Autonomous multi-target tracking method based on dynamic adjustment of position relation between targets

Abstract

The invention discloses an autonomous multi-target tracking method based on dynamic adjustment of a position relation between targets, and belongs to the field of remote sensing satellite task execution. According to the method, a plurality of targets in an imaging area are selected as follow-up tracking targets according to user requirements, and an initial tracking task is formed; and when each tracking target cannot be located in the same view field along with the movement in each direction, a new task is adaptively and dynamically split while the original task is continued to be executed. And the new task covers the target which cannot be continuously covered in the view field of the original task by adjusting the satellite imaging pointing area. And a plurality of imaging tasks are alternately executed, and imaging direction adjustment is carried out among the tasks, so that continuous coverage imaging and tracking of all tracking targets are realized. According to the invention, the efficiency of multiple target tracking tasks is improved, and the satellite target tracking capacity is improved.

Description

Technical Field

[0001] The present invention relates to an autonomous multi-target tracking method based on dynamic adjustment of the positional relationship between targets, belonging to the technical field of remote sensing satellite mission execution. Background Art

[0002] A remote sensing satellite can discover and locate targets within the imaging area by performing sequential imaging on a specified area and based on on-board data processing capabilities. After selecting a specified target, it can autonomously decide to adjust the satellite imaging pointing at an appropriate time so that the target is always within the imaging field of view, realizing autonomous tracking of the target.

[0003] However, when the number of targets that the user wants to track within the imaging area changes from single to multiple, the satellite becomes difficult to adapt. At the initial stage of the mission, all targets concerned by the user are within the imaging field of view, and at this time, the satellite can effectively obtain the positions and trajectories of each target. As time goes by, each target moves in different directions until all targets cannot be covered within the satellite imaging field of view. At this time, the basic processing solution is that the satellite only selects one of the targets and adjusts the imaging pointing according to its position to ensure that it is within the field of view. If other targets cannot be kept within the field of view, the ability to continue tracking is lost, and the user's multi-target tracking requirements cannot be met. An improved solution is that at the initial stage of the mission, a tracking task is established for each tracking target concerned by the user, and imaging is performed for each target in a time-sharing round-robin manner and its tracking is maintained. The disadvantage of this method is low efficiency, because when the distance between targets is relatively close, a single imaging can cover multiple targets at the same time. If independent imaging time periods are arranged for each target, the imaging interval of each target will be greatly lengthened, and target tracking requires associating the targets in the front and back imaging, and there are strict restrictions on the imaging interval, which results in a still small total number of tracking targets that this solution can support.

[0004] The above existing methods greatly restrict the multi-target tracking mission execution ability of the satellite and weaken the practicality of the current satellite autonomous tracking function. Summary of the Invention

[0005] The technical problem solved by the present invention is: overcoming the deficiencies of the prior art, an autonomous multi-target tracking method based on dynamic adjustment of the positional relationship between targets is proposed. During the multi-target tracking process, based on the positional relationship between targets, the number of tracking tasks for targets is dynamically adjusted to solve the problems of low target tracking capacity, poor adaptability and flexibility in the prior art, and improve the multi-target tracking efficiency of remote sensing satellites.

[0006] The technical solution of the present invention is:

[0007] An autonomous multi-target tracking method based on dynamic adjustment of the positional relationship between targets, comprising:

[0008] Step 1: The satellite performs continuous imaging on a specified area according to the remote sensing imaging mission to generate an image sequence. The on-board data processing module simultaneously performs on-orbit processing on the image sequence, detects all suspected targets within the imaging area, completes the positioning calculation of the suspected targets, and completes the trajectory association of the targets in the sequence images. The suspected target information is transmitted to the ground.

[0009] Step 2: The ground user selects m targets of interest from all suspected targets as tracking targets and sends a tracking task for tracking all the tracking targets to the satellite.

[0010] Step 3: The satellite receives the tracking task, continuously images and classifies and marks the detected targets. The first category is the tracking targets selected by the user, and the second category is the remaining targets, denoted as general targets. At the same time, the satellite continuously calculates the distance of each tracking target from the edge of the satellite imaging field of view: The area in the imaging area with a distance greater than the tracking task adjustment threshold P from the imaging edge is defined as the stable area, and the remaining area of the imaging area is defined as the adjustment area. When all tracking targets are located within the stable area, the satellite continuously images to detect the targets and calculates the position and trajectory information of each target to achieve continuous tracking of each tracking target.

[0011] Step 4: If any one of the tracking targets enters the adjustment area within the corresponding imaging area of the task, trigger the tracking task adjustment and calculate the relationship between the position of the target that has entered the adjustment area and all task imaging areas:

[0012] If there is only one tracking task currently, execute Step 7;

[0013] If there are multiple tracking tasks currently, perform the following 1) - 3) processing:

[0014] 1) If the target that has entered the adjustment area is located in the stable area of other task imaging areas currently, execute Step 5;

[0015] 2) If the target that has entered the adjustment area is not in the stable area of all task imaging areas, and if the imaging pointing of the task to which the target currently belongs is adjusted with the position of the target as the center, it can ensure that all tracking targets within the task are located in the stable area of the new imaging area, execute Step 6;

[0016] 3) If the target that has entered the adjustment area is not in the stable area of all task imaging areas, and if the imaging pointing of the task to which the target currently belongs is adjusted with the position of the target as the center, not all tracking targets within the task are in the stable area, execute Step 7;

[0017] Step 5: In the target's current task that enters the adjustment area in Step 4, change the target type of the target to a general target; select any task where the target is located in the stable area of its imaging region, change the target type of the target in the task to a tracking target, and then execute Step 3;

[0018] Step 6: Adjust the pointing of the current task to which the target that enters the adjustment area in Step 4 belongs. After adjustment, the imaging region is centered on the target position, and all tracking targets in the task are located in the new imaging region. Based on the position information of each target calculated before the adjustment pointing, complete the target association of each tracking target in the new imaging region, and then execute Step 3;

[0019] Step 7: Without changing the imaging region of the task to which the target that enters the adjustment area belongs, perform dynamic task splitting, that is, the satellite autonomously adds a new task. The satellite imaging center of the new task is set to the position of the tracking target that enters the adjustment area, and then execute Step 3.

[0020] Further, in Step 5, after the target that enters the adjustment area is adjusted to other tasks, if there are no tracking targets in the original task, cancel the original task to achieve dynamic task merging and release satellite resources.

[0021] Further, in Step 3, when there are multiple tasks, the satellite autonomously allocates the imaging time periods of all tasks and alternately executes each task, so that each task takes turns to occupy the satellite time to carry out imaging and data processing tasks.

[0022] Further, to ensure the necessary inter-scene target association for target tracking, the interval between each round of execution of any task needs to be less than 15 minutes.

[0023] Further, in Step 3, if the same target is marked as a tracking target in different tasks, change the mark of any one of the tasks to a general target, and only one target remains as a tracking target in only one task.

[0024] Further, in Step 3, the tracking task adjustment threshold P is preferably 5% - 30% of the imaging field of view.

[0025] Further, imaging includes imaging of various remote sensing satellite payload systems, including but not limited to optical imaging and SAR imaging.

[0026] Further, task settings include but are not limited to imaging time, imaging interval, imaging pointing position, and task duration; among them, the imaging interval setting needs to comprehensively consider the processing ability of the on-board data processing module and the requirement for achieving inter-scene association of the target.

[0027] Further, the satellite uses object detection algorithms for object detection, and the object detection algorithms include but are not limited to Yolo, SSD, and RetinaNet algorithms.

[0028] Further, the satellite uses object tracking algorithms to complete object trajectory association and subsequent tracking, and the object tracking algorithms include but are not limited to Kalman filtering and particle filtering algorithms.

[0029] The advantages of the present invention compared with the prior art are as follows:

[0030] (1) According to the relative motion and position relationship of each tracking target, when the tracking targets cannot be in the same field of view due to their respective movements, while the on-board data processing module continues to execute the original task, it will adaptively and dynamically split out a new task. The new task adjusts the satellite imaging pointing area to cover the targets that can no longer be covered within the field of view of the original task. The two imaging tasks are executed alternately, and the imaging pointing is adjusted between the tasks, so as to achieve continuous coverage imaging and tracking of all tracking targets. By analogy, according to the relative motion relationship of each tracking target, the number of tasks can be further split into more to achieve flexible and continuous tracking of each tracking target.

[0031] (2) According to the relative motion and position relationship of each tracking target, when the tracking targets have a tendency to move closer to each other, it is also possible to achieve adaptive dynamic task merging under the premise of ensuring coverage of all targets. While ensuring effective tracking of all targets, the total number of targets is reduced, the resources of the satellite are released, and the interval of each task is correspondingly shortened, which is beneficial to better complete the association of targets between scenes.

[0032] (3) The present invention avoids the problem of wasting satellite resources and low tracking target capacity caused by allocating a task to each target, and utilizes the characteristic of covering multiple targets with one imaging to achieve high efficiency of tracking multiple targets with a small number of tasks.

[0033] (4) The present invention is no longer a multi-target tracking task execution method that fixedly allocates resources at the initial stage of a task. Not only can the task be adaptively and dynamically adjusted according to the movement position of the target, but it also supports users to adjust and set the tracking targets, greatly improving the flexibility of task execution. Description of the Drawings

[0034] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0035] Figure 1Flowchart of the autonomous multi-target tracking method for remote sensing satellites based on dynamic adjustment of the positional relationship between targets in the embodiments of the present invention;

[0036] Figure 2 Schematic diagram of the dynamic splitting of tasks during the multi-target tracking process in the embodiments of the present invention;

[0037] Figure 3 Schematic diagram of the dynamic merging of tasks during the multi-target tracking process in the embodiments of the present invention;

[0038] Figure 4 Schematic diagram of the dynamic adjustment of task pointing during the multi-target tracking process in the embodiments of the present invention. Detailed implementation manners

[0039] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0040] The present invention proposes an autonomous multi-target tracking method based on dynamic adjustment of the positional relationship between targets. For all targets discovered by the on-board data processing module in the imaging area of the remote sensing satellite, multiple of them can be selected as subsequent tracking targets according to user requirements to form an initial tracking task. When the tracking targets cannot be in the same field of view due to their respective movements in various directions, the on-board data processing module will adaptively and dynamically split out a new task while continuing to execute the original task. The new task adjusts the satellite imaging pointing area to cover the targets that can no longer be covered within the field of view of the original task. The two imaging tasks are executed alternately, and the imaging pointing is adjusted between the tasks, so as to achieve continuous coverage imaging and tracking of all tracking targets. By analogy, according to the relative motion relationship of each tracking target, the number of tasks can be further split into more; and when the tracking targets have a tendency to move closer to each other, some tasks can also be merged on the premise of ensuring coverage of all targets, reducing the task burden and the imaging interval of each area.

[0041] Through the above autonomous multi-target tracking method that dynamically splits, merges, and adjusts tracking tasks based on the positional relationship between targets, the efficiency of multiple target tracking tasks can be improved, the satellite target tracking capacity can be increased, and the application efficiency of the satellite can be enhanced. Among them, the satellite imaging areas of each target tracking task are different, and each tracking task includes one or more tracking targets.

[0042] This method is as Figure 1 shown and specifically includes the following steps:

[0043] Step 1: According to the user's requirements, the satellite performs continuous imaging tasks on the specified area. The imaging includes the imaging of various remote sensing satellite payload systems, such as optical (visible light, infrared, hyperspectral, etc.), SAR, etc. The task settings should include imaging time, imaging interval, imaging pointing position, task duration, etc. Among them, the imaging interval setting should comprehensively consider the processing capacity of the on-board data processing module and the need for scene-to-scene correlation of the target. It is preferably less than 15 minutes. In this embodiment, it is set to 1 minute, corresponding to imaging once per minute.

[0044] The on-board data processing module synchronously completes the on-orbit processing of the image sequence, discovers m targets (m is an integer greater than or equal to 1, and in this embodiment, it is assumed that m = 10) in the imaging area, and completes the positioning calculation and association of the target trajectories in the sequence images. Transmit each target information (including but not limited to target number, target position information) to the ground user.

[0045] Step 2: The user selects n targets (n is an integer greater than or equal to 1, n ≤ m, and in this embodiment, it is assumed that n = 3) of concern from the m targets discovered by the satellite as the subsequent tracking targets. The user sends the tracking targets to the satellite. The satellite continues to image and perform on-board data processing. The on-board data processing module classifies the targets it discovers into 2 categories. The first category is the tracking targets uploaded by the user (3 in total in this embodiment), and the second category, that is, the remaining targets, are marked as general targets (7 in this embodiment). The on-board data processing module will calculate the position, trajectory, and other information of all targets, but only the tracking targets will trigger the imaging pointing adjustment action of the satellite in the subsequent steps.

[0046] Among them, the on-board data processing module needs to use target detection algorithms, including but not limited to various algorithms suitable for target detection such as Yolo, SSD, RetinaNet, etc., to complete target discovery; and needs to use target tracking algorithms, including but not limited to algorithms such as Kalman filtering and particle filtering, to complete target trajectory association and subsequent tracking.

[0047] Step 3: The satellite continuously calculates the distance of each tracking target from the edge of the satellite imaging field of view. Set the tracking task adjustment threshold P (the set value can be adjusted as needed). Define the area in the imaging area with a distance greater than P from the imaging edge as the stable area; define the area in the imaging area with a distance less than or equal to P from the imaging edge as the adjustment area.

[0048] When all tracking targets are located in the stable area, continue to execute Step 3 to maintain the covering imaging and processing calculation of all tracking targets;

[0049] When any one tracking target (defined as O x ) enters the adjustment area, it triggers the tracking task adjustment: Based on the calculation and judgment, if O xAfter adjusting the pointing centered on O, if all the tracking targets can be located in the stable area within the new imaging area, then step 4 is executed; if after adjusting the pointing centered on O x not all the tracking can be located in the stable area within the imaging area, step 5 is executed.

[0050] Step 4, as Figure 4 shown, adjust the satellite imaging pointing so that the center of the imaging field of view is adjusted to the target O x position, and based on the position information of each target calculated before adjusting the pointing, complete the target association of each tracking target within the new imaging area.

[0051] It should be noted that in steps 3 and 4, the tracking of the target needs to ensure that the target is always within the imaging area. As the target moves and time goes by, the target may gradually approach the edge of the satellite imaging field of view and leave the field of view. It is necessary to adjust the satellite imaging pointing before the target leaves the field of view and adjust the center of the imaging area back to near the target position. Therefore, a tracking task adjustment threshold P is defined. The set value of P is preferably 5% - 30% of the imaging field of view. The area with a distance less than or equal to P from the edge of the field of view is defined as the adjustment area, and other areas are defined as the stable area. When the target enters the adjustment area, it means that the target is about to leave the field of view and needs to be adjusted in time.

[0052] Step 5, as Figure 2 shown, without changing the imaging area (defined as A1) of the original task (defined as T1), perform task dynamic splitting, that is, the satellite autonomously adds 1 new task T2, and the satellite imaging center of T2 is set to the position where the tracking target O x is located (the imaging area is defined as A2). The satellite autonomously allocates the imaging time periods of T1 and T2, alternately executes tasks T1 and T2, and adjusts the satellite imaging pointing between tasks T1 and T2 to achieve alternate imaging between areas A1 and A2. In this embodiment, an autonomous allocation method of evenly distributing time is adopted, evenly distributing the time of T1 and T2. Each round, T1 occupies 1 minute and T2 occupies 1 minute, and the polling cycle is 2 minutes. This autonomous allocation method of evenly distributing time allocates equal time for each tracking task and in a fixed order, which can ensure that in each task, no task has a particularly long execution interval. A longer interval may cause the target to move too far and miss the target. Therefore, the most ideal situation is to image alternately in a fixed order and evenly distribute the time.

[0053] In task T2, after being processed by the on-board data processing module, multiple targets are found in area A2. Among them, the targets found in the overlapping area of A1 and A2 can also be found in task T1, including the target O x . Based on the calculation of the target position information, complete the association of the same target in different imaging areas. To prevent O xIs double-counted by two tasks, T1 and T2, for pointing adjustment. In task T1, change the type of O x from a tracked target marker to a general target; in task T2, mark O x as a tracked target, and mark other targets found in task T2 as general targets.

[0054] Among them, through the calculation based on the positional relationship between targets and their relationship with the imaging area, and through the dynamic splitting of tasks, the problem of tracking multiple moving and separating targets is solved. When there are multiple tracking tasks, time allocation is required so that each task takes turns occupying satellite time to carry out imaging and data processing tasks. To ensure the necessary inter-view target association for target tracking, the interval of any task in each round of execution should be less than 15 minutes.

[0055] Step 6, continuously execute tasks T1 and T2 in turn. When any one of the tracked targets (defined as O y ) in tasks T1 and T2 enters the adjustment area within the imaging area corresponding to its task, trigger the tracking task adjustment, and calculate the relationship between the position of O y and all task imaging areas (A1, A2, etc.):

[0056] If O y is located in the adjustment area within the imaging area of the current task (defined as T i ), but O y is also located in the stable area of the imaging area of other tasks, execute step 7;

[0057] If O y is not in the stable area of all task imaging areas, and O y the current task T i If adjusting the imaging pointing with the position of O y as the center can ensure that all tracked targets within the task are located in the stable area of the new imaging area, then execute step 8;

[0058] If O y is not in the stable area of all task imaging areas, and O y the current task T i If adjusting the imaging pointing with the position of O y as the center, not all tracked targets within the task are in the stable area, then execute step 9.

[0059] Step 7, as Figure 3 shown, in the current task T y of O i , change the target type of O y from a tracked target to a general target; select any one of the tasks T y where O is located in the stable area of its imaging area T j, change the target type of O in the task from a general target to a tracking target. That is, for target O y the change of the task to which it belongs is completed, and the continuous tracking of O is maintained without adding new tasks or adjusting the pointing. y If O y is adjusted to another task, and task T y no longer has a tracking target, then the subsequent T i task is cancelled, the dynamic merging of tasks is realized, and satellite resources are released. i Based on the calculation of the positional relationship between targets and the relationship with the imaging area, through the dynamic merging of tasks, while ensuring the effective tracking of all targets, the total number of targets is reduced, satellite resources are released, and the interval of each task is correspondingly shortened, which is conducive to better completing the association of targets between scenes.

[0060]

[0061] Step 8, the current task T y to which O belongs adjusts the pointing, and after adjustment, the imaging area is centered on the position of O i . All tracking targets in T y are located in the new imaging area, and the tracking task continues to be executed. i

[0062] Step 9, execute Step 5, a new task is added centered on the position of O y to achieve the tracking of O y . The original task T i continues to maintain the tracking of other targets within its task.

[0063] Step 10, continuously execute Steps 6, 7, 8, and 9, and adaptively and dynamically adjust (including splitting, merging, and pointing adjustment) each tracking task according to the relative motion and positional relationship of each tracking target to achieve flexible and continuous tracking of each tracking target.

[0064] During the target tracking process, the user can adjust the type of the target (from a general target to a tracking target or vice versa) according to the needs, which improves the flexibility of the tracking task.

[0065] The above-described embodiments are only relatively preferred specific embodiments of the present invention, and the common changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included in the protection scope of the present invention.​​

