Track association method and device based on multi-dimensional feature clustering in parameter space

Abstract

The present invention provides a track association method and device based on parameter space multidimensional feature clustering, which relates to the field of image processing technology, and is used to solve the technical problems that the traditional Hough transform cannot utilize spatiotemporal multidimensional features in the process of track association application, and the peak extraction is difficult due to the peak clustering phenomenon. The method includes: obtaining simulated observation data of the motion trajectory of the space target object, and adding Poisson distribution clutter to generate a data set; converting the data set from the observation space to the parameter space; superimposing the position information of each space target object in the parameter space on its position information in the observation space to obtain a multidimensional feature data set; using a pre-trained multidimensional spatiotemporal feature density clustering model, clustering to obtain coordinate parameters located on the same motion trajectory; reversely converting the coordinate parameters into trajectory parameters in the observation space; based on the trajectory parameters, associating the space target objects that meet the preset motion trajectory trend to obtain associated arc segments.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and can be applied to the field of machine learning and artificial intelligence technology. More specifically, it relates to a track association method and device based on multi-dimensional feature clustering in parameter space. Background Art

[0002] Since the first artificial satellite was launched, human beings have been exploring space for more than 60 years. So far, there have been more than 8,000 launches and more than 12,000 satellites have been sent into space orbit. With frequent space activities, space objects have spread all over the Earth's orbit. Therefore, effective monitoring of space objects has important security significance.

[0003] The detection methods of near-Earth orbit targets are mainly divided into radar detection systems and photoelectric detection systems according to the detection principles and equipment. At present, photoelectric equipment is the main means of detecting medium and high orbit space debris, near-Earth objects and other space objects. Sky survey observations by photoelectric detection systems are the main observation mode for discovering unknown targets. They can discover uncataloged space debris and near-Earth asteroids without prior information. The sky survey observation mode faces complex astronomical observation scenes, but due to the complex astronomical observation background, a set of data often faces multiple types of multiple space objects, and the signal-to-noise ratio of the target is relatively low. The extracted candidate target set still includes a certain number of stars, comets, abnormal noise and other interference signals, which increases the difficulty of target detection.

[0004] Image processing in the sky survey observation mode generally includes: image preprocessing, stellar background extraction and space target detection, and target track association. Among them, the image preprocessing and target extraction processes already have relatively mature algorithm flows, and stellar background extraction and space target detection can generally be implemented based on mature astronomical software. The track association process is to achieve: from the candidate data set containing multiple types and multiple targets and with strong noise obtained by target detection from the time series image, the time series data points belonging to the same target are associated to form arc segments. The arc segments generated by track association are the data basis for the subsequent processes of orbit determination, cataloging, and early warning of space targets, but there is currently no relevant mature technology to implement this process, especially the track association of multiple types and multiple targets (referred to as multi-targets) in complex astronomical observation scenarios in the sky survey mode, which is still a research hotspot and a technical difficulty that needs to be solved in the field of target detection. Summary of the invention

[0005] In view of this, the present invention provides a track association method and device based on parameter space multidimensional feature clustering, which is used to solve the technical problems that traditional Hough transform cannot utilize spatiotemporal multidimensional features in the track association application process, and the peak extraction is difficult due to the peak clustering phenomenon.

[0006] One aspect of the present invention provides a track association method based on parameter space multidimensional feature clustering, comprising: obtaining simulated observation data of the motion trajectory of a space target object, and adding Poisson distribution clutter to the simulated observation data to generate a data set, wherein the Poisson distribution clutter is used to simulate the interference clutter of the space target object in a real observation environment; converting the data set from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space, wherein the distribution characteristics represent the set of all possible motion trajectories of each space target object in the observation space; based on the distribution characteristics, superimposing the position information of each space target object in the parameter space onto its position information in the observation space to obtain a multidimensional feature data set; according to the multidimensional feature data set, using a pre-trained multidimensional spatiotemporal feature density clustering model, clustering to obtain the coordinate parameters corresponding to the space target objects located on the same motion trajectory; reversely converting the coordinate parameters into trajectory parameters of the space target object in the observation space, wherein the trajectory parameters represent the slope and intercept of the motion trajectory of the space target object; based on the trajectory parameters, associating the space target objects in the observation space that meet the preset motion trajectory trend to obtain associated arc segments.

[0007] According to an embodiment of the present invention, a data set is converted from an observation space to a parameter space to obtain distribution characteristics of each space target object in the parameter space, including: using a linear slope-intercept transformation function to convert the observation coordinates of the space target object in the observation space into parameter coordinates in the parameter space; based on the parameter coordinates, a sinusoidal curve corresponding to each space target object in the parameter space is generated, wherein the sinusoidal curves corresponding to the space target objects located on the same motion trajectory have local intersections.

[0008] According to an embodiment of the present invention, the straight line slope-intercept transformation function is:

[0009]

[0010] in,( , ) represents the observation coordinates, ( , ) represents the parameter coordinates, ∈(0,π).

[0011] According to an embodiment of the present invention, based on a multidimensional feature data set, using a pre-trained multidimensional spatiotemporal feature density clustering model, before clustering to obtain the coordinate parameters corresponding to the spatial target objects located on the same motion trajectory, it includes: extracting the time information, morphological information and brightness information of each spatial target object in the data set; superimposing the time information, morphological information and brightness information on the multidimensional feature data set to obtain a multidimensional feature array; wherein the feature dimensions of each spatial target object in the multidimensional feature array include: position information in the observation space, position information in the parameter space, time information, morphological information and brightness information.

