A point cloud target grouping method and device based on optimal transmission flow consistency

Abstract

This invention provides a point cloud target grouping method and apparatus based on optimal transmission flow consistency. The method involves acquiring radar point cloud data and processing it to obtain first data; performing optimal transmission modeling based on the first data to obtain a transmission matrix; calculating a global flow consistency metric based on the transmission matrix; and performing clustering processing based on the first data and the global flow consistency metric to obtain target grouping results. By performing distribution modeling on radar point cloud data and utilizing optimal transmission to characterize the quality distribution relationship between point clouds, stable grouping and continuous representation of physical targets can be achieved even under conditions of point count variation and detection instability.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and in particular to a point cloud target grouping method and apparatus based on optimal transmission flow consistency. Background Technology

[0002] With the development of mobile robots and autonomous driving systems, environmental perception has become one of the key factors affecting system performance. Current mainstream robot perception systems typically integrate multiple sensors such as vision, LiDAR, and millimeter-wave radar. Among them, millimeter-wave radar is gradually becoming an important component of robot environmental perception due to its strong robustness in rain, fog, low light, and complex environments.

[0003] In millimeter-wave radar sensing, the raw echo signal is typically processed by range, Doppler, and angle measurements, and represented as a discrete point cloud—a set of detection points containing target range, azimuth, and amplitude information. However, compared to lidar, radar point clouds suffer from lower resolution, unstable detection, and significant variations in the number of points over time. The same physical target may appear as a set of scattering points of varying numbers in different time frames, and the position and amplitude of these points fluctuate due to noise, changes in viewing angle, and changes in target attitude.

[0004] Against this backdrop, achieving stable target grouping and association in multi-frame radar point clouds has become a key issue in robot perception systems. Existing methods mainly include target tracking methods based on spatial clustering (such as density clustering) and data association (such as nearest neighbor matching and joint probabilistic data association). However, these methods typically rely on the following assumptions: first, the same target has a stable distribution of points and positions in adjacent frames; and second, there is sufficient spatial separation between targets. In practical robot applications, these assumptions are often difficult to meet. For example, due to radar resolution limitations or changes in observation angle, the same target may degenerate from multiple points to a few points in different frames, or even experience point disappearance and re-emergence; simultaneously, multiple targets may be spatially close to each other, making it difficult for traditional methods to accurately distinguish between different targets.

[0005] Furthermore, traditional data association methods typically employ a "point-to-point" matching mechanism, assuming that a detection point in one frame can be matched one-to-one with a detection point in the next frame. However, in radar perception, due to the complex scattering characteristics of targets, the same target often exhibits a "many-to-one" or "one-to-many" relationship in different frames. This leads to the problem that association methods based on one-to-one matching are prone to erroneous associations or unstable associations, thereby affecting the accuracy of subsequent target tracking and motion estimation.

[0006] Therefore, there is an urgent need for a point cloud target grouping method and apparatus based on optimal transmission flow consistency to improve the above problems. Summary of the Invention

[0007] The purpose of this invention is to provide a point cloud target grouping method and apparatus based on optimal transmission flow consistency, which can achieve stable grouping and continuous representation of physical targets under conditions of point number variation and detection instability.

[0008] In a first aspect, the present invention provides a point cloud target grouping method based on optimal transmission flow consistency, comprising the steps of: acquiring radar point cloud data and processing it to obtain first data; performing optimal transmission modeling based on the first data to obtain a transmission matrix; calculating a global flow consistency metric based on the transmission matrix; and performing clustering processing based on the first data and the global flow consistency metric to obtain target grouping results.

[0009] Optionally, acquiring radar point cloud data and processing it to obtain the first data includes: acquiring radar point cloud data of continuous time frames and performing uniform discrete distribution processing to obtain the first data.

[0010] Optionally, performing optimal transmission modeling based on the first data to obtain the transmission matrix includes: performing optimal transmission modeling based on the first data to obtain the correspondence between point clouds in adjacent time frames, thereby obtaining the transmission matrix.

[0011] Optionally, calculating the flow consistency metric based on the transmission matrix includes: normalizing the transmission matrix to calculate the local flow consistency metric of different point clouds within the same time frame; and calculating the global flow consistency metric of point clouds between different time frames based on the local flow consistency metric.

[0012] Optionally, clustering based on the first data and the global flow consistency metric to obtain the target grouping result includes: constructing a weighted graph based on the first data using the global flow consistency metric, and performing clustering on the weighted graph to obtain the target grouping result.

[0013] Optionally, after obtaining the target grouping results, the method further includes: obtaining the motion trajectory of each target through weighted fusion based on the target grouping results.

[0014] Secondly, the present invention provides a point cloud target grouping device based on optimal transmission flow consistency, the device comprising modules / units for executing any of the possible design methods described in the first aspect above. These modules / units can be implemented in hardware or by hardware executing corresponding software.

[0015] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a program executable on the processor, and when the program is executed by the processor, the electronic device implements a method for performing any of the possible designs described above.

[0016] Fourthly, the present invention provides a readable storage medium storing a program, which, when executed, implements a method of any possible design of any of the above aspects.

[0017] Fifthly, the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0018] The beneficial effects of the method of the present invention are as follows: acquiring radar point cloud data and processing it to obtain first data; performing optimal transmission modeling based on the first data to obtain a transmission matrix; calculating a global flow consistency metric based on the transmission matrix; and performing clustering processing based on the first data and the global flow consistency metric to obtain target grouping results. By performing distribution modeling on radar point cloud data and using optimal transmission to characterize the quality distribution relationship between point clouds, stable grouping and continuous representation of physical targets can be achieved even under conditions of point number variation and detection instability. Attached Figure Description

[0019] Figure 1 A flowchart illustrating a point cloud target grouping method based on optimal transmission flow consistency, provided for an embodiment of the present invention;

[0020] Figure 2 A schematic diagram of a point cloud target grouping device based on optimal transmission flow consistency is provided for an embodiment of the present invention;

[0021] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention. Unless otherwise defined, the technical or scientific terms used herein should have the ordinary meaning understood by those skilled in the art. The terms "comprising" and similar expressions used herein mean that the element or object preceding the word covers the element or object listed following the word and its equivalents, but do not exclude other elements or objects.

[0023] The technical solutions of the embodiments of the present invention will be described below with reference to the accompanying drawings. In the description of the embodiments of the present invention, the terminology used in the following embodiments is for the purpose of describing specific embodiments only and is not intended to limit the present invention. The singular expressions “a,” “the,” “the,” and “this” are intended to also include expressions such as “one or more,” unless the context clearly indicates otherwise. It should also be understood that in the following embodiments of the present invention, “at least one” and “one or more” refer to one or more (including two). The term “and / or” is used to describe the relationship between related objects, indicating that three relationships can exist; for example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character “ / ” generally indicates that the preceding and following related objects are in an “or” relationship.

[0024] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in connection with that embodiment is included in one or more embodiments of the invention. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," and "in still other embodiments" appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including, but not limited to," unless otherwise specifically emphasized. The term "connection" includes both direct and indirect connections, unless otherwise stated. "First" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated.

[0025] In embodiments of the present invention, "exemplarily" or "for example" are used to indicate that they are examples, illustrations, or descriptions. Any embodiment or design described as "exemplarily" or "for example" in embodiments of the present invention should not be construed as being more preferred or advantageous than other embodiments or design solutions. Rather, the use of "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0026] like Figure 1 As shown, this invention provides a point cloud target grouping method based on optimal transmission flow consistency, including the following steps:

[0027] S101: Acquire radar point cloud data and process it to obtain the first data.

[0028] In some embodiments, acquiring radar point cloud data and processing it to obtain first data includes: acquiring radar point cloud data of continuous time frames and performing uniform discrete distribution processing to obtain first data.

[0029] S102, perform optimal transmission modeling based on the first data to obtain the transmission matrix.

[0030] In some embodiments, performing optimal transmission modeling based on the first data to obtain a transmission matrix includes: performing optimal transmission modeling based on the first data to obtain the correspondence between point clouds in adjacent time frames, thereby obtaining a transmission matrix.

