the target according to the initial guidance information of the target, recursively deduce the predicted position and speed information of the target, obtain the payload pointing information, and adjust the payload pointing to the target;

[0013] Step S2: With the star vector as the center, the sinusoidal control variable is superimposed to control the target to perform circular motion and obtain the target trajectory;

[0014] Step S3: extract target crossing threshold points, connect target crossing threshold points and calculate gravity center, form target candidate points and perform multi-frame association, and calculate confidence;

[0015] Step S4: Perform continuous frame judgment on the target candidate points, identify the target candidate points whose column positions remain unchanged in the image coordinate system as abnormal elements and remove them;

[0016] Step S5: Based on the payload pointing information, the image row and column coordinates of the target candidate points after removing the outliers are converted into the celestial coordinate system, and through continuous frame judgment, the target candidate points whose positions remain unchanged in the celestial coordinate system are identified as star targets;

[0017] Step S6: Eliminate the star targets, select the target with the highest confidence in the target trajectory as the tracking target, and update the target detection results in real time with the tracking target vector as the center.

[0018] Preferably, in step S1, the predicted position and velocity information of the target are predicted by the orbital dynamics equation according to the initial guidance information of the target, and the payload visual axis is guided to point to the predicted position of the target, and the target motion trajectory is guided from the edge of the payload's field of view to the payload's field of view. The target stops when the angle between the payload's field of view center vector and the star sight vector is less than one-quarter of the payload's field of view angle, and the payload is pointed into place.

[0019] After the payload is pointed into place in step S2, a sinusoidal control variable is superimposed on the image plane perpendicular to the star-eye vector with the star-eye vector as the center. The control center controls the target to move in a circular motion to obtain the target trajectory:

[0020]

[0021] Among them, (x k ,y k ,1) represents the control center vector that changes with time t;

[0022] x k Represents the x-direction component of the image plane control center;

[0023] y k Indicates the y-direction component of the image plane control center;

[0024] (x0,y0,1) represents the target vector;

[0025] x0 represents the x-direction component of the image target;

[0026] y0 represents the y-direction component of the image target;

[0027] r represents the circumferential control radius;

[0028] ω represents the circular control angular velocity.

[0029] Preferably, in step S3, the grayscale values ​​of the current frame image and the background differential frame interval frame image are subtracted. If the differential grayscale value after the subtraction is greater than or equal to 0, the differential grayscale value is retained as the differential pixel grayscale value. If the differential grayscale value is less than 0, the differential pixel grayscale value is set to 0:

[0030]

[0031] Among them, m represents the row vector of pixel grayscale;

[0032] n represents the column vector of pixel grayscale;

[0033] P k (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kth frame;

[0034] P k-d (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kd-th frame;

[0035] d represents the background difference frame interval;

[0036] P diff (m,n) represents the pixel grayscale value after dynamic background subtraction.

[0037] Connected domain association is performed based on the extracted target crossing threshold points. The connected target crossing threshold points are regarded as pixels belonging to the same target. The target center of gravity is calculated and used as the target candidate point for multi-frame association. The confidence of the target trajectory is obtained based on the consistency of the target trajectory and grayscale energy.

[0038] Preferably, in step S4, each target crossing threshold point is judged in a continuous frame. If the pixel coordinates of the target crossing threshold point change, it is skipped. If the pixel coordinates of the target crossing threshold point remain unchanged, it is identified as an abnormal element, marked on the image and smoothed.

[0039] In step S5, the image coordinates of the candidate target point are converted into the payload body coordinate system vector according to the built-in payload imaging model. The payload body coordinate system vector is converted into an inertial system vector according to the inertial system pointing angle information of the payload at the imaging time, and then converted into an associated celestial coordinate system vector. Multi-frame judgment is performed on the associated candidate target point. If the change in the target celestial coordinate system vector is greater than the angular resolution corresponding to one pixel, it is skipped. If the change in the target celestial coordinate system vector is not greater than the angular resolution corresponding to one pixel, it is identified as a stellar target.

[0040] In step S6, outliers and star targets are eliminated, the remaining target trajectories are sorted by confidence, the target with the highest confidence is selected as the tracking target, the target trajectory is converted into an inertial system vector, the vector of the tracking target is used as the center, and step S2 is executed to update the target detection result and maintain circular motion control.

[0041] A high-confidence satellite-to-moving target acquisition and tracking system provided by the present invention comprises:

[0042] Module M1: Schedules the payload to point to the target based on the initial guidance information of the target, recursively calculates the predicted position and velocity information of the target, obtains the payload pointing information, and adjusts the payload pointing to the target;

[0043] Module M2: With the star vector as the center, the sinusoidal control variable is superimposed to control the target to perform circular motion and obtain the target trajectory;

[0044] Module M3: Extract target crossing threshold points, connect target crossing threshold points and calculate the center of gravity, form target candidate points and perform multi-frame association, and calculate confidence;

[0045] Module M4: Perform continuous frame judgment on target candidate points, identify target candidate points whose column positions remain unchanged in the image coordinate system as abnormal elements and remove them;

[0046] Module M5: Combined with the payload pointing information, the image row and column coordinates of the target candidate points after removing the outliers are converted to the celestial coordinate system. Through continuous frame judgment, the target candidate points whose positions remain unchanged in the celestial coordinate system are identified as star targets.

[0047] Module M6: Eliminate star targets, select the target with the highest confidence in the target trajectory as the tracking target, and update the target detection results in real time with the tracking target vector as the center.

[0048] Preferably, the module M1 predicts the predicted position and velocity information of the target through the orbital dynamics equation based on the initial guidance information of the target, guides the payload visual axis to point to the predicted position of the target, and guides the target motion trajectory from the edge of the payload's field of view to the payload's field of view. It stops when the angle between the payload's field of view center vector and the star sight vector is less than one-quarter of the payload's field of view angle, and the payload is pointed into place.

[0049] After the payload in the module M2 is pointed into place, the sinusoidal control quantity is superimposed on the image plane perpendicular to the star-eye vector with the star-eye vector as the center. The control center controls the target to move in a circular motion to obtain the target trajectory:

[0050]

[0051] Among them, (x k ,y k ,1) represents the control center vector that changes with time t;

[0052] x k Represents the x-direction component of the image plane control center;

[0053] y k Indicates the y-direction component of the image plane control center;

[0054] (x0,y0,1) represents the target vector;

[0055] x0 represents the x-direction component of the image target;

[0056] y0 represents the y-direction component of the image target;

[0057] r represents the circumferential control radius;

[0058] ω represents the circular control angular velocity.

[0059] Preferably, the module M3 subtracts the grayscale value of the current frame image from the grayscale value of the background differential frame interval frame image. If the differential grayscale value after subtraction is greater than or equal to 0, the differential grayscale value is retained as the differential pixel grayscale value. If the differential grayscale value is less than 0, the differential pixel grayscale value is set to 0:

[0060]

[0061] Among them, m represents the row vector of pixel grayscale;

[0062] n represents the column vector of pixel grayscale;

[0063] P k (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kth frame;

[0064] P k-d (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kd-th frame;

[0065] d represents the background difference frame interval;

[0066] P diff (m,n) represents the pixel grayscale value after dynamic background subtraction.

[0067] Connected domain association is performed based on the extracted target crossing threshold points. The connected target crossing threshold points are regarded as pixels belonging to the same target. The target center of gravity is calculated and used as the target candidate point for multi-frame association. The confidence of the target trajectory is obtained based on the consistency of the target trajectory and grayscale energy.

[0068] Preferably, the module M4 performs continuous frame judgment on each target crossing the threshold point. If the pixel coordinates of the target crossing the threshold point change, it is skipped. If the pixel coordinates of the target crossing the threshold point remain unchanged, it is identified as an abnormal element, marked on the image and smoothed.

[0069] In the module M5, the image coordinates of the target candidate point are converted into the payload body coordinate system vector based on the built-in payload imaging model. The payload body coordinate system vector is converted into an inertial system vector based on the inertial system pointing angle information of the payload at the imaging moment, and then converted into an associated celestial coordinate system vector. Multi-frame judgment is performed on the associated target candidate point. If the change in the target celestial coordinate system vector is greater than the angular resolution corresponding to one pixel, it is skipped. If the change in the target celestial coordinate system vector is not greater than the angular resolution corresponding to one pixel, it is identified as a stellar target.

[0070] In the module M6, abnormal elements and star targets are eliminated, the remaining target trajectories are sorted by confidence, the target with the highest confidence is selected as the tracking target, the target trajectory is converted into an inertial system vector, and the vector of the tracking target is used as the center to trigger the module M2 to update the target detection result and maintain circular motion control.

[0071] According to a computer-readable storage medium storing a computer program provided by the present invention, when the computer program is executed by a processor, the steps of the method for high-confidence capture and tracking of a moving target by a satellite are implemented.

[0072] A spacecraft provided according to the present invention includes the above-mentioned satellite high-confidence capture and tracking system for moving targets, or the above-mentioned computer-readable storage medium storing a computer program.

[0073] Compared with the prior art, the present invention has the following beneficial effects:

[0074] 1. The present invention uses a dynamic background difference method to remove slowly varying background fluctuations, abnormal elements, stars and other image interference sources in the image background, thereby improving target tracking stability and achieving high-confidence capture and tracking capabilities.

[0075] 2. The present invention adopts active circular control means and superimposes a sinusoidal control amount on the basis of the moving target observation vector to make the target show circular motion in the field of view, which is more practical for the capture and tracking of space moving targets.

[0076] 3. The present invention takes into account various interference sources in imaging payload image detection, and uses the target image plane motion characteristics and coordinate transformation to effectively identify false alarm sources such as image plane anomalies and stars, reducing the impact of various false alarm sources during target tracking and improving tracking confidence. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0078] Figure 1 Flowchart of high-confidence capture and tracking method for moving targets;

[0079] Figure 2 Schematic diagram of the target's motion trajectory in the payload image plane. DETAILED DESCRIPTION

[0080] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0081] In order to solve the problem of interference from image background fluctuations, image plane anomalies, stars and other interference sources on the continuous detection of targets during the tracking of space moving targets, and to improve the confidence of target detection and tracking stability, a high-confidence capture and tracking method for satellites to capture and track moving targets is proposed according to the present invention. It effectively eliminates the influence of various false alarm sources of targets in deep space background, realizes high-confidence capture and tracking of space moving targets, and improves the reliability of autonomous closed-loop tracking on board. Figure 1 For example, the specific steps are as follows:

[0082] Step S1: Schedule the space-based payload to point to the target according to the initial guidance information of the target, and recursively predict the position and velocity information of the target in real time, so that the payload's line of sight gradually approaches the predicted position of the target, and the target enters the payload's field of view and gradually remains stable.

[0083] In more preferred embodiments, step S1 includes:

[0084] by Figure 2 For example, based on the target's initial guidance information, i.e., the target's initial position and velocity, the orbital dynamics equation is used to predict the subsequent target's real-time position and velocity, and the payload's line of sight is guided gradually toward the target's predicted position. The target's motion trajectory is gradually guided from the edge of the field of view to the payload's field of view. When the angle between the payload's field of view center vector and the satellite-to-target vector, i.e., the star-eye vector, is less than one-quarter of the payload's field of view angle, the payload is considered to be in position.

[0085] Step S2: After the payload is pointed into place, with the star-eye vector as the control center, a sinusoidal control variable is superimposed on the image plane perpendicular to the star-eye vector to make the target exhibit circular motion on the image plane;

[0086] In more preferred embodiments, step S2 includes:

[0087] After the payload is pointed into place, the control center actively controls the target to move in a circular motion with the star vector as the center and the sinusoidal control quantity superimposed. The formula is as follows.

[0088]

[0089] Where: (x k ,y k ,1) is the control center vector that changes with time t, where x k is the x-direction component of the image plane, y kis the y-direction component of the image plane, (x0, y0, 1) is the guiding target vector, where x0 is the x-direction component of the image plane, y0 is the y-direction component of the image plane, r is the circumferential control radius, and ω is the circumferential control angular velocity.

[0090] Step S3: After the payload is pointed in place, the dynamic background difference method is used to extract the target threshold points, the threshold points are connected and the center of gravity is calculated to form the target candidate points, and multi-frame correlation detection is performed on the candidate points;

[0091] In more preferred examples, a dynamic background difference method is used to extract target crossing threshold points. Dynamic background difference is to subtract the grayscale value of the current frame image from the image before the fixed interval frame. For each pixel: if the grayscale value after subtraction is ≥0, the differential grayscale value is retained as the differential pixel grayscale value; if the grayscale value after subtraction is <0, the differential grayscale value is taken as 0. The formula is as follows.

[0092]

[0093] Where: m represents the row vector of pixel grayscale, n represents the column vector of pixel grayscale, P k (m,n) is the grayscale of the pixel with row and column number (m,n) in the kth frame, P k-d (m,n) is the grayscale of the pixel with row and column number (m,n) in the kdth frame, d is the background difference frame interval, P diff (m,n) is the pixel grayscale value after dynamic background difference.

[0094] An appropriate background difference frame interval is selected to ensure that, at circular motion speeds, target crossing threshold pixels (i.e., target crossing threshold points) do not overlap between frames. Connected domain association is performed based on target crossing threshold points extracted from the difference image. Connected target crossing threshold points are treated as pixels belonging to the same target, and the target centroid is calculated as a candidate target point. Multi-frame association is performed on the candidate target points, and the target trajectory confidence is calculated based on the local signal-to-noise ratio, anomaly similarity, and target historical state information, along with the target image motion (i.e., the obtained target trajectory) and grayscale energy consistency.

[0095] Step S4: identifying target candidate points with unchanged column positions in the image coordinate system as abnormal elements through continuous frame judgment;

[0096] In more preferred embodiments, step S4 includes:

[0097] Each candidate target point is judged in multiple frames. If the pixel coordinates of the candidate target point remain unchanged, it is identified as an outlier, marked on the image and smoothed. The method includes taking the average grayscale value of its four adjacent pixels as the grayscale of the pixel.

[0098] Step S5: Based on the payload pointing information, the image row and column coordinates of the remaining target candidate points after removing the abnormal elements are converted to celestial right ascension and declination, and the candidate points with unchanged positions in the celestial coordinate system are identified as star targets through continuous frame judgment;

[0099] In more preferred embodiments, step S5 includes:

[0100] Based on the payload imaging model, the image coordinates of the candidate target point are first converted to a vector in the payload's coordinate system. Then, based on the payload's inertial frame pointing angle at the time of imaging, the payload's coordinate system vector is converted to an inertial frame vector. The corresponding celestial right ascension and declination of the candidate target point are then determined. Multi-frame judgment is performed on the associated candidate target point. If the change in right ascension and declination corresponding to the target does not exceed the angular resolution corresponding to one pixel, the target is identified as a stellar target.

[0101] Step S6: Eliminate abnormal elements and stars, select the target with the highest confidence in the target trajectory as the tracking target, and use the tracking target vector as the center to update the target detection results in real time, maintain the circular control of the payload, and realize autonomous closed-loop tracking on the satellite under circular motion.

[0102] In more preferred embodiments, step S6 includes:

[0103] Eliminate outliers and stars, sort the remaining target trajectories by confidence, and select the target with the highest confidence as the tracking target. Convert the target image plane trajectory into an inertial vector. With the tracking target vector as the center, superimpose the sinusoidal control variable as described in step S2, update the target detection results in real time, maintain circular control of the payload, and achieve autonomous closed-loop tracking on the satellite under circular motion.

[0104] The present invention also provides a high-confidence capture and tracking system for a satellite to a moving target. The high-confidence capture and tracking system for a satellite to a moving target can be implemented by executing the process steps of the high-confidence capture and tracking method for a satellite to a moving target, that is, those skilled in the art can understand the high-confidence capture and tracking method for a satellite to a moving target as a preferred implementation of the high-confidence capture and tracking system for a satellite to a moving target.

[0105] A high-confidence satellite-to-moving target acquisition and tracking system provided by the present invention comprises:

[0106] Module M1: Schedules the payload to point to the target based on the initial guidance information of the target, recursively calculates the predicted position and velocity information of the target, obtains the payload pointing information, and adjusts the payload pointing to the target;

[0107] Module M2: With the star vector as the center, the sinusoidal control variable is superimposed to control the target to perform circular motion and obtain the target trajectory;

[0108] Module M3: Extract target crossing threshold points, connect target crossing threshold points and calculate the center of gravity, form target candidate points and perform multi-frame association, and calculate confidence;

[0109] Module M4: Perform continuous frame judgment on target candidate points, identify target candidate points whose column positions remain unchanged in the image coordinate system as abnormal elements and remove them;

[0110] Module M5: Combined with the payload pointing information, the image row and column coordinates of the target candidate points after removing the outliers are converted to the celestial coordinate system. Through continuous frame judgment, the target candidate points whose positions remain unchanged in the celestial coordinate system are identified as star targets.

[0111] Module M6: Eliminate star targets, select the target with the highest confidence in the target trajectory as the tracking target, and update the target detection results in real time with the tracking target vector as the center.

[0112] In more preferred examples, the module M1 predicts the predicted position and velocity information of the target based on the initial guidance information of the target through the orbital dynamics equation, guides the payload visual axis to point to the predicted position of the target, and guides the target motion trajectory from the edge of the payload's field of view to the payload's field of view. It stops when the angle between the payload's field of view center vector and the star sight vector is less than one-quarter of the payload's field of view angle, and the load is pointed into place.

[0113] After the payload in the module M2 is pointed into place, the sinusoidal control quantity is superimposed on the image plane perpendicular to the star-eye vector with the star-eye vector as the center. The control center controls the target to move in a circular motion to obtain the target trajectory:

[0114]

[0115] Among them, (x k ,y k ,1) represents the control center vector that changes with time t;

[0116] x k Represents the x-direction component of the image plane control center;

[0117] y k Indicates the y-direction component of the image plane control center;

[0118] (x0,y0,1) represents the target vector;

[0119] x0 represents the x-direction component of the image target;

[0120] y0 represents the y-direction component of the image target;

[0121] r represents the circumferential control radius;

[0122] ω represents the circular control angular velocity.

[0123] In more preferred examples, the module M3 subtracts the grayscale value of the current frame image from the grayscale value of the background differential frame interval frame image. If the differential grayscale value after subtraction is greater than or equal to 0, the differential grayscale value is retained as the differential pixel grayscale value. If the differential grayscale value is less than 0, the differential pixel grayscale value is set to 0:

[0124]

[0125] Among them, m represents the row vector of pixel grayscale;

[0126] n represents the column vector of pixel grayscale;

[0127] P k (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kth frame;

[0128] P k-d (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kd-th frame;

[0129] d represents the background difference frame interval;

[0130] P diff (m,n) represents the pixel grayscale value after dynamic background subtraction.

[0131] Connected domain association is performed based on the extracted target crossing threshold points. The connected target crossing threshold points are regarded as pixels belonging to the same target. The target center of gravity is calculated and used as the target candidate point for multi-frame association. The confidence of the target trajectory is obtained based on the consistency of the target trajectory and grayscale energy.

[0132] In more preferred examples, the module M4 performs continuous frame judgment on each target crossing the threshold point. If the pixel coordinates of the target crossing the threshold point change, it is skipped. If the pixel coordinates of the target crossing the threshold point remain unchanged, it is identified as an abnormal element, marked on the image and smoothed.

[0133] In the module M5, the image coordinates of the target candidate point are converted into the payload body coordinate system vector based on the built-in payload imaging model. The payload body coordinate system vector is converted into an inertial system vector based on the inertial system pointing angle information of the payload at the imaging moment, and then converted into an associated celestial coordinate system vector. Multi-frame judgment is performed on the associated target candidate point. If the change in the target celestial coordinate system vector is greater than the angular resolution corresponding to one pixel, it is skipped. If the change in the target celestial coordinate system vector is not greater than the angular resolution corresponding to one pixel, it is identified as a stellar target.

[0134] In the module M6, abnormal elements and star targets are eliminated, the remaining target trajectories are sorted by confidence, the target with the highest confidence is selected as the tracking target, the target trajectory is converted into an inertial system vector, and the vector of the tracking target is used as the center to trigger the module M2 to update the target detection result and maintain circular motion control.

[0135] According to a computer-readable storage medium storing a computer program provided by the present invention, when the computer program is executed by a processor, the steps of the method for high-confidence capture and tracking of a moving target by a satellite are implemented.

[0136] A spacecraft provided according to the present invention includes the above-mentioned satellite high-confidence capture and tracking system for moving targets, or the above-mentioned computer-readable storage medium storing a computer program.

[0137] Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, and units provided by the present invention in purely computer-readable program code, it is entirely possible to implement the same functions of the system and its various devices, modules, and units provided by the present invention in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; the devices, modules, and units for implementing various functions can also be considered as both software modules implementing the method and structures within the hardware component.

[0138] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. The embodiments of this application and the features in the embodiments may be combined with each other in any manner unless there is a conflict.

Claims

1. A method for capturing and tracking a moving target with high confidence by a satellite, characterized in that: include: Step S1: Schedule the payload to point to the target according to the initial guidance information of the target, recursively deduce the predicted position and speed information of the target, obtain the payload pointing information, and adjust the payload pointing to the target; Step S2: With the star vector as the center, the sinusoidal control variable is superimposed to control the target to perform circular motion and obtain the target trajectory; Step S3: extract target crossing threshold points, connect target crossing threshold points and calculate gravity center, form target candidate points and perform multi-frame association, and calculate confidence; Step S4: Perform continuous frame judgment on the target candidate points, identify the target candidate points whose column positions remain unchanged in the image coordinate system as abnormal elements and remove them; Step S5: Based on the payload pointing information, the image row and column coordinates of the target candidate points after removing the outliers are converted into the celestial coordinate system, and through continuous frame judgment, the target candidate points whose positions remain unchanged in the celestial coordinate system are identified as star targets; Step S6: Eliminate the star targets, select the target with the highest confidence in the target trajectory as the tracking target, and update the target detection results in real time with the tracking target vector as the center.

2. The method for capturing and tracking a moving target with high confidence by a satellite according to claim 1, wherein: In step S1, the target's predicted position and velocity are predicted using the orbital dynamics equation based on the target's initial guidance information, and the payload's visual axis is guided to point to the target's predicted position. The target's motion trajectory is guided from the edge of the payload's field of view to the payload's field of view. The target stops when the angle between the payload's field of view center vector and the star-eye vector is less than one-quarter of the payload's field of view angle, and the payload is pointed into position. After the payload is pointed into place in step S2, a sinusoidal control variable is superimposed on the image plane perpendicular to the star-eye vector with the star-eye vector as the center. The control center controls the target to move in a circular motion to obtain the target trajectory: Among them, (x k ,y k ,1) represents the control center vector that changes with time t; x k Represents the x-direction component of the image plane control center; y k Indicates the y-direction component of the image plane control center; (x0,y0,1) represents the target vector; x0 represents the x-direction component of the image target; y0 represents the y-direction component of the image target; r represents the circumferential control radius; ω represents the circular control angular velocity.

3. The method for capturing and tracking a moving target with high confidence by a satellite according to claim 1, wherein: In step S3, the grayscale values ​​of the current frame image and the background differential frame interval frame image are subtracted. If the differential grayscale value after subtraction is greater than or equal to 0, the differential grayscale value is retained as the differential pixel grayscale value. If the differential grayscale value is less than 0, the differential pixel grayscale value is set to 0: Among them, m represents the row vector of pixel grayscale; n represents the column vector of pixel grayscale; P k (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kth frame; P k-d (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kd-th frame; d represents the background difference frame interval; P diff (m,n) represents the pixel gray value after dynamic background difference; Connected domain association is performed based on the extracted target crossing threshold points. The connected target crossing threshold points are regarded as pixels belonging to the same target. The target center of gravity is calculated and used as the target candidate point for multi-frame association. The confidence of the target trajectory is obtained based on the consistency of the target trajectory and grayscale energy.

4. The method for capturing and tracking a moving target with high confidence by a satellite according to claim 1, wherein: In step S4, each target crossing threshold point is judged in consecutive frames. If the pixel coordinates of the target crossing threshold point change, it is skipped. If the pixel coordinates of the target crossing threshold point remain unchanged, it is identified as an abnormal element, marked on the image and smoothed. In step S5, the image coordinates of the candidate target point are converted to the payload body coordinate system vector according to the built-in payload imaging model. The payload body coordinate system vector is converted to an inertial system vector according to the inertial system pointing angle information of the payload at the imaging time, and then converted to an associated celestial coordinate system vector. The associated candidate target point is subjected to multi-frame judgment. If the target celestial coordinate system vector changes by more than the angular resolution corresponding to one pixel, it is skipped. If the target celestial coordinate system vector changes by less than the angular resolution corresponding to one pixel, it is identified as a stellar target. In step S6, outliers and star targets are eliminated, the remaining target trajectories are sorted by confidence, the target with the highest confidence is selected as the tracking target, the target trajectory is converted into an inertial system vector, the vector of the tracking target is used as the center, and step S2 is executed to update the target detection result and maintain circular motion control.