[0012] According to an embodiment of the present invention, based on a multidimensional feature data set, using a pre-trained multidimensional spatiotemporal feature density clustering model, clustering to obtain coordinate parameters corresponding to spatial target objects located on the same motion trajectory includes: based on a multidimensional feature array, using a pre-trained multidimensional spatiotemporal feature density clustering model, extracting coordinate values ​​corresponding to intersection points in a sine curve; wherein the coordinate values ​​corresponding to the intersection points represent coordinate parameters corresponding to spatial target objects located on the same motion trajectory.

[0013] According to an embodiment of the present invention, based on a multidimensional feature array and using a pre-trained multidimensional spatiotemporal feature density clustering model, the coordinate values ​​corresponding to the intersection points in the sine curve are extracted, including: calculating the multidimensional Euclidean distance between each pair of spatial target objects, wherein the multidimensional Euclidean distance characterizes the similarity of the multidimensional features of each pair of spatial target objects; preliminarily dividing the multidimensional feature array according to the multidimensional Euclidean distance to obtain a plurality of potential subclusters; within each potential subcluster, recalculating the local density of each spatial target object according to the neighborhood radius and the minimum number of samples; selecting spatial target objects above a density threshold as core points, and expanding density-connected spatial target objects to form clusters, wherein the spatial target objects in the same cluster are located on the same motion trajectory; and obtaining coordinate parameters corresponding to the spatial target objects located on the same motion trajectory according to the clusters.

[0014] According to an embodiment of the present invention, the multi-dimensional Euclidean distance is calculated as follows:

[0015]

[0016] in, Represents data points and The multidimensional Euclidean distance between The distance between two data points ( ) dimension, represents the time constraint function, Represents the multidimensional feature weight.

[0017] According to an embodiment of the present invention, the reverse conversion of coordinate parameters into trajectory parameters of the space target object in the observation space includes: solving the coordinate values ​​corresponding to the intersection points according to the straight line slope-intercept transformation function to obtain the trajectory parameters of the space target object in the observation space.

[0018] According to an embodiment of the present invention, based on trajectory parameters, spatial target objects in the observation space that meet preset motion trajectory trends are associated to obtain associated arc segments, including: matching all spatial target objects in the observation space with candidate motion trajectories to obtain an associated point data set corresponding to each candidate motion trajectory; in response to the number of spatial target objects in the associated point data set corresponding to each candidate motion trajectory exceeding a threshold, the corresponding candidate motion trajectory is successfully associated to obtain an associated arc segment.

[0019] Another aspect of the present invention provides a track association device based on multidimensional feature clustering in parameter space, comprising: an acquisition module, used to acquire simulated observation data of the motion trajectory of a space target object, and add Poisson distribution clutter to the simulated observation data to generate a data set, wherein the Poisson distribution clutter is used to simulate the interference clutter of the space target object in a real observation environment; a conversion module, used to convert the data set from the observation space to the parameter space, and obtain the distribution characteristics of each space target object in the parameter space, wherein the distribution characteristics represent the set of all possible motion trajectories of each space target object in the observation space; a superposition module, used to superimpose each space target object based on the distribution characteristics. The position information of the spatial target object in the parameter space is superimposed on its position information in the observation space to obtain a multidimensional feature data set; a clustering module is used to cluster the coordinate parameters corresponding to the spatial target objects located on the same motion trajectory according to the multidimensional feature data set using a pre-trained multidimensional spatiotemporal feature density clustering model; a reverse conversion module is used to reversely convert the coordinate parameters into trajectory parameters of the spatial target object in the observation space, wherein the trajectory parameters represent the slope and intercept of the motion trajectory of the spatial target object; an association module is used to associate the spatial target objects that meet the preset motion trajectory trend in the observation space based on the trajectory parameters to obtain associated arc segments.

[0020] Another aspect of the present invention provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the above method.

[0021] Another aspect of the present invention provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the above method when executed.

[0022] Another aspect of the present invention provides a computer program product, the computer program product comprising computer executable instructions, and the instructions are used to implement the above method when being executed.

[0023] Compared with the prior art, the track association method and device based on parameter space multidimensional feature clustering provided by the present invention have at least the following beneficial effects:

[0024] (1) The track association method and device based on parameter space multidimensional feature clustering provided by the present invention are different from traditional parameter space conversion methods (such as the classic Hough transform). Instead, the method converts the peak extraction problem with low accuracy into a multidimensional feature clustering problem, and clusters the multidimensional features using a multidimensional spatiotemporal feature density clustering model (MDST-DBSCAN model), thereby solving the technical problem of peak extraction difficulties caused by track clustering. To a certain extent, the method has a certain degree of adaptability to the discretization interval value of the parameter space, and is more suitable for engineering applications.

[0025] (2) Compared with the traditional track association method which only considers the location information of the data set but ignores the multi-dimensional spatiotemporal characteristics of the data set, the track association method and device provided by the present invention introduces multi-dimensional feature information including the time characteristics, morphological characteristics and brightness characteristics of the space target objects to participate in clustering, thereby solving the problem of local area observation attribution caused by the proximity and intersection of multiple target tracks, making the track association result more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0027] Figure 1 A flowchart of a track association method based on multi-dimensional feature clustering in parameter space according to an embodiment of the present invention is schematically shown;

[0028] Figure 2 A schematic diagram showing a principle diagram of obtaining a multi-dimensional feature array by superimposing multi-dimensional feature information according to an embodiment of the present invention;

[0029] Figure 3 A diagram schematically shows a result of clustering using a multi-dimensional spatiotemporal feature density clustering model according to an embodiment of the present invention;

[0030] Figure 4 The arc segment diagram obtained by successfully associating the track association method based on multi-dimensional feature clustering in parameter space according to an embodiment of the present invention is schematically shown;

[0031] Figure 5A structural block diagram of a track association device based on parameter space multi-dimensional feature clustering according to an embodiment of the present invention is schematically shown;

[0032] Figure 6 The structural block diagram of an electronic device suitable for implementing a track association method based on multi-dimensional feature clustering in parameter space according to an embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0033] Below, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of concepts of the present invention.

[0034] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0035] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0036] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0037] In the embodiments of the present invention, the collection, updating, analysis, processing, use, transmission, provision, disclosure, storage and other aspects of the data involved (for example, including but not limited to user personal information) are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information and network security.

[0038] Since the first artificial satellite was launched, human beings have been exploring space for more than 60 years. So far, there have been more than 8,000 launches and more than 12,000 satellites have been sent into space orbit. With frequent space activities, space objects have spread all over the Earth's orbit. Therefore, effective monitoring of space objects has important security significance.

[0039] The detection methods of near-Earth orbit targets are mainly divided into radar detection systems and photoelectric detection systems according to the detection principles and equipment. At present, photoelectric equipment is the main means of detecting medium and high orbit space debris, near-Earth objects and other space objects. Sky survey observations by photoelectric detection systems are the main observation mode for discovering unknown targets. They can discover uncataloged space debris and near-Earth asteroids without prior information. The sky survey observation mode faces complex astronomical observation scenes, but due to the complex astronomical observation background, a set of data often faces multiple types of multiple space objects, and the signal-to-noise ratio of the target is relatively low. The extracted candidate target set still includes a certain number of stars, comets, abnormal noise and other interference signals, which increases the difficulty of target detection.

[0040] Image processing in the sky survey observation mode generally includes: image preprocessing, stellar background extraction and space target detection, and target track association. Among them, the image preprocessing and target extraction processes already have relatively mature algorithm flows, and stellar background extraction and space target detection can generally be implemented based on mature astronomical software. The track association process is to achieve: from the candidate data set containing multiple types and multiple targets and with strong noise obtained by target detection from the time series image, the time series data points belonging to the same target are associated to form arc segments. The arc segments generated by track association are the data basis for the subsequent processes of orbit determination, cataloging, and early warning of space targets, but there is currently no relevant mature technology to implement this process, especially the track association of multiple types and multiple targets (referred to as multi-targets) in complex astronomical observation scenarios in the sky survey mode, which is still a research hotspot and a technical difficulty that needs to be solved in the field of target detection.

[0041] Based on this, the present invention provides a track association method and device based on parameter space multidimensional feature clustering, which is used to solve the technical problems that the traditional Hough transform cannot use spatiotemporal multidimensional features in the track association application process, and the peak extraction is difficult due to the peak clustering phenomenon. The track association method based on parameter space multidimensional feature clustering includes:

[0042] Acquire simulated observation data of the motion trajectory of the space target object, and add Poisson distribution clutter to the simulated observation data to generate a data set, wherein the Poisson distribution clutter is used to simulate the interference clutter of the space target object in a real observation environment;

[0043] The data set is converted from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space, wherein the distribution characteristics represent the set of all possible motion trajectories of each space target object in the observation space;

[0044] Based on the distribution characteristics, the position information of each spatial target object in the parameter space is superimposed on its position information in the observation space to obtain a multi-dimensional feature data set;

[0045] According to the multi-dimensional feature data set, the coordinate parameters corresponding to the spatial target objects on the same motion trajectory are clustered using the pre-trained multi-dimensional spatiotemporal feature density clustering model;

[0046] The coordinate parameters are reversely converted into trajectory parameters of the space target object in the observation space, wherein the trajectory parameters represent the slope and intercept of the motion trajectory of the space target object;

[0047] Based on the trajectory parameters, the spatial target objects that meet the preset motion trajectory trend in the observation space are associated to obtain associated arc segments.

[0048] The track association method based on multidimensional feature clustering in parameter space provided by the present invention is different from the traditional parameter space conversion method (such as the classic Hough transform). Instead, it converts the peak extraction problem with low accuracy into a multidimensional feature clustering problem, and clusters the multidimensional features using a multidimensional spatiotemporal feature density clustering model (MDST-DBSCAN model), thereby solving the technical problem of difficulty in peak extraction caused by track clustering. To a certain extent, it has a certain degree of adaptability to the value of the discretization interval of the parameter space, and is more suitable for engineering applications.

[0049] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0050] Figure 1 The flowchart of the track association method based on parameter space multi-dimensional feature clustering according to an embodiment of the present invention is schematically shown.

[0051] like Figure 1 As shown, the track association method based on parameter space multi-dimensional feature clustering in this embodiment may include operations S1 to S5.

[0052] In operation S1, simulated observation data of the motion trajectory of the space target object is obtained, and Poisson distribution clutter is added to the simulated observation data to generate a data set, wherein the Poisson distribution clutter is used to simulate the interference clutter of the space target object in a real observation environment.

[0053] In operation S2, the data set is converted from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space, wherein the distribution characteristics represent a set of all possible motion trajectories of each space target object in the observation space.

[0054] In operation S3, based on the distribution characteristics, the position information of each spatial target object in the parameter space is superimposed on its position information in the observation space to obtain a multi-dimensional feature data set.

[0055] In operation S4, based on the multi-dimensional feature data set, the coordinate parameters corresponding to the spatial target objects located on the same motion trajectory are clustered using a pre-trained multi-dimensional spatiotemporal feature density clustering model.

[0056] In operation S5 , the coordinate parameters are reversely converted into trajectory parameters of the space target object in the observation space, wherein the trajectory parameters represent the slope and intercept of the motion trajectory of the space target object.

[0057] In operation S6, based on the trajectory parameters, the spatial target objects that meet the preset motion trajectory trend in the observation space are associated to obtain associated arc segments.

[0058] In this embodiment, operation S1 obtains simulated observation data of the motion trajectory of the space target object, and adds Poisson distribution clutter to the simulated observation data to generate a data set. For example, 4 approximate linear trajectories / arc segments can be simulated, a total of 136 real data points (i.e., space target objects) are simulated, and random noise satisfying the Poisson distribution λ=60 is added thereto to construct a data set.

[0059] In the track association method based on parameter space multidimensional feature clustering provided in the embodiment of the present invention, the technical details of space debris and near-Earth asteroid image processing and target extraction involved are common knowledge in the field and will not be repeated here.

[0060] Assuming that there are N space target objects in the observation space moving in a straight line at a uniform speed, and the motions of each space target object are independent of each other, a data set with significant motion trajectory trends is formed through observation by the photoelectric detection system and conventional processing.

[0061] Considering the noise, the motion equation of the space target object can be described as:

[0062]

[0063]

[0064] in, represents the state vector of the target, , Indicates the target location, , represents the target movement speed, represents the state transfer matrix, represents a Gaussian white noise sequence with a mean of 0 and a known variance, and T represents a transpose operation.

[0065] According to an embodiment of the present invention, operation S2 converts the data set from the observation space to the parameter space, and obtains the distribution characteristics of each space target object in the parameter space, which may include:

[0066] Using the linear slope-intercept transformation function, the observed coordinates of the space target object in the observation space are converted into parameter coordinates in the parameter space;

[0067] According to the parameter coordinates, a sine curve corresponding to each space target object in the parameter space is generated, wherein the sine curves corresponding to the space target objects located on the same motion trajectory have local intersections.

[0068] In this embodiment, the parameter space conversion function, that is, the transformation function, is an important medium for converting the motion trajectory detection problem of the target in the observation space into the local peak detection problem in the parameter space.

[0069] The selection of transformation functions is based on the trajectory characteristics of the moving target. Since the movement speed of near-Earth asteroids and high-orbit space debris is relatively low, their movement trajectories can be assumed to be approximate straight lines during the observation period. Therefore, the straight line slope-intercept transformation function can be used as the parameter space conversion function to convert the observation coordinates (α, δ) of the target to be selected into the parameter coordinates (ρ, θ) in the parameter space.

[0070] The slope-intercept transformation function of the line is:

[0071]

[0072] in,( , ) represents the observation coordinates, ( , ) represents the parameter coordinates, ∈(0,π).

[0073] Assume that there is a set of n observation points that are collinear (i.e., located on the same motion trajectory) in the observation space, and the coordinates are expressed as (α i , δ i ), then in the ideal state without considering noise and clutter, there will be a unique set of parameters (ρ0, θ0) in the parameter space, so that this set of coordinates satisfies the following conditions:

[0074]

[0075] Among them, i∈(1,n).

[0076] That is, a group of observation data points with a collinear relationship will form a group of sinusoidal curves with a common intersection in the parameter space after Hough transform, and the coordinates of the intersection determine the slope and intercept of the motion trajectory of the high-orbit, low-speed space target object in the observation space.

[0077] The voting space is a discretized space similar to the parameter space, which can be understood as a two-dimensional matrix storing the number of votes, and its size is consistent with the parameter space. The data set is transformed into the parameter space through the linear slope-intercept transformation function. Each observation point forms a sine curve in the parameter space. The observation points on the same linear trajectory are represented in the parameter space as local intersections of their sine curves. The original values ​​of the pixels in the voting space are all 0. The curve converted from the observation space to the parameter space will pass through a series of discrete intervals. Each time a discrete interval is received, it is regarded as a vote, and the corresponding pixel value count in the voting space is increased by 1.

[0078] Spatial transformation in complex multi-target scenarios will form multiple peaks in the voting space, and due to the discrete properties of the error and parameter space itself, local peak "clusters" rather than peak "points" often appear. Therefore, in this embodiment, multi-dimensional spatiotemporal feature density clustering is used to obtain cluster peaks.

[0079] According to an embodiment of the present invention, operation S3 superimposes the position information of each spatial target object in the parameter space onto its position information in the observation space based on the distribution characteristics to obtain a multidimensional feature data set. For example, the parameter coordinates of each spatial target object in the parameter space are superimposed onto its observation coordinates in the observation space.

[0080] The closer the observed data points are in the voting space, the more likely they are to belong to the same target collinear trajectory. Therefore, the rational use of the voting space position information helps to transform the linear trajectory detection problem of moving targets into the local peak detection problem. Each accumulation of discrete interval pixel values ​​in the voting space comes from a sine curve in the parameter space, that is, from an observed data point in the observation space. This means that the closer the voting space position is, the more likely they are to belong to the collinear trajectory in the observation space.

[0081] Therefore, in this embodiment, each accumulation of pixel values ​​in the voting space is regarded as an independent point, and then the position attribute (parameter coordinate) of this point in the voting space is superimposed on the position attribute (observation coordinate) in its corresponding observation space, which is equivalent to adding a feature dimension to the data of the original data set to obtain a multi-dimensional feature data set.

[0082] According to an embodiment of the present invention, before operation S4 clustering the coordinate parameters corresponding to the space target objects located on the same motion trajectory using a pre-trained multi-dimensional spatiotemporal feature density clustering model according to the multi-dimensional feature data set, the following steps are included:

[0083] Extract the time information, morphological information and brightness information of each spatial target object in the data set;

[0084] The time information, morphological information and brightness information are superimposed on a multidimensional feature data set to obtain a multidimensional feature array; wherein the feature dimensions of each spatial target object in the multidimensional feature array include: position information in the observation space, position information in the parameter space, time information, morphological information and brightness information.

[0085] Although the data points with closer voting space coordinates are more likely to belong to the same linear trajectory of the target, the trajectory distribution of multiple targets in complex astronomical scenes may be crossed or close to each other, so the simple voting space position information is not enough to achieve track association. For example, the traditional track association method generally only considers the location information of the data set, but ignores the multi-dimensional spatiotemporal characteristics of the data set.