[0031] S103, calculate the global flow consistency metric value based on the transmission matrix.

[0032] In some embodiments, calculating the flow consistency metric based on the transmission matrix includes: normalizing the transmission matrix to calculate the local flow consistency metric of different point clouds within the same time frame; and calculating the global flow consistency metric of point clouds between different time frames based on the local flow consistency metric.

[0033] S104, perform clustering processing based on the first data and the global flow consistency metric to obtain the target grouping result.

[0034] In some embodiments, clustering based on the first data and the global flow consistency metric to obtain the target grouping result includes: constructing a weighted graph based on the first data using the global flow consistency metric, and performing clustering on the weighted graph to obtain the target grouping result.

[0035] In other embodiments, after obtaining the target grouping results, the method further includes: obtaining the motion trajectory of each target by weighted fusion based on the target grouping results.

[0036] To facilitate understanding, this embodiment further elaborates on the specific implementation process of the above method in conjunction with a specific application scenario, which includes the following steps:

[0037] Step 1: Distribution Modeling of Radar Point Clouds

[0038] Suppose the robot operates in consecutive time frames. The radar point cloud data is acquired, and each frame of the point cloud is represented as follows:

[0039]

[0040] in, Indicates the position of a point (distance-angle or Cartesian coordinates). This represents the amplitude at the corresponding point. To perform uniform discrete distribution processing, the point cloud is represented in a discrete distribution form, yielding the first data:

[0041]

[0042] The above modeling transforms the original discrete point set into a weighted distribution, providing a unified representation for subsequent cross-frame relationship modeling.

[0043] Step 2: Optimal Transmission Modeling Across Frames

[0044] Optimal transmission modeling is performed based on the first data to obtain point clouds of adjacent time frames. and The correspondence between them:

[0045]

[0046] satisfy:

[0047]

[0048] The cost function is defined as follows:

[0049]

[0050] By solving this optimization problem, the transfer matrix is ​​obtained. This matrix describes the first The energy of each point in the frame at the th The allocation of points in a frame is used to establish the overall correspondence between point clouds across frames.

[0051] Step 3: Constructing Local Flow Consistency

[0052] In obtaining the transmission matrix Then, normalize it:

[0053]

[0054] For any two points in the same frame and By comparing their distribution structure in the next frame, the local flow consistency metric of different point clouds within the same time frame is calculated:

[0055]

[0056] When two points have similar allocation patterns during cross-frame transmission, their corresponding The value is relatively large, so it can be used to characterize the potential homology relationship between points.

[0057] Step 4: Global Flow Consistency Propagation

[0058] To further enhance the stability of target grouping, flow consistency is propagated across multiple frames, and a global flow consistency metric for the point cloud across different time frames is calculated based on the local flow consistency metric.

[0059]

[0060] This propagation process extends local inter-frame relationships to a longer time frame, thereby improving the robustness of grouping under conditions of point count changes and detection fluctuations.

[0061] Step 5: Target grouping based on global flow consistency

[0062] Using points in all time frames as nodes, global flow consistency is utilized. A weighted graph structure is constructed, and clustering is performed on the graph to group points with high consistency into the same category, resulting in target grouping. Each category corresponds to a physical target, realizing the transformation from point-level representation to target-level representation.

[0063] Step 6: Target Trajectory Construction

[0064] Based on the target grouping results, for the first... The first goal, in the first The set of points corresponding to the frame is denoted as Its spatial location is obtained through weighted fusion:

[0065]

[0066] The trajectory representation of the target over continuous time is obtained through the above method, thereby providing a stable input for subsequent motion estimation and decision-making.

[0067] This invention addresses the problems of point count variation, unstable detection, and complex observation conditions in multi-frame scenarios of robot radar point clouds. The key aspects are as follows:

[0068] (1) Unified representation of point clouds based on distributed modeling

[0069] Unlike traditional methods that directly process discrete points, this invention first models each frame of radar point cloud as a weighted discrete distribution:

[0070]

[0071] This representation transforms the discrete problem, which originally depended on the number of points and the stability of point positions, into a problem of modeling the relationship between distributions. This allows for processing within a unified framework even if there are changes in the number of points or perturbations in point positions between different frames, fundamentally reducing the impact of point-level instability on subsequent processing.

[0072] (2) Cross-frame overall association mechanism based on optimal transmission

[0073] This invention obtains the transmission matrix by solving the optimal transmission problem. Using the following constraints:

[0074]

[0075] This mechanism establishes a global association between two frames at the distribution level. Unlike the "point-to-point" matching in traditional target tracking methods, this mechanism allows for many-to-one and one-to-many relationships between points, ensuring that the association structure remains consistent even when the number of points for the same target changes in different frames, thereby avoiding association errors caused by mismatched point counts.

[0076] (3) Target discrimination criterion based on flow consistency

[0077] In obtaining the transmission matrix Subsequently, this invention constructs a normalized flow distribution. And define a local flow consistency metric:

[0078]

[0079] By leveraging the similarity of the allocation structure of different points during cross-frame transmission, the same target can be identified. Compared with traditional clustering methods based on spatial distance or density, this identification method introduces temporal dimension information, so that target grouping no longer depends on the spatial structure of a single frame, but on the consistency of cross-frame evolution, thereby significantly improving the stability and accuracy of grouping.

[0080] (4) Global flow consistency propagation mechanism

[0081] To further enhance the robustness of the grouping results, this invention achieves the following:

[0082]

[0083] Propagating flow consistency across multiple frames extends local inter-frame relationships into global flow consistency constraints over a long timeframe. This mechanism effectively suppresses the impact of single-frame detection errors or short-term fluctuations, resulting in more stable and continuous target grouping results over time.

[0084] (5) Construction of target-level representation based on global flow consistency

[0085] By clustering points with high flow consistency, this invention achieves the transformation from point-level representation to target-level representation, and further constructs the target trajectory through weighted fusion:

[0086]

[0087] Compared to traditional target tracking which directly outputs discrete point trajectories, this method can generate continuous and smooth target representations even when the number of points changes and detection is unstable, providing a more reliable input for subsequent motion estimation.

[0088] (6) Overcoming key limitations of traditional methods

[0089] Compared with existing methods, the present invention achieves breakthroughs in the following aspects:

[0090] At the data association level, the process has been expanded from "point-to-point matching" to "distribution-to-distribution association," avoiding matching difficulties caused by inconsistent point numbers.

[0091] At the target grouping level, the approach has been expanded from "spatial clustering" to "spatiotemporal consistency modeling," which improves the grouping capability in complex scenarios.

[0092] At the model representation level, the "discrete point representation" has been expanded to "continuous distribution evolution," enhancing the adaptability to changes in the target structure.

[0093] In terms of cross-frame target grouping and representation of robot radar point clouds, the embodiments of the present invention have the following significant effects and advantages compared with traditional methods:

[0094] (1) Significantly improves grouping stability under scenarios with varying point counts.

[0095] In actual robot perception, the same physical target often appears as a set of scattering points with inconsistent point counts across different time frames due to factors such as changes in observation angle, resolution limitations, or occlusion. Traditional point-to-point matching-based tracking methods rely on stable point count relationships, which are prone to association errors when "many-to-one" or "one-to-many" relationships occur. This invention establishes a quality allocation relationship between distributions through an optimal transmission model, enabling stable association between different frames without satisfying point count consistency constraints, thereby maintaining the consistency of target grouping even when point counts change significantly.

[0096] (2) Improve robustness in complex environments

[0097] In the presence of noise interference, weak targets, or unstable detection, traditional spatial clustering methods (such as distance- or density-based clustering) rely solely on single-frame spatial information, making them susceptible to misgrouping due to local outliers. This invention incorporates cross-frame optimal transmission relationships and flow consistency metrics into the grouping process, ensuring that the target grouping result depends not only on the current frame's spatial structure but also on the consistency of multi-frame evolution. This effectively suppresses the interference of random noise and isolated points on the grouping results, significantly improving system robustness.

[0098] (3) Effectively avoid data association errors

[0099] Traditional target tracking methods typically employ nearest neighbor or probabilistic association strategies, which are prone to mismatches or target swapping when multiple targets are close together or when point cloud distributions are complex. This invention constructs a transfer matrix to describe the overall relationship between two frames of point clouds within a globally optimal framework. This eliminates the reliance on locally optimal matching for data association, effectively reducing the probability of misassociations and preventing erroneous crossovers and confusion between targets.

[0100] (4) Improve the temporal continuity of target groups

[0101] Because this invention introduces a cross-frame consistency propagation mechanism, target grouping is based not only on information from adjacent frames but also comprehensively considers the transmission relationships between multiple frames, thus enabling the formation of a continuous and stable target representation in the time dimension. Compared to methods that rely solely on information from a single frame or adjacent frames, this mechanism effectively mitigates the impact of short-term detection omissions or local anomalies, making the target grouping results smoother and more consistent over time.

[0102] (5) Improve the quality of target trajectory and subsequent parameter estimation

[0103] By weighted fusion of the set of points corresponding to the same target, the target trajectory constructed by this invention has better continuity and stability. Compared with tracking using single points or discrete detection points directly, this method can reduce trajectory jitter caused by point-level fluctuations, thus providing more reliable input for subsequent motion parameter estimation (such as velocity, Doppler frequency, or direction estimation).

[0104] (6) Applicable to complex robot application scenarios

[0105] The method of this invention can be adapted to the following typical robot perception scenarios:

[0106] Multi-target densely distributed environment;

[0107] Scenarios where point cloud resolution limitations cause fluctuations in point count;

[0108] Dynamic environments with occlusion or partial missing targets;

[0109] Long-term continuous sensing tasks.

[0110] In the above scenarios, the present invention can maintain a stable target grouping effect, thereby improving the robot's perception ability and decision-making reliability in complex environments.

[0111] (7) Summary of overall performance improvement

[0112] In summary, this invention achieves a shift from "point-level matching" to "distribution-level modeling" and an improvement from "spatial clustering" to "spatiotemporally consistent grouping" by modeling radar point clouds as a distribution and establishing cross-frame relationships using optimal transmission. This method exhibits higher stability and accuracy under conditions of varying point counts, unstable detection, and complex environments, providing a more robust means of target grouping and representation for robot perception systems.

[0113] like Figure 2 As shown, based on the above method, the present invention provides a point cloud target grouping device based on optimal transmission flow consistency, comprising: an acquisition unit 201, used to acquire radar point cloud data and process it to obtain first data; a modeling unit 202, used to perform optimal transmission modeling based on the first data to obtain a transmission matrix; a calculation unit 203, used to calculate a global flow consistency metric value based on the transmission matrix; and a processing unit 204, used to perform clustering processing based on the first data and the global flow consistency metric value to obtain target grouping results.

[0114] It should be understood that all relevant content of each step involved in the above method embodiments can be referenced to the functional description of the corresponding functional module, and will not be repeated here. Furthermore, the use of suffixes such as "module," "component," or "unit" to represent elements is merely for the purpose of illustrative purposes and has no specific meaning in itself. Therefore, "module," "component," or "unit" can be used interchangeably. Terminals can be implemented in various forms. For example, the terminals described in this invention may include mobile terminals such as mobile phones, tablets, laptops, handheld computers, personal digital assistants (PDAs), portable media players (PMPs), navigation devices, wearable devices, smart bracelets, pedometers, etc., as well as fixed terminals such as digital TVs and desktop computers. The following description will use mobile terminals as examples; those skilled in the art will understand that, in addition to elements specifically designed for mobile purposes, the construction according to embodiments of the present invention can also be applied to fixed-type terminals.

[0115] In other embodiments of the present invention, an electronic device 300 is disclosed, such as... Figure 3 As shown, the device may include: one or more processors 301; memory 302; display 303; one or more application programs (not shown); and one or more computer programs 304. These devices can be connected via one or more communication buses 305. The one or more computer programs 304 are stored in the memory 302 and configured to be executed by the one or more processors 301. The one or more computer programs 304 include instructions that can be used to perform actions such as... Figure 1 Each step in the corresponding embodiment.

[0116] Processor 301 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0117] The memory 302 can be an internal storage unit of the electronic device 300, such as a hard disk or RAM of the electronic device 300. The memory 302 can also be an external storage device of the electronic device 300, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or FlashCard equipped on the electronic device 300. Furthermore, the memory 302 can include both internal and external storage units of the electronic device 300. The memory 302 is used to store computer programs and other programs and data required by the electronic device. The memory 302 can also be used to temporarily store data that has been output or will be output.

[0118] The computer program 304 can be divided into one or more modules / units. The one or more modules / units can be a series of computer program instruction segments that can perform a specific function. The instruction segments are used to describe the execution process of the computer program 304 in the electronic device 300.

[0119] In addition to the above-described structure, those skilled in the art will understand that Figure 3This is merely an example of electronic device 300 and does not constitute a limitation on electronic device 300. Electronic device 300 may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device may also include input / output devices, network access devices, buses, etc.

[0120] Those skilled in the art will understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the functions described above can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0121] Based on the above embodiments, the present invention also discloses a computer-readable storage medium having at least one computer program stored thereon, wherein the computer program, when executed by a processor, implements the methods described in the foregoing embodiments.

[0122] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof. The storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. This available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state drive (SSD)).

[0123] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0124] Although the embodiments of the present invention have been described in detail above, it will be apparent to those skilled in the art that various modifications and variations can be made to these embodiments. The above descriptions are merely embodiments of the present invention and do not limit the patent scope of the present invention. However, it should be understood that such modifications and variations fall within the scope and spirit of the present invention. Moreover, the present invention described herein may have other embodiments and can be implemented or realized in various ways. All equivalent transformations made based on the content of this specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of this invention.

Claims

1. A point cloud target grouping method based on optimal transmission flow consistency, characterized in that, Including the following steps: Acquire radar point cloud data and process it to obtain the first data; Based on the first data, optimal transmission modeling is performed to obtain the transmission matrix; The global flow consistency metric is calculated based on the aforementioned transfer matrix; Clustering is performed based on the first data and the global flow consistency metric to obtain the target grouping results.

2. The method according to claim 1, characterized in that, The first data obtained by acquiring and processing radar point cloud data includes: The radar point cloud data of continuous time frames is acquired and subjected to uniform discrete distribution processing to obtain the first data.

3. The method according to claim 1, characterized in that, Based on the first data, optimal transmission modeling is performed, resulting in a transmission matrix including: Based on the first data, optimal transmission modeling is performed to obtain the correspondence between point clouds of adjacent time frames, and the transmission matrix is ​​obtained.

4. The method according to claim 1, characterized in that, The flow consistency metric values ​​calculated based on the aforementioned transfer matrix include: Based on the transmission matrix, normalization processing is performed to calculate the local flow consistency metric value of different point clouds within the same time frame; Calculate the global flow consistency metric of the point cloud across different time frames based on the local flow consistency metric.

5. The method according to claim 1, characterized in that, Clustering is performed based on the first data and the global flow consistency metric to obtain the target grouping results, including: Based on the first data, a weighted graph is constructed using the global flow consistency metric, and the weighted graph is clustered to obtain the target grouping result.

6. The method according to any one of claims 1-5, characterized in that, After obtaining the target grouping results, the following is also included: The motion trajectory of each target is obtained by weighted fusion based on the target grouping results.

7. A point cloud target grouping device based on optimal transmission flow consistency, used in the method of any one of claims 1-6, characterized in that, include: The acquisition unit is used to acquire radar point cloud data and process it to obtain the first data. A modeling unit is used to perform optimal transmission modeling based on the first data to obtain a transmission matrix; The calculation unit is used to calculate the global flow consistency metric value based on the transmission matrix; The processing unit is used to perform clustering based on the first data and the global flow consistency metric to obtain the target grouping result.

8. An electronic device, characterized in that, The device includes a memory and a processor, wherein the memory stores a program that can run on the processor, and when the program is executed by the processor, causes the electronic device to perform the method of any one of claims 1-6.

9. A readable storage medium storing a program, characterized in that, When the program is executed, it implements the method of any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1-6.

Images