Claims

1. An autonomous multi-target tracking method based on dynamic adjustment of the positional relationship between targets, characterized in that Including: Step 1: The satellite continuously images a specified area according to the remote sensing imaging mission, generating an image sequence; The on-board data processing module synchronously performs on-orbit processing on the image sequence, detects all suspected targets within the imaging area, completes the positioning calculation of the suspected targets, and completes the trajectory association of the targets in the sequence images; Transmit the suspected target information to the ground; Step 2: The ground user selects m targets of concern from all suspected targets as tracking targets, and sends the tracking task of tracking all the tracking targets to the satellite; Step 3: The satellite receives the tracking task, continuously images and classifies and marks the detected targets. The first category is the tracking targets selected by the user, and the second category is the remaining targets, denoted as general targets. At the same time, the satellite continuously calculates the distance of each tracking target from the edge of the satellite imaging field of view: Define the area in the imaging area with a distance greater than the tracking task adjustment threshold P from the imaging edge as the stable area, and define the remaining area of the imaging area as the adjustment area. When all tracking targets are located within the stable area, the satellite continuously images to detect the targets and calculates the position and trajectory information of each target, realizing the continuous tracking of each tracking target; Step 4: If any one of the tracking targets enters the adjustment area within the corresponding imaging area of the task, trigger the tracking task adjustment, and calculate the relationship between the position of the target that currently enters the adjustment area and all task imaging areas: If there is only one tracking task currently, execute Step 7; If there are multiple tracking tasks currently, perform the following 1)-3) processing: 1) If the target that currently enters the adjustment area is located in the stable area of other task imaging areas, execute Step 5; 2) If the target that currently enters the adjustment area is not in the stable area of all task imaging areas, and if the task to which the target currently belongs can ensure that all tracking targets within the task are located in the stable area of the new imaging area after adjusting the imaging pointing with the target position as the center, execute Step 6; 3) If the target that currently enters the adjustment area is not in the stable area of all task imaging areas, and if the task to which the target currently belongs, after adjusting the imaging pointing with the target position as the center, not all tracking targets within the task are in the stable area, execute Step 7; Step 5: Change the target type of the target in the task to which the target that enters the adjustment area in Step 4 belongs to a general target; Select any task where the target is located in the stable area of its imaging area, change the target type of the target in the task to a tracking target, and then execute Step 3; Step 6: Adjust the pointing of the task to which the target that enters the adjustment area in Step 4 belongs. After adjustment, the imaging area is centered on the target position, and all tracking targets within the task are located in the new imaging area. Based on the position information of each target calculated before the adjustment of the pointing, complete the target association of each tracking target in the new imaging area, and then execute Step 3; Step 7: Without changing the imaging area of the task to which the target that enters the adjustment area belongs, perform task dynamic splitting, that is, the satellite autonomously adds a task, and set the satellite imaging center of the new task to the position of the tracking target that enters the adjustment area, and then execute Step 3.

2. The autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, wherein In step 5, after the target entering the adjustment area is adjusted to other tasks, if there is no tracking target for the original task, the original task is cancelled to achieve dynamic merging of tasks and release satellite resources.

3. The autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, wherein, In step 3, when there are multiple tasks, the satellite autonomously allocates the imaging time periods of all tasks and alternately executes each task, so that each task takes turns to occupy the satellite time to carry out imaging and data processing tasks.

4. An autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 3, characterized in that, To ensure the necessary inter-scene target association for target tracking, the interval of any task in each round of execution needs to be less than 15 minutes.

5. The autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, wherein In step 3, if the same target is marked as a tracking target in different tasks, the mark of any one of the tasks is changed to a general target, and only one target remains as the tracking target in only one task.

6. The autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, wherein, In step 3, the tracking task adjustment threshold P is preferably 5% - 30% of the imaging field of view.

7. The autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, wherein Imaging includes imaging by various remote sensing satellite payload systems, including but not limited to optical imaging and SAR imaging.

8. An autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, characterized in that, Task settings include but are not limited to imaging time, imaging interval, imaging pointing position, and task duration; among them, the imaging interval setting needs to comprehensively consider the processing capacity of the on-board data processing module and the requirement for achieving inter-scene association of the target.

9. The autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, wherein, The satellite uses a target detection algorithm for target detection, and the target detection algorithm includes but is not limited to the Yolo, SSD, and RetinaNet algorithms.

10. The autonomous multi-target tracking method based on dynamic adjustment according to the positional relationship between targets as claimed in claim 1, wherein The satellite uses a target tracking algorithm to complete target trajectory association and subsequent tracking, and the target tracking algorithm includes but is not limited to the Kalman filter and particle filter algorithms.

Images