[0086] Therefore, in this embodiment, it is assumed that multiple observations belonging to the same target are also similar in shape and brightness, and multi-dimensional feature information including target shape characteristics and brightness characteristics is introduced to participate in clustering, such as Figure 2 shown.

[0087] Figure 2 The schematic diagram shows a principle diagram of obtaining a multi-dimensional feature array by superimposing multi-dimensional feature information according to an embodiment of the present invention.

[0088] like Figure 2 As shown, the data set information obtained after target extraction and astronomical positioning includes multi-dimensional information such as position, time, target shape, brightness, etc., and also includes position information of parameter space superimposed by operation S3. In this embodiment, this information is comprehensively utilized, and the feature dimensions of each space target object in the multi-dimensional feature array finally obtained include: position information in the observation space, position information in the parameter space, time information, shape information and brightness information (magnitude).

[0089] Magnitude is a measure of the brightness of a celestial body, that is, a measure of the brightness of a star. The smaller the magnitude, the brighter the star, and the larger the magnitude, the dimmer the star.

[0090] The pixel value in the voting space represents the accumulation of peak values, and each pixel value time series data set is transformed by Hough and peak value accumulation, which can be expressed in the voting space as follows:

[0091] TSM={TSM1,TSM2,…,TSM n}

[0092] Where n represents the total number of observations in the data set, and any data point can be represented as TSM i .

[0093] Assume that the multiple targets in the observation space can be expressed as:

[0094] TSO={TSO1,TSO2,…,TSO m}

[0095] Where m represents the total number of targets / tracks in the dataset, and any target is represented by TSO j Considering multidimensional features, any target in the data set can be represented by a multidimensional feature array:

[0096]

[0097] Among them, position represents the position information, but it should be noted that the attribute of position is the position attribute of the voting space, not the position attribute of the observation space, time represents the time information, shape index represents the shape information, and MAG AUTO represents the brightness information. The multidimensional feature array contains k elements, representing the dimensions of the multidimensional features of the parameter space.

[0098] In this embodiment, multi-dimensional features such as position, time, brightness, major semi-axis, minor semi-axis and angle in morphological index can be further analyzed to screen and determine features and dimensions that are efficient for subsequent clustering algorithms.

[0099] The track association method based on parameter space multidimensional feature clustering provided by the embodiment of the present invention is different from the traditional track association method which only considers the location information of the data set but ignores the multidimensional spatiotemporal characteristics of the data set. In this method, multidimensional feature information including the time characteristics, morphological characteristics and brightness characteristics of the space target objects is introduced to participate in clustering, thereby solving the problem of local area observation attribution caused by the proximity and intersection of multiple target tracks, and making the track association result more accurate.

[0100] According to an embodiment of the present invention, operation S4 may cluster coordinate parameters corresponding to space target objects located on the same motion trajectory based on the multidimensional feature data set using a pre-trained multidimensional spatiotemporal feature density clustering model, for example, including:

[0101] According to the multidimensional feature array, the coordinate values ​​corresponding to the intersection points in the sine curve are extracted using a pre-trained multidimensional spatiotemporal feature density clustering model; wherein the coordinate values ​​corresponding to the intersection points represent the coordinate parameters corresponding to the spatial target objects located on the same motion trajectory.

[0102] In this embodiment, a multidimensional spatiotemporal feature density clustering model is selected to extract intersection points in a sine curve. The multidimensional spatiotemporal feature density clustering model is constructed based on the MDST-DBSCAN algorithm. The MDST-DBSCAN algorithm is an improvement based on the DBSCAN algorithm and is the core clustering algorithm of the method of this embodiment.

[0103] The MDST-DBSCAN algorithm needs to determine the following related elements before clustering:

[0104] (1) Determine the multidimensional feature array TSO j The features and dimensions k included, that is, which features are considered in the multidimensional feature clustering process to support the identification of data points, and then construct the multidimensional feature array TSO j .

[0105] (2) Determine the multidimensional feature weight w ι (ι∈[1,k]), must satisfy .

[0106] (3) Determine the neighborhood radius Eps and the minimum number of samples MinPts. Considering the complexity of multi-dimensional spatiotemporal characteristics, the neighborhood radius is determined by experimental methods. By referring to the life cycle threshold criterion in the field of signal detection, a Monte Carlo simulation experiment is designed to estimate and optimize the threshold, and the feasibility of a dynamic threshold screening method based on data is explored. In this embodiment, the parameters of the neighborhood radius Eps and the minimum number of samples MinPts can be set to, for example, Eps = 2 and MinPts = 15.

[0107] (4) Determine the similarity measurement algorithm. In this embodiment, the multi-dimensional Euclidean distance Distant (TSM a ,TSM b ) describes the similarity of the comprehensive multi-dimensional characteristics of data points.

[0108] Based on the above related elements, the multidimensional feature array is clustered, the number of clusters is adaptively obtained, and outliers are removed to obtain the clustering parameter value. Specifically:

[0109] The multidimensional Euclidean distance between each pair of spatial target objects is calculated, wherein the multidimensional Euclidean distance represents the similarity of the multidimensional features of each pair of spatial target objects.

[0110] Considering that the clustering algorithm is based on multi-dimensional spatiotemporal feature clustering, the traditional one-dimensional spatial distance similarity measurement method is no longer applicable. In this embodiment, a multi-dimensional Euclidean distance is used to comprehensively express the data point TSM a With TSM bIt is a comprehensive state of the similarity of k dimensional features, including time, position, brightness, major semi-axis, minor semi-axis and angle in the morphological index.

[0111] The multidimensional Euclidean distance is calculated as:

[0112]

[0113] in, Represents data points and The multidimensional Euclidean distance between The distance between two data points ( ) dimension, represents the time constraint function, Represents the multidimensional feature weight.

[0114] Define when TSM a With TSM b It comes from the time when the image is taken at the same time, f T (time a ,time b )=∞, otherwise f T (time a ,time b ) = 1. Through the time constraint function, the similarity between data points from images taken at the same time can be defined as positive infinity. In other words, only data points from images taken at different times can be identified as the same cluster.

[0115] Then, according to the multidimensional Euclidean distance, the multidimensional feature array is preliminarily divided to obtain multiple potential subclusters.

[0116] Within each potential subcluster, the local density of each spatial target object is recalculated based on the neighborhood radius and the minimum number of samples.

[0117] Then, the space target objects with a density higher than the threshold are selected as core points, and the space target objects connected by density are expanded to form clusters, wherein the space target objects in the same cluster are located on the same motion trajectory.

[0118] In this embodiment, one cluster corresponds to one intersection point. Finally, the coordinate value corresponding to the intersection point is obtained according to the cluster, that is, the coordinate parameter corresponding to the space target object located on the same motion trajectory is obtained. The specific clustering result is as follows: Figure 3 shown.

[0119] Figure 3 The diagram schematically shows a result of clustering using a multi-dimensional spatiotemporal feature density clustering model according to an embodiment of the present invention.

[0120] like Figure 3 As shown, in this embodiment, the multidimensional spatiotemporal feature density clustering model finally outputs 4 clusters, that is, 4 intersections are extracted, and the data points in the clusters are regarded as a group of collinear observations of the same target. Subsequently, the effect of track association is quantitatively evaluated using the silhouette coefficient and the Davidson-Botting index, and the multidimensional spatiotemporal feature density clustering model is adjusted according to the evaluation results to achieve the optimal effect, and finally the coordinate parameters (ρ, θ) of the parameter space are output.

[0121] According to an embodiment of the present invention, operation S5 of reversely converting the coordinate parameters into the trajectory parameters of the space target object in the observation space may include, for example:

[0122] According to the straight line slope-intercept transformation function, the coordinate values ​​corresponding to the intersection points are solved and processed to obtain the trajectory parameters of the space target object in the observation space.

[0123] In this embodiment, the linear parameters of the observation space can be calculated according to the linear slope-intercept transformation function: , solve to get the coordinate value of the intersection point in parameter space The slope-intercept of the corresponding observation space line .

[0124] Specifically, in this embodiment, four cluster peaks (intersection points) are extracted to obtain four corresponding sets of coordinate values: , after solving, 4 sets of slope-intercept straight line parameters are obtained.

[0125] According to an embodiment of the present invention, operation S6 associates the space target objects that meet the preset motion trajectory trend in the observation space based on the trajectory parameters, and the associated arc segments may include:

[0126] Match all spatial target objects in the observation space with the candidate motion trajectories to obtain a data set of associated points corresponding to each candidate motion trajectory;

[0127] In response to the number of space target objects in the associated point data set corresponding to each candidate motion trajectory exceeding a threshold, the corresponding candidate motion trajectory is successfully associated to obtain an associated arc segment.

[0128] In this embodiment, after the slope-intercept line parameters are obtained by inverse solution, the observation space data set is finally Perform data track association, specifically:

[0129] Match all data points in the observation space with the candidate straight line trajectory to obtain the associated point data set corresponding to each candidate straight line trajectory, where a data point Belong to a straight line trajectory The conditions are:

[0130]

[0131] Among them, A represents the lateral coefficient of the straight line trajectory, B represents the longitudinal coefficient of the straight line trajectory, and C represents the constant term. Represents the association tolerance.

[0132] After associating all data points in the observation space, each linear trajectory obtains a corresponding data set of associated points. When the amount of associated point data of a trajectory exceeds the threshold P, the trajectory / arc segment is considered to be successfully associated, and finally the associated arc segment is obtained, such as Figure 4 shown.

[0133] Figure 4 The arc segment graph obtained by successfully associating the track association method based on multi-dimensional feature clustering in parameter space according to an embodiment of the present invention is schematically shown.

[0134] like Figure 4 As shown in the figure, a backtracking operation is performed in the observation space. For each set of candidate straight line parameters obtained, the distance between all data points in the data set ∑(α, δ) and the straight line is calculated. After the association tolerance parameter ℶ=0.01 is constrained, the four arc segments are successfully associated.

[0135] In order to verify the feasibility and advancement of the track association method based on parameter space multidimensional feature clustering provided by the embodiment of the present invention, in this embodiment, the association results are quantitatively evaluated using two common parameters, TPR (True Positive Rate) and FPR (False Positive Rate), and the program running time is statistically analyzed. The results are shown in Table 1:

[0136] Table 1 Track association results

[0137]

[0138] As can be seen from Table 1, the TPR accuracy of the track association method based on parameter space multidimensional feature clustering provided by the embodiment of the present invention is higher than 0.9 and the FPR accuracy is about 0.1, indicating that the track association accuracy is high and the false alarm rate is at a low level. It can be seen that the method is feasible and advanced.

[0139] The track association method based on multi-dimensional feature clustering in parameter space provided by the present invention is different from the traditional parameter space conversion method (such as the classic Hough transform). It converts the peak extraction problem with low accuracy into a multi-dimensional feature clustering problem, and clusters the multi-dimensional features using a multi-dimensional spatio-temporal feature density clustering model, thereby solving the technical problem of difficulty in peak extraction caused by track clustering. To a certain extent, it has a certain degree of adaptability to the value of the discretization interval of the parameter space, and is more suitable for engineering applications.

[0140] Figure 5 The structure block diagram of the track association device based on parameter space multi-dimensional feature clustering according to an embodiment of the present invention is schematically shown.

