A target detection method and device based on density perception topology analysis

Abstract

The application provides a target detection method and device based on density perception topological structure analysis, acquires radar point cloud data, and calculates corresponding density perception distance values; multi-scale topological structures are constructed according to the density perception distance values, and connected components under different scales are obtained; structure stability measurement values are calculated according to the connected components; target detection is performed according to the structure stability measurement values, and a detection result is output. By introducing density information in a multi-scale topological analysis framework and combining structure stability and geometric features, the relationship between points is adaptively modulated, so that the expression ability of the topological structure is maintained while false connections are suppressed, and the effective distinction between targets and clutter is realized by using structure stability, thereby realizing the robust detection and classification of targets.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and in particular to a target detection method and apparatus based on density-sensing topology analysis. Background Technology

[0002] Sea surface target detection and classification is one of the key technologies in marine monitoring, maritime security, unmanned system perception, and shipborne and airborne radar applications. With the development of millimeter-wave radar, synthetic aperture radar (SAR), and multi-dimensional sensing radar technologies, radar systems can acquire higher resolution sea surface echo information. However, at the same time, the complex scattering environment of the sea surface also places higher demands on target detection and classification algorithms.

[0003] Compared to land-based scenarios, the sea surface exhibits significant dynamic characteristics. Its echoes primarily originate from wave undulations, wind fields, and variations in sea surface roughness. Sea surface scattering typically displays a strong non-Gaussian distribution, significant time-varying characteristics, and Doppler broadening, accompanied by both localized strong reflections and random weak scattering. This complex scattering characteristic means that sea surface clutter not only possesses high energy but also exhibits a structured yet unstable distribution pattern in both space and time, thus severely interfering with the detection of weak targets such as small vessels and buoys.

[0004] Existing methods for sea surface target detection largely rely on statistical models or signal features for processing. For example, the Constant False Alarm Rate (CFAR) method adaptively sets a detection threshold by estimating the statistical characteristics of local clutter. However, such methods typically depend on the accuracy of the clutter distribution model, while actual sea surface clutter often fails to meet ideal statistical assumptions, leading to high false alarm rates or missed detections in complex sea conditions. Methods based on Doppler features distinguish targets from the sea surface by utilizing the velocity difference in the velocity domain. However, because waves themselves introduce Doppler spread, small targets at low velocities are often difficult to distinguish effectively from the sea surface, resulting in decreased detection performance. Furthermore, methods based on clustering or geometric structures, such as DBSCAN or K-means, achieve target segmentation through point cloud spatial distribution. However, these methods are sensitive to parameters and are easily affected by sparse point connections in sea surface scenarios, producing a "bridging effect"—that is, different targets or targets and sea surface clutter are incorrectly connected, thus reducing the accuracy of detection and classification.

[0005] Essentially, the methods described above rely primarily on signal energy or local density characteristics for discrimination, failing to effectively utilize the differences in structural stability between targets and the sea surface. In reality, while sea clutter possesses a certain spatial structure within a local area, this structure exhibits significant instability with scale variations and temporal evolution; whereas real targets typically correspond to relatively stable spatial structures, demonstrating stronger connectivity maintenance capabilities across multiple scales. Therefore, relying solely on energy or local density is insufficient for effectively distinguishing targets from sea clutter.

[0006] In recent years, topological data analysis methods have provided new approaches for modeling complex structural data. By characterizing the connectivity changes of data at multiple scales, stable structural features can be extracted. However, traditional topological analysis methods typically construct relationships between data based on Euclidean distance, which is susceptible to the influence of sparse points in marine scenarios, leading to the incorrect connection of unrelated structures and weakening the discriminative power of topological features. Furthermore, these methods lack characterization of local density information, making it difficult to effectively suppress bridging effects, and thus still have certain limitations in complex marine environments.

[0007] Therefore, there is an urgent need for a target detection method and device based on density-aware topology analysis to improve the above problems. Summary of the Invention

[0008] The purpose of this invention is to provide a target detection method and device based on density-sensing topology analysis, which can effectively suppress clutter interference and improve detection accuracy.

[0009] In a first aspect, the present invention provides a target detection method based on density-aware topology analysis, comprising the steps of: acquiring radar point cloud data and calculating the corresponding density-aware distance value; constructing a multi-scale topology based on the density-aware distance value to obtain connected components at different scales; calculating a structural stability metric based on the connected components; performing target detection based on the structural stability metric and outputting the detection result.

[0010] Optionally, constructing a multi-scale topology based on the density-sensing distance value to obtain connected components at different scales includes: setting a scale parameter, establishing connectivity between scattering points in the point cloud data based on the density-sensing distance value, and forming a topology filtering process as the scale parameter value increases to obtain connected components at different scales.

[0011] Optionally, acquiring radar point cloud data and calculating the corresponding density sensing distance value includes: acquiring radar point cloud data and performing local density estimation to obtain a local density value; and calculating the corresponding density sensing distance value based on the local density value.

[0012] Optionally, calculating the structural stability metric based on the connected components includes: calculating a first scale based on the connected components; calculating a second scale based on the connected components; calculating the structural stability metric based on the first scale and the second scale; and / or the structural stability metric includes structural stability margin and scale separation ratio.

[0013] Optionally, target detection is performed based on the structural stability metric, and the output detection result includes: threshold discrimination based on the structural stability metric to obtain the corresponding target structure; extraction of multidimensional feature values ​​through principal component analysis based on the target structure and classification, and output detection result; and / or the multidimensional feature values ​​include geometric structure feature values, motion consistency feature values ​​and spatial scale feature values; the detection result includes target location, category and related feature information.

[0014] Optionally, the density sensing distance value is:

[0015]

[0016] in, For the point cloud data The eigenvectors of each scattering point; For the point cloud data The eigenvectors of each scattering point; To adjust the parameters; For the first Local density values ​​at each scattering point; For the first Local density values ​​of scattering points; the point cloud data is , This represents the total number of scattering points in the point cloud data. For feature dimensions; , The number of nearest neighbors. It is a set of nearest neighbors; the feature vector includes one or more of the following: distance, azimuth angle, and Doppler velocity.

[0017] Secondly, the present invention provides a target detection device based on density-aware topology analysis, 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.

[0018] 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.

[0019] 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.

[0020] Fifthly, the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0021] The beneficial effects of the method of this invention are as follows: acquiring radar point cloud data and calculating the corresponding density-sensing distance value; constructing a multi-scale topology based on the density-sensing distance value to obtain connected components at different scales; calculating a structural stability metric based on the connected components; performing target detection based on the structural stability metric and outputting the detection result. By introducing density information into the multi-scale topology analysis framework and combining structural stability and geometric features, the relationships between points are adaptively modulated, thereby suppressing erroneous connectivity while maintaining the expressive power of the topology, and effectively distinguishing between targets and clutter using structural stability, thus achieving robust target detection and classification. Attached Figure Description

[0022] Figure 1 A flowchart illustrating a target detection method based on density-aware topology analysis provided in an embodiment of the present invention;

[0023] Figure 2 This is a schematic diagram of a target detection device based on density-sensing topology analysis provided in an embodiment of the present invention;

[0024] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0025] 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.

[0026] 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.

[0027] 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.

[0028] 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.

[0029] like Figure 1 As shown, this invention provides a target detection method based on density-aware topology analysis, comprising the following steps:

[0030] S101: Acquire radar point cloud data and calculate the corresponding density sensing distance value.

[0031] In some embodiments, acquiring radar point cloud data and calculating the corresponding density sensing distance value includes: acquiring radar point cloud data and performing local density estimation to obtain a local density value; and calculating the corