Parameter adaptive space multi-target track association method, device, equipment and medium

Abstract

The invention provides a parameter adaptive space multi-target track association method, device and equipment and a medium, relates to the technical field of image processing, and aims to solve the problem that the prior art is difficult to adapt to non-linear track association in multi-target track association application. And a technical problem that a traditional density clustering algorithm needs to manually intervene in determination of clustering parameters and cannot be suitable for engineering application with variable density is solved. The method comprises the following steps: acquiring space-time position observation data of a space multi-target object, and preprocessing the space-time position observation data to generate a data set; according to the data set, extracting features capable of representing the distance relation between each object and the object close to the k object; determining the neighborhood radius of the target according to the distance relation characteristics; according to the neighborhood radius, identifying objects in the same cluster in the data set by using a pre-trained density clustering model; and associating the objects in each cluster to obtain a track association segmental arc of each target.

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 parameter adaptive spatial multi-target track association method, device, equipment and medium. Background Art

[0002] Since the first launch of artificial satellites, space objects have spread all over the Earth's orbit along with frequent space activities. Effective monitoring of space objects has important safety significance. 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. During sky survey observations, there are usually multiple types of meaningful targets in the image. How to timely discover meaningful targets in the sky area in image processing methods is still in the stage of continuous exploration and improvement.

[0003] Image processing in the sky survey observation mode generally includes: image preprocessing, star 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 star 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 for the complex astronomical observation scenarios of the sky survey mode. The track association of multiple types and multiple targets (referred to as multi-targets) with nonlinear trajectory morphology is still a research hotspot and a technical difficulty that needs to be solved in the field of target detection. Summary of the invention

[0004] In view of this, the present invention provides a parameter-adaptive spatial multi-target track association method, device, equipment and medium to solve the technical problems that the prior art is difficult to adapt to nonlinear trajectory association in multi-target track association applications, and the traditional density clustering algorithm requires human intervention in the determination of clustering parameters and cannot be applied to engineering applications with variable density.

[0005] One aspect of the present invention provides a parameter-adaptive spatial multi-target track association method, comprising: obtaining spatiotemporal position observation data of spatial multi-target objects, and preprocessing the spatiotemporal position observation data to generate a data set; based on the data set, extracting a distance relationship feature that can characterize the distance between each object and the object closest to it, wherein k is a positive integer greater than zero; determining the neighborhood radius of the target based on the distance relationship feature; based on the neighborhood radius, using a pre-trained density clustering model, identifying objects in the same cluster in the data set, wherein the objects in the same cluster represent trajectory points belonging to the same target in different time series; and respectively associating the objects in each cluster to obtain a track association arc segment for each target.

[0006] According to an embodiment of the present invention, the spatiotemporal position observation data are preprocessed to generate a data set, including: processing missing values ​​of the spatiotemporal position observation data; processing data outliers of the spatiotemporal position observation data after the missing values ​​processing; and standardizing the spatiotemporal position observation data after the data outliers processing.

[0007] According to an embodiment of the present invention, based on a data set, extracting features that can characterize the distance relationship between each object and the object that is k-th closest to it includes: extracting the Euclidean distance between each object and other objects, wherein the Euclidean distance characterizes the true distance between two objects; generating a distance distribution matrix based on the Euclidean distance; and generating a distance relationship feature between each object and the object that is k-th closest to it based on the distance distribution matrix.

[0008] According to an embodiment of the present invention, generating the distance relationship characteristics between each object and the object closest to it by the kth distance based on the distance distribution matrix includes: sorting the distance elements in the distance distribution matrix; using the kth column elements after the sorting as the vertical coordinate and the number of objects as the horizontal coordinate to generate a k-distance curve, wherein any point on the k-distance curve represents the distance between the point and the object closest to it by the kth distance.

[0009] According to an embodiment of the present invention, the distance distribution matrix is:

[0010]

[0011] in, represents the distance distribution matrix, Represents any object in the dataset With other objects The Euclidean distance between Indicates the number of objects in the dataset.

[0012] According to an embodiment of the present invention, determining the neighborhood radius of a target based on distance relationship characteristics includes: extracting the distance corresponding to an inflection point in a k-distance curve, wherein the inflection point represents a sudden change in density between objects; and determining the neighborhood radius of the target based on the distance corresponding to the inflection point.

[0013] According to an embodiment of the present invention, based on the neighborhood radius, using a pre-trained density clustering model, identifying objects in the same cluster in a data set includes: starting from any unvisited object in the data set, judging whether the number of all objects within the neighborhood radius is greater than or equal to a predetermined minimum number of samples; in response to the number of all objects within the neighborhood radius being greater than or equal to the predetermined minimum number of samples, creating a cluster with it as the core point, and adding all objects within the neighborhood radius to the current cluster; continuing to search for new unvisited objects in the current cluster, and judging whether the number of all objects within the neighborhood radius is greater than or equal to the predetermined minimum number of samples; in response to the number of all objects within the neighborhood radius being greater than or equal to the predetermined minimum number of samples, adding all objects within the neighborhood radius to the current cluster until no new objects are added to the current cluster; repeating the above operations until all objects in the data set are classified into corresponding clusters.

[0014] Another aspect of the present invention provides a parameter-adaptive space multi-target track association device, including: an acquisition module, used to acquire the spatiotemporal position observation data of space multi-target objects, and pre-process the spatiotemporal position observation data to generate a data set; an extraction module, used to extract, based on the data set, a distance relationship feature that can characterize the distance between each object and the object closest to it at kth distance, wherein k is a positive integer greater than zero; a determination module, used to determine the neighborhood radius of the target based on the distance relationship feature; an identification module, used to identify objects in the same cluster in the data set based on the neighborhood radius using a pre-trained density clustering model, wherein objects in the same cluster characterize their trajectory points belonging to the same target in different time series; an association module, used to associate the objects in each cluster respectively to obtain the track association arc segment of each target.

[0015] 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.

[0016] 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.

[0017] 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.

[0018] Compared with the prior art, the parameter adaptive spatial multi-target track association method, device, equipment and medium provided by the present invention have at least the following beneficial effects:

[0019] (1) The parameter-adaptive spatial multi-target track association method, device, equipment and medium provided by the present invention use the density clustering model DBSCAN (density based spatial clustering of applications with noise) as the basic theory for track association. Unlike the traditional k-means clustering algorithm, the density clustering model does not need to predict the number of target tracks contained in the time series observation data in advance during the clustering process. Instead, it identifies the number of adaptive clusters through a clustering process under radius and density constraints. Therefore, clustering can be achieved in the absence of prior knowledge of the number of targets / clustering clusters. The algorithm has strong anti-interference ability and can eliminate false target outliers. At the same time, it has strong tolerance for the shape of the clusters and can adapt to nonlinear trajectory application scenarios.

[0020] (2) The parameter-adaptive spatial multi-target track association method, device, equipment and medium provided by the present invention can simultaneously perform global processing on a set of data sets obtained in the sky survey mode, avoiding the traversal operations in the sequential algorithm and the Hough transform algorithm and its improved algorithm, reducing the computing cost, running fast and in parallel, and being suitable for processing large-scale data sets. Compared with traditional methods, it is more suitable for daily engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] 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:

[0022] Figure 1 A flow chart of a parameter adaptive spatial multi-target track association method according to an embodiment of the present invention is schematically shown;

[0023] Figure 2 A k-dist curve diagram in a parameter adaptive spatial multi-target track association method according to an embodiment of the present invention is schematically shown;

[0024] Figure 3 A diagram schematically shows an association effect diagram of a parameter adaptive spatial multi-target track association method according to an embodiment of the present invention;

[0025] Figure 4 A structural block diagram of a parameter adaptive spatial multi-target track association device according to an embodiment of the present invention is schematically shown;

[0026] Figure 5 The structure block diagram of an electronic device suitable for implementing a parameter adaptive spatial multi-target track association method according to an embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0027] 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.

[0028] 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.

[0029] 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.

[0030] 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.).

[0031] 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.

[0032] Since the first launch of artificial satellites, space objects have spread all over the Earth's orbit along with frequent space activities. Effective monitoring of space objects has important safety significance. 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. During sky survey observations, there are usually multiple types of meaningful targets in the image. How to timely discover meaningful targets in the sky area in image processing methods is still in the stage of continuous exploration and improvement.

[0033] Image processing in the sky survey observation mode generally includes: image preprocessing, star background extraction and space target detection, and target track association. Among them, the image preprocessing and target extraction process already have relatively mature algorithm flows. Stars appear in point-like forms on the image, and when space targets pass through the observation area, they present different morphological features from stars in the image because of their relative speed. Stars and targets can be efficiently distinguished by mathematical morphology. Generally, star background extraction and space target detection operations can be performed based on mature astronomical software. The track association process is to achieve: from the candidate data set containing multiple types of multiple targets and 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 realize this process, especially for the complex astronomical observation scenarios of the sky survey mode. The track association of multiple types of multiple targets (referred to as multi-targets) with nonlinear trajectory morphology is still a research hotspot and a technical difficulty that needs to be solved in the field of target detection.

[0034] Based on this, the present invention provides a parameter-adaptive spatial multi-target track association method, device, equipment and medium to solve the technical problems that the prior art is difficult to adapt to nonlinear track association in multi-target track association applications, and the traditional density clustering algorithm requires human intervention in the determination of clustering parameters and cannot be applied to engineering applications with variable density. The parameter-adaptive spatial multi-target track association method includes:

[0035] Obtaining the spatiotemporal position observation data of multiple target objects in space, and preprocessing the spatiotemporal position observation data to generate a data set;

[0036] According to the data set, extract the distance relationship feature that can characterize each object and its k-th closest object, where k is a positive integer greater than zero; determine the neighborhood radius of the target according to the distance relationship feature;

[0037] According to the neighborhood radius, the pre-trained density clustering model is used to identify the objects in the same cluster in the data set, where the objects in the same cluster represent the trajectory points of the same target in different time series;

[0038] The objects in each cluster are associated respectively to obtain the track association arc of each target.

[0039] The parameter adaptive spatial multi-target track association method provided in an embodiment of the present invention selects the density based spatial clustering model DBSCAN (density based spatial clustering of applications with noise) as the basic theory for track association. Unlike the traditional k-means clustering algorithm, the density clustering model does not need to predict the number of target tracks contained in the time series observation data in advance during the clustering process. Instead, it identifies the number of adaptive clusters through a clustering process under radius and density constraints. Therefore, clustering can be achieved in the absence of prior knowledge of the number of targets / clustering clusters. The algorithm has strong anti-interference ability and can eliminate false target outliers. At the same time, it has strong tolerance for the shape of the clusters and can adapt to nonlinear trajectory application scenarios.

[0040] 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.

[0041] Figure 1 The flowchart of the parameter adaptive spatial multi-target track association method according to an embodiment of the present invention is schematically shown.

[0042] like Figure 1 As shown, the parameter adaptive spatial multi-target track association method of this embodiment may include operations S1 to S5, for example.

[0043] In operation S1, the spatiotemporal position observation data of the spatial multi-target objects are acquired, and the spatiotemporal position observation data are preprocessed to generate a data set.

[0044] In operation S2, a feature capable of characterizing the distance relationship between each object and its k-th closest object is extracted according to the data set, where k is a positive integer greater than zero.

[0045] In operation S3, the neighborhood radius of the target is determined according to the distance relationship feature.

[0046] In operation S4, objects in the same cluster in the data set are identified according to the neighborhood radius using a pre-trained density clustering model, wherein the objects in the same cluster represent trajectory points belonging to the same target in different time series.

[0047] In operation S5, objects in each cluster are associated respectively to obtain track associated arcs of each target.

[0048] In this embodiment, in operation S1, the spatiotemporal position observation data of multiple target objects in space are obtained. For example, the spatiotemporal position observation data of multiple target objects obtained in a complex astronomical observation scene in a sky survey mode can be selected. In addition, it can also be simulation data.

[0049] Data preprocessing plays an important fundamental role in the parameter adaptive spatial multi-target track association method of this embodiment, and is a prerequisite for ensuring the implementation of the subsequent clustering parameter adaptive algorithm.

[0050] According to an embodiment of the present invention, preprocessing the spatiotemporal position observation data to generate a data set may include, for example:

[0051] Process missing values ​​for spatiotemporal location observation data;

[0052] The spatiotemporal location observation data after the data missing value processing is subjected to data outlier processing; the spatiotemporal location observation data after the data outlier processing is subjected to standardization processing.

[0053] Since a large amount of missing data will interfere with the operation of the model, this embodiment adopts the method of replacing missing values ​​with the mean value or deleting them for processing. If the number of missing values ​​reaches more than 5%, they are replaced by the mean, and if it is less than 5%, they can be directly deleted.

[0054] Each attribute feature has its own reasonable value range based on its specific physical meaning, and data beyond the value range can be considered as invalid values ​​and need to be deleted.

[0055] Standardization of data can make multiple parameters involved at the same level. In this embodiment, the standardization method is as follows:

[0056]

[0057] in, represents the sample value of the attribute feature, and Represents the maximum value of the attribute feature in the sample population.

[0058] According to an embodiment of the present invention, operation S2 extracts, based on the data set, features that can characterize the distance relationship between each object and the object closest to it by the kth distance, for example, including:

[0059] Extract the Euclidean distance between each object and other objects, where the Euclidean distance represents the true distance between two objects;

[0060] Generate a distance distribution matrix based on the Euclidean distance;

[0061] According to the distance distribution matrix, the distance relationship feature between each object and the object closest to it is generated.

[0062] In this embodiment, based on the k-distance algorithm, the Euclidean distance between each object in the data set and other objects is first calculated.

[0063] Euclidean distance, also known as Euclidean metric, is a distance definition that refers to the real distance between two points in m-dimensional space, or the natural length of a vector (that is, the distance from the point to the origin). The Euclidean distance in two-dimensional and three-dimensional space is the actual distance between two points.

[0064] Then, based on the Euclidean distance, the distance distribution matrix Dist is generated. n*n :

[0065]

[0066] Among them, Dist n*n represents the distance distribution matrix, Represents any object in the dataset With other objects The Euclidean distance between Indicates the number of objects in the dataset.

[0067] Finally, based on the distance distribution matrix, the distance relationship feature between each object and its kth closest object can be generated.

[0068] According to an embodiment of the present invention, generating the distance relationship feature between each object and the object closest to it by the kth distance according to the distance distribution matrix may include, for example:

[0069] Sorting the distance elements in the distance distribution matrix;

[0070] The elements of the kth column after sorting are used as the ordinate and the number of objects is used as the abscissa to generate a k-distance curve, wherein any point on the k-distance curve represents the distance between the point and the object closest to it.

[0071] In this embodiment, the distance relationship feature can be represented by drawing a k-dist curve (k-distance curve), specifically:

[0072] Arrange the distance elements in the distance distribution matrix in ascending order, and then use the kth column element in the sorted matrix as the ordinate of the k-dist curve, and the number of objects X (X∈[1,n]) as the abscissa of the k-dist curve. The physical meaning of any point on the k-dist curve is: the distance between the Xth object in the data set H and the object closest to it. Different values ​​of parameter k will generate different k-dist curves. Generally, the k-dist curves of the same set of data have the same trend.

[0073] According to an embodiment of the present invention, operation S3 determines the neighborhood radius of the target according to the distance relationship feature, including:

[0074] Extract the distance corresponding to the inflection point in the k-distance curve, where the inflection point represents a sudden change in the density between objects;

[0075] According to the distance corresponding to the inflection point, the neighborhood radius of the target is determined.

[0076] In this embodiment, the k-dist curve represents the distance ranking between the Xth object in the data set H and other objects in the data set. The smooth part of the curve indicates that the data density is high and the distance between objects is close, while the inflection point reflects the sudden change of the relative distance.

[0077] According to the principle of clustering: distance is an important indicator to measure whether two objects belong to the same class of targets, so it can be considered that the distance corresponding to the inflection point is the best distance to distinguish whether an object and other objects belong to the same class of targets. The inflection point in this embodiment can be obtained by the elbow method, and the distance corresponding to the inflection point is the adaptive value of the neighborhood radius Eps of the density clustering model parameter.

[0078] After determining the value of the neighborhood radius Eps, the data set can be clustered according to the trained density clustering model.

[0079] Since the optimal value of the minimum number of samples MinPts has been determined during the training process, clustering can be performed directly based on the determined value of the neighborhood radius Eps.

[0080] The density clustering model defines the core object as the object that contains the minimum number of samples MinPts within a given neighborhood radius Eps. That is, the qualified object requires the number of objects within the neighborhood radius Eps to exceed a certain threshold level.

[0081] The minimum number of samples MinPts (Minimum Points) is a key parameter in the density clustering model. It defines the minimum number of objects that must be contained within the neighborhood radius Eps before an object is considered a core object.

[0082] In this embodiment, judging whether an object is a core object is to judge whether the number of all objects within the neighborhood radius Eps is greater than or equal to the minimum sample number MinPts. Specifically:

[0083] An object is a core object if it has at least MinPts objects (including itself) within its neighborhood radius Eps.

[0084] If an object itself is not a core object, but is within the neighborhood radius Eps of a core object, then the object is a boundary object.

[0085] If an object is neither a core object nor a boundary object, it will be marked as a noise object.

[0086] According to an embodiment of the present invention, based on the neighborhood radius, using a pre-trained density clustering model, identifying objects in the same cluster in a data set may specifically include:

[0087] Starting from any unvisited object in the data set, determine whether the number of all objects within its neighborhood radius is greater than or equal to the predetermined minimum number of samples;

[0088] In response to the number of all objects within the neighborhood radius being greater than or equal to a predetermined minimum number of samples, a cluster is created using the object as a core point, and all objects within the neighborhood radius are added to the current cluster;

[0089] Continue to search for new objects that have not been visited in the current cluster, and determine whether the number of all objects within the neighborhood radius is greater than or equal to the predetermined minimum number of samples;

[0090] In response to the number of all objects within the neighborhood radius being greater than or equal to a predetermined minimum number of samples, all objects within the neighborhood radius are added to the current cluster until no new objects are added to the current cluster;

[0091] Repeat the above steps until all objects in the data set are classified into corresponding clusters.

[0092] In this embodiment, the density clustering model identifies clusters based on the two parameters of neighborhood radius Eps and minimum number of samples MinPts, and finally classifies all objects into clusters, where objects in the same cluster represent the trajectory points of the same target in different time series, and finally associate the objects in each cluster to obtain the track association arc of each target. Objects that are free from each cluster are not included in the neighborhood radius of any core point and are regarded as abnormal noise objects.

[0093] The parameter adaptive spatial multi-target track association method provided by the present invention selects a density clustering model as the basic theory for track association. Unlike traditional k-means clustering algorithms, the density clustering model does not need to predict the number of target tracks contained in the time series observation data in advance during the clustering process. Instead, it identifies the number of adaptive clusters through a clustering process under radius and density constraints. Therefore, clustering can be achieved in the absence of prior knowledge of the number of targets / clustering clusters. The algorithm has strong anti-interference ability and can eliminate false target anomalies. At the same time, it has strong tolerance for the shape of clusters and can adapt to nonlinear trajectory application scenarios.

[0094] On the basis of the above embodiment, after the track association of each target is successful, the association effect can also be evaluated by screening the quantitative indicator Rand Index (RI).

[0095] For example, in the evaluation phase, the following data can be used to calculate the precision quantitative index, including:

[0096] (1) TP, which represents the number of data points whose true target points are correctly predicted as true.

[0097] (2) TN, represents the number of clutter points that are correctly predicted as false data points.

[0098] (3) FP, which represents the number of data points where clutter points are incorrectly predicted as true.

[0099] (4) FN, represents the number of data points where the true target point is incorrectly predicted as a false one.

[0100]

[0101] The RI indicator can measure the similarity between the clustering results and the labels / true categories. The value range is [0,1]. The larger the value, the more similar the clustering results are to the true categories, and the better the clustering effect.

[0102] The parameter-adaptive spatial multi-target track association method provided by the present invention can simultaneously perform global processing on a group of data sets obtained in the sky survey mode, avoiding the traversal operations in the calculation process of the sequential algorithm, the Hough transform algorithm and its improved algorithm, reducing the computing cost, running at a fast speed and parallelization, and is suitable for processing large-scale data sets. Compared with traditional methods, it is more suitable for daily engineering applications.

[0103] Figure 2 The k-dist curve diagram in the parameter adaptive spatial multi-target track association method according to an embodiment of the present invention is schematically shown.

[0104] like Figure 2As shown, the parameter adaptive space multi-target track association method of an embodiment of the present invention simulates 5 tracks / arcs in the sky survey observation mode, including 2 nonlinear tracks, and adds random clutter satisfying the Poisson distribution λ=60 to verify the track association method and analyze its effect.

[0105] After preprocessing, the dataset contains a total of 186 objects, of which 139 labels are true and the rest are random noise. After standardization, the horizontal and vertical coordinates of the objects are in the range of [0,1].

[0106] Considering that the objects associated in the same arc segment must come from different observation times, or extracted from different images, the time constraint f is added when calculating the distance matrix. T (time a , time b ), then the object TSM a With Object TSM b The distance between It can be expressed as:

[0107]

[0108] Among them, f T (time a , time b ) represents the time constraint function.

[0109] In this embodiment, the object TSM is defined as a With Object TSM b It comes from the time when the image is taken at the same time, f T (time a , time b )=9999, otherwise f T (time a , time b )=1.

[0110] 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. The distance distribution matrix of the experimental data set is calculated, and after eliminating invalid values, the k-dist curve is drawn based on the distance distribution matrix with k=4.

[0111] Based on the k-dist curve, the inflection point is extracted according to the elbow method, and the optimal neighborhood radius Eps parameter value is 0.1412.

[0112] Figure 3The figure schematically shows the association effect diagram of the parameter adaptive spatial multi-target track association method according to an embodiment of the present invention.

[0113] like Figure 3 As shown in the figure, the density clustering model clustered the data set under the parameter settings of neighborhood radius Eps=0.1412 and minimum number of samples MinPts=5.

[0114] In this embodiment, the density clustering model identified a total of 5 arcs. The unassociated object points in the figure are outliers identified by the density clustering model, which represent the clutter predicted by the algorithm; the object points associated with the arcs represent the true targets predicted by the algorithm, and different arcs represent different clusters. The normalized data set is reversed to obtain the observation space arc information.

[0115] In order to verify the accuracy of the density clustering model, the results in this embodiment use the quantitative indicator Rand Index (RI) to evaluate its correlation effect, and the results are shown in Table 1:

[0116] Table 1 Results of track association algorithm

[0117]

[0118] It can be seen from Table 1 that the RI accuracy of track association using the parameter adaptive spatial multi-target track association method of the embodiment of the present invention is higher than 0.95, indicating that the similarity between the clustering results and the labels / true categories is high, and the track association effect is good. It can be seen that this method is feasible and advanced.