[0141] like Figure 5 As shown, the track association device 500 based on parameter space multi-dimensional feature clustering according to the embodiment of the present invention includes: an acquisition module 510 , a conversion module 520 , a superposition module 530 , a clustering module 540 , a reverse conversion module 550 and an association module 560 .

[0142] The acquisition module 510 is used to acquire simulated observation data of the motion trajectory of the space target object, and add Poisson distribution clutter to the simulated observation data to generate a data set, wherein the Poisson distribution clutter is used to simulate the interference clutter of the space target object in a real observation environment.

[0143] The conversion module 520 is used to convert the data set from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space, wherein the distribution characteristics represent the set of all possible motion trajectories of each space target object in the observation space.

[0144] The superposition module 530 is used to superimpose the position information of each space target object in the parameter space onto its position information in the observation space based on the distribution characteristics, so as to obtain a multi-dimensional feature data set.

[0145] The clustering module 540 is used to cluster the coordinate parameters corresponding to the spatial target objects located on the same motion trajectory according to the multi-dimensional feature data set and using a pre-trained multi-dimensional spatio-temporal feature density clustering model.

[0146] The reverse conversion module 550 is used to reversely convert the coordinate parameters into the trajectory parameters of the space target object in the observation space, wherein the trajectory parameters represent the slope and intercept of the motion trajectory of the space target object.

[0147] The association module 560 is used to associate the space target objects that meet the preset motion trajectory trend in the observation space based on the trajectory parameters to obtain associated arc segments.

[0148] According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (ASICs), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging the circuit, or by any one of the three implementation methods of software, hardware, and firmware, or by a proper combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be executed.

[0149] For example, any multiple of the acquisition module 510, the conversion module 520, the superposition module 530, the clustering module 540, the reverse conversion module 550, and the association module 560 can be combined in one module / unit / sub-unit for implementation, or any one of the modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of these modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units and implemented in one module / unit / sub-unit. According to an embodiment of the present invention, at least one of the acquisition module 510, the conversion module 520, the superposition module 530, the clustering module 540, the reverse conversion module 550 and the association module 560 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in an appropriate combination of any of them. Alternatively, at least one of the acquisition module 510, the conversion module 520, the superposition module 530, the clustering module 540, the reverse conversion module 550 and the association module 560 can be at least partially implemented as a computer program module, and when the computer program module is run, the corresponding function can be performed.

[0150] It should be noted that the part of the track association device based on multi-dimensional feature clustering in parameter space in the embodiment of the present invention corresponds to the part of the track association method based on multi-dimensional feature clustering in parameter space in the embodiment of the present invention. The description of the part of the track association device based on multi-dimensional feature clustering in parameter space specifically refers to the part of the track association method based on multi-dimensional feature clustering in parameter space, which will not be repeated here.

[0151] Figure 6 The structural block diagram of an electronic device suitable for implementing a track association method based on multi-dimensional feature clustering in parameter space according to an embodiment of the present invention is schematically shown. Figure 6 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0152] like Figure 6 As shown, the electronic device 600 according to an embodiment of the present invention includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 to a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include an onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0153] In the storage part 608, various programs and data required for the operation of the electronic device 600 are stored. The processor 601, the ROM 602 and the storage part 608 are connected to each other through the bus 604. The processor 601 performs various operations of the method flow according to the embodiment of the present invention by executing the program in the ROM 602 and / or the storage part 608. It should be noted that the program can also be stored in one or more memories other than the ROM 602 and the storage part 608. The processor 601 can also perform various operations of the method flow according to the embodiment of the present invention by executing the program stored in the one or more memories.

[0154] According to an embodiment of the present invention, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to the bus 604. The electronic device 600 may further include one or more of the following components connected to the input / output (I / O) interface 605: an input portion 606 including a keyboard, a mouse, etc.; an output portion 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 608 including a hard disk, etc.; and a communication portion 609 including a network interface card such as a LAN card, a modem, etc. The communication portion 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage portion 608 as needed.

[0155] According to an embodiment of the present invention, the method flow according to an embodiment of the present invention can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, the above-mentioned functions defined in the system of the embodiment of the present invention are executed. According to an embodiment of the present invention, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0156] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0157] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device.

[0158] For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 602 and / or the storage portion 608 described above and / or one or more memories other than the ROM 602 and the storage portion 608 .

[0159] An embodiment of the present invention also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method provided by the embodiment of the present invention. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the method provided by the embodiment of the present invention.

[0160] When the computer program is executed by the processor 601, the above functions defined in the system / device of the embodiment of the present invention are executed. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0161] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 609, and / or installed from a removable medium 611. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0162] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0163] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments of the present invention can be combined and / or combined in various ways, even if such a combination or combination is not explicitly recorded in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features recorded in the various embodiments of the present invention can be combined and / or combined in various ways. All these combinations and / or combinations fall within the scope of the present invention.