5. A high-confidence satellite tracking system for moving targets, characterized in that: include: Module M1: Schedules the payload to point to the target based on the initial guidance information of the target, recursively calculates the predicted position and velocity information of the target, obtains the payload pointing information, and adjusts the payload pointing to the target; Module M2: With the star vector as the center, the sinusoidal control variable is superimposed to control the target to perform circular motion and obtain the target trajectory; Module M3: Extract target crossing threshold points, connect target crossing threshold points and calculate the center of gravity, form target candidate points and perform multi-frame association, and calculate confidence; Module M4: Perform continuous frame judgment on target candidate points, identify target candidate points whose column positions remain unchanged in the image coordinate system as abnormal elements and remove them; Module M5: Combined with the payload pointing information, the image row and column coordinates of the target candidate points after removing the outliers are converted to the celestial coordinate system. Through continuous frame judgment, the target candidate points whose positions remain unchanged in the celestial coordinate system are identified as star targets. Module M6: Eliminate star targets, select the target with the highest confidence in the target trajectory as the tracking target, and update the target detection results in real time with the tracking target vector as the center.

6. The satellite-to-moving-target high-confidence acquisition and tracking system according to claim 5, characterized in that: In the module M1, the predicted position and velocity of the target are predicted by the orbital dynamics equation according to the initial guidance information of the target, and the payload visual axis is guided to point to the predicted position of the target, and the target motion trajectory is guided from the edge of the payload's field of view to the payload's field of view. When the angle between the payload's field of view center vector and the star sight vector is less than one-quarter of the payload's field of view angle, the target stops and the payload is pointed into place. After the payload in the module M2 is pointed into place, the sinusoidal control quantity is superimposed on the image plane perpendicular to the star-eye vector with the star-eye vector as the center. The control center controls the target to move in a circular motion to obtain the target trajectory: Among them, (x k ,y k ,1) represents the control center vector that changes with time t; x k Represents the x-direction component of the image plane control center; y k Indicates the y-direction component of the image plane control center; (x0,y0,1) represents the target vector; x0 represents the x-direction component of the image target; y0 represents the y-direction component of the image target; r represents the circumferential control radius; ω represents the circular control angular velocity.

7. The satellite-to-moving target high-confidence acquisition and tracking system according to claim 5, characterized in that: In the module M3, the grayscale value of the current frame image is subtracted from the grayscale value of the background differential frame interval frame image. If the differential grayscale value after subtraction is greater than or equal to 0, the differential grayscale value is retained as the differential pixel grayscale value. If the differential grayscale value is less than 0, the differential pixel grayscale value is set to 0: Among them, m represents the row vector of pixel grayscale; n represents the column vector of pixel grayscale; P k (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kth frame; P k-d (m,n) represents the grayscale of the pixel with row and column number (m,n) in the kd-th frame; d represents the background difference frame interval; P diff (m,n) represents the pixel gray value after dynamic background difference; Connected domain association is performed based on the extracted target crossing threshold points. The connected target crossing threshold points are regarded as pixels belonging to the same target. The target center of gravity is calculated and used as the target candidate point for multi-frame association. The confidence of the target trajectory is obtained based on the consistency of the target trajectory and grayscale energy.

8. The satellite-to-moving target high-confidence acquisition and tracking system according to claim 5, characterized in that: In the module M4, each target crossing the threshold point is judged in consecutive frames. If the pixel coordinates of the target crossing the threshold point change, it is skipped. If the pixel coordinates of the target crossing the threshold point remain unchanged, it is identified as an abnormal element, marked on the image and smoothed. In the module M5, the image coordinates of the candidate target point are converted to the payload body coordinate system vector according to the built-in payload imaging model. The payload body coordinate system vector is converted to an inertial system vector according to the inertial system pointing angle information of the payload at the imaging time, and then converted to an associated celestial coordinate system vector. The associated candidate target point is judged in multiple frames. If the change of the target celestial coordinate system vector is greater than the angular resolution corresponding to one pixel, it is skipped. If the change of the target celestial coordinate system vector is not greater than the angular resolution corresponding to one pixel, it is identified as a star target. In the module M6, abnormal elements and star targets are eliminated, the remaining target trajectories are sorted by confidence, the target with the highest confidence is selected as the tracking target, the target trajectory is converted into an inertial system vector, and the vector of the tracking target is used as the center to trigger the module M2 to update the target detection result and maintain circular motion control.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for high-confidence capture and tracking of a moving target by a satellite according to any one of claims 1 to 4 are implemented.

10. A spacecraft, characterized in that: The method comprises the satellite high-confidence capture and tracking system for moving targets according to any one of claims 5 to 8, or the computer-readable storage medium storing a computer program according to claim 9.

Images