[0119] Figure 4 The structure block diagram of the parameter adaptive spatial multi-target track association device according to an embodiment of the present invention is schematically shown.

[0120] like Figure 4 As shown, the parameter adaptive spatial multi-target track association device 400 of the embodiment of the present invention includes: an acquisition module 410, an extraction module 420, a determination module 430, an identification module 440 and an association module 450.

[0121] The acquisition module 410 is used to acquire the spatiotemporal position observation data of the spatial multi-target objects, and pre-process the spatiotemporal position observation data to generate a data set.

[0122] The extraction module 420 is used to extract, based on the data set, features that can characterize the distance relationship between each object and its k-th closest object, where k is a positive integer greater than zero.

[0123] The determination module 430 is used to determine the neighborhood radius of the target according to the distance relationship feature.

[0124] The identification module 440 is used to identify objects in the same cluster in the data set according to the neighborhood radius using a pre-trained density clustering model, wherein the objects in the same cluster represent trajectory points belonging to the same target in different time series.

[0125] The association module 450 is used to associate the objects in each cluster respectively to obtain the track association arc segment of each target.

[0126] 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.

[0127] For example, any multiple of the acquisition module 410, the extraction module 420, the determination module 430, the identification module 440, and the association module 450 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 410, the extraction module 420, the determination module 430, the identification module 440, and the association module 450 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 by any one of the three implementation methods of software, hardware, and firmware, or by a suitable combination of any of them. Alternatively, at least one of the acquisition module 410 , the extraction module 420 , the determination module 430 , the identification module 440 , and the association module 450 may be at least partially implemented as a computer program module, and when the computer program module is executed, a corresponding function may be performed.

[0128] It should be noted that the parameter adaptive space multi-target track association device part in the embodiment of the present invention corresponds to the parameter adaptive space multi-target track association method part in the embodiment of the present invention. The description of the parameter adaptive space multi-target track association device part specifically refers to the parameter adaptive space multi-target track association method part, which will not be repeated here.

[0129] Figure 5 The structure block diagram of an electronic device suitable for implementing a parameter adaptive spatial multi-target track association method according to an embodiment of the present invention is schematically shown. Figure 5 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.

[0130] like Figure 5As shown, the electronic device 500 according to an embodiment of the present invention includes a processor 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage part 508 to the random access memory (RAM) 503. The processor 501 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 501 may also include an onboard memory for caching purposes. The processor 501 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.

[0131] In the storage part 508, various programs and data required for the operation of the electronic device 500 are stored. The processor 501, the ROM 502 and the storage part 508 are connected to each other through the bus 504. The processor 501 performs various operations of the method flow according to the embodiment of the present invention by executing the program in the ROM 502 and / or the storage part 508. It should be noted that the program can also be stored in one or more memories other than the ROM 502 and the storage part 508. The processor 501 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.

[0132] According to an embodiment of the present invention, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to the bus 504. The electronic device 500 may further include one or more of the following components connected to the input / output (I / O) interface 505: an input portion 506 including a keyboard, a mouse, etc.; an output portion 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 508 including a hard disk, etc.; and a communication portion 509 including a network interface card such as a LAN card, a modem, etc. The communication portion 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that the computer program read therefrom is installed into the storage portion 508 as needed.

[0133] 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 509, and / or installed from the removable medium 511. When the computer program is executed by the processor 501, 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.

[0134] 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.

[0135] 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.

[0136] For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 502 and / or the storage portion 508 described above and / or one or more memories other than the ROM 502 and the storage portion 508 .

[0137] 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.

[0138] When the computer program is executed by the processor 501, 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.

[0139] 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 509, and / or installed from the removable medium 511. 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.

[0140] 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).

[0141] 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 such combinations and / or combinations fall within the scope of the present invention.

[0142] 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 parameter adaptive spatial multi-target track association method, characterized in that: The method comprises: Acquire spatiotemporal position observation data of multiple spatial target objects, and preprocess the spatiotemporal position observation data to generate a data set; Extracting, from the data set, features that can characterize the distance relationship between each object and its k-th closest object, where k is a positive integer greater than zero; Determining the neighborhood radius of the target according to the distance relationship characteristics; According to the neighborhood radius, using a pre-trained density clustering model, objects in the same cluster in the data set are identified, wherein the objects in the same cluster represent trajectory points belonging to the same target in different time series; The objects in each cluster are associated respectively to obtain the track association arc of each target.

2. The method according to claim 1, characterized in that: The preprocessing of the spatiotemporal position observation data to generate a data set includes: Processing missing values ​​of the spatiotemporal position observation data; Perform data outlier processing on the spatiotemporal location observation data after the data missing values ​​are processed; The spatiotemporal location observation data after data outliers are processed are standardized.

3. The method according to claim 1, characterized in that The extracting, based on the data set, features that can characterize the distance relationship between each object and the object closest to it by the kth distance includes: Extracting the Euclidean distance between each object and other objects, wherein the Euclidean distance represents the true distance between the two objects; Generate a distance distribution matrix according to the Euclidean distance; According to the distance distribution matrix, a distance relationship feature between each object and the object closest to it by the kth distance is generated.

4. The method according to claim 3, characterized in that Generating the distance relationship feature between each object and the object closest to it by the kth distance according to the distance distribution matrix includes: Sorting the distance elements in the distance distribution matrix; The elements of the kth column after sorting are used as the ordinate and the number of objects is used as the abscissa to generate a k-distance curve, wherein any point on the k-distance curve represents the distance between the point and the object closest to it.

5. The method according to claim 3, characterized in that: The distance distribution matrix is: in, represents the distance distribution matrix, Represents any object in the dataset With other objects The Euclidean distance between Indicates the number of objects in the dataset.

6. The method according to claim 4, characterized in that Determining the neighborhood radius of the target according to the distance relationship feature includes: Extracting the distance corresponding to the inflection point in the k-distance curve, wherein the inflection point represents a sudden change in density between objects; The neighborhood radius of the target is determined according to the distance corresponding to the inflection point.

7. The method according to claim 1, characterized in that The identifying objects in the same cluster in the data set using a pre-trained density clustering model according to the neighborhood radius includes: Starting from any unvisited object in the data set, determine whether the number of all objects within its neighborhood radius is greater than or equal to a predetermined minimum number of samples; In response to the number of all objects within the neighborhood radius being greater than or equal to a predetermined minimum number of samples, a cluster is created using the object as a core point, and all objects within the neighborhood radius are added to the current cluster; Continue to search for new objects that have not been visited in the current cluster, and determine whether the number of all objects within the neighborhood radius is greater than or equal to a predetermined minimum number of samples; In response to the number of all objects within the neighborhood radius being greater than or equal to a predetermined minimum number of samples, all objects within the neighborhood radius are added to the current cluster until no new objects are added to the current cluster; The above operation is repeated until all objects in the data set are classified into corresponding clusters.

8. A parameter adaptive spatial multi-target track association device, characterized in that: The device comprises: An acquisition module is used to acquire the spatiotemporal position observation data of multiple target objects in space, and preprocess the spatiotemporal position observation data to generate a data set; an extraction module, configured to extract, based on the data set, features that can characterize the distance relationship between each object and its k-th closest object, where k is a positive integer greater than zero; A determination module, used to determine the neighborhood radius of the target according to the distance relationship feature; An identification module, used to identify objects in the same cluster in the data set according to the neighborhood radius and using a pre-trained density clustering model, wherein the objects in the same cluster represent trajectory points belonging to the same target in different time series; The association module is used to associate the objects in each cluster respectively to obtain the track association arc segment of each target.

9. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having executable instructions stored thereon, characterized in that: When the instruction is executed by a processor, the processor executes the method according to any one of claims 1 to 7.

Images