[0164] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although each embodiment is described above, it does not mean that the measures in each embodiment cannot be used in combination advantageously. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A track association method based on multidimensional feature clustering in parameter space, characterized in that: The method comprises: Acquire simulated observation data of the motion trajectory of the space target object, and add Poisson distribution clutter to the simulated observation data to generate a data set, wherein the Poisson distribution clutter is used to simulate the interference clutter of the space target object in a real observation environment; The data set is converted from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space, wherein the distribution characteristics represent a set of all possible motion trajectories of each space target object in the observation space, and the data set is converted from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space, including: Using a linear slope-intercept transformation function, the observed coordinates of the space target object in the observation space are converted into parameter coordinates in the parameter space; Generate a sine curve corresponding to each space target object in the parameter space according to the parameter coordinates, wherein the sine curves corresponding to the space target objects located on the same motion trajectory have local intersection points; Based on the distribution characteristics, the position information of each space target object in the parameter space is superimposed on its position information in the observation space to obtain a multi-dimensional feature data set; Extracting time information, morphological information and brightness information of each spatial target object in the data set; The time information, the morphology information and the brightness information are superimposed on the multidimensional feature data set to obtain a multidimensional feature array, wherein the feature dimensions of each space target object in the multidimensional feature array include: position information in the observation space, position information in the parameter space, the time information, the morphology information and the brightness information; According to the multidimensional feature data set, using a pre-trained multidimensional spatiotemporal feature density clustering model, clustering to obtain coordinate parameters corresponding to spatial target objects located on the same motion trajectory; Reversely converting the coordinate parameters into trajectory parameters of the space target object in the observation space, wherein the trajectory parameters represent the slope and intercept of the motion trajectory of the space target object; Based on the trajectory parameters, the space target objects in the observation space that meet the preset motion trajectory trend are associated to obtain associated arc segments, including: Matching all spatial target objects in the observation space with candidate motion trajectories to obtain a correlation point data set corresponding to each candidate motion trajectory; In response to the number of space target objects in the associated point data set corresponding to each candidate motion trajectory exceeding a threshold, the corresponding candidate motion trajectory is successfully associated to obtain an associated arc segment.

2. The method according to claim 1, characterized in that: The straight line slope-intercept transformation function is: in,( , ) represents the observation coordinates, ( , ) represents the parameter coordinates, ∈(0,π).

3. The method according to claim 1, characterized in that The method of clustering the coordinate parameters corresponding to the space target objects on the same motion trajectory by using the pre-trained multi-dimensional spatiotemporal feature density clustering model according to the multi-dimensional feature data set includes: According to the multidimensional feature array, using a pre-trained multidimensional spatiotemporal feature density clustering model, extracting the coordinate values ​​corresponding to the intersection points in the sine curve; The coordinate values ​​corresponding to the intersection points represent the coordinate parameters corresponding to the space target objects located on the same motion trajectory.

4. The method according to claim 3, characterized in that: The method of extracting the coordinate values ​​corresponding to the intersection points in the sine curve according to the multidimensional feature array and using a pre-trained multidimensional spatiotemporal feature density clustering model includes: Calculating the multi-dimensional Euclidean distance between each pair of space target objects, wherein the multi-dimensional Euclidean distance represents the similarity of the multi-dimensional features of each pair of space target objects; Preliminarily dividing the multidimensional feature array according to the multidimensional Euclidean distance to obtain a plurality of potential subclusters; In each potential subcluster, the local density of each spatial target object is recalculated based on the neighborhood radius and the minimum number of samples; Selecting space target objects with a density higher than a threshold as core points, expanding density-connected space target objects to form clusters, where the space target objects in the same cluster are located on the same motion trajectory; According to the clusters, coordinate parameters corresponding to the space target objects located on the same motion trajectory are obtained.

5. The method according to claim 4, characterized in that The multi-dimensional Euclidean distance is calculated as follows: in, Represents data points and The multidimensional Euclidean distance between The distance between two data points The Euclidean distance in dimension, where , k represents the dimension of multidimensional features, represents the time constraint function, Represents the multidimensional feature weight.

6. The method according to claim 3, characterized in that: The reverse conversion of the coordinate parameters into the trajectory parameters of the space target object in the observation space comprises: According to the straight line slope-intercept transformation function, the coordinate values ​​corresponding to the intersection points are solved and processed to obtain the trajectory parameters of the space target object in the observation space.

7. A track association device based on multi-dimensional feature clustering in parameter space, characterized in that: The device comprises: An acquisition module, used to acquire simulated observation data of the motion trajectory of the space target object, and add Poisson distribution clutter to the simulated observation data to generate a data set, wherein the Poisson distribution clutter is used to simulate the interference clutter of the space target object in a real observation environment; A conversion module is used to convert the data set from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space, wherein the distribution characteristics represent all possible motion trajectory sets of each space target object in the observation space, and the conversion of the data set from the observation space to the parameter space to obtain the distribution characteristics of each space target object in the parameter space includes: Using a linear slope-intercept transformation function, the observed coordinates of the space target object in the observation space are converted into parameter coordinates in the parameter space; Generate a sine curve corresponding to each space target object in the parameter space according to the parameter coordinates, wherein the sine curves corresponding to the space target objects located on the same motion trajectory have local intersection points; A superposition module is used to superimpose the position information of each space target object in the parameter space onto its position information in the observation space based on the distribution characteristics to obtain a multidimensional feature data set; extract the time information, shape information and brightness information of each space target object in the data set; superimpose the time information, the shape information and the brightness information onto the multidimensional feature data set to obtain a multidimensional feature array, wherein the feature dimensions of each space target object in the multidimensional feature array include: the position information in the observation space, the position information in the parameter space, the time information, the shape information and the brightness information; A clustering module, for clustering coordinate parameters corresponding to space target objects located on the same motion trajectory according to the multidimensional feature data set using a pre-trained multidimensional spatiotemporal feature density clustering model; A reverse conversion module, used for reversely converting the coordinate parameters into trajectory parameters of the space target object in the observation space, wherein the trajectory parameters represent the slope and intercept of the motion trajectory of the space target object; An association module is used to associate the space target objects that meet the preset motion trajectory trend in the observation space based on the trajectory parameters to obtain associated arc segments, including: Matching all spatial target objects in the observation space with candidate motion trajectories to obtain a correlation point data set corresponding to each candidate motion trajectory; In response to the number of space target objects in the associated point data set corresponding to each candidate motion trajectory exceeding a threshold, the corresponding candidate motion trajectory is successfully associated to obtain an associated arc segment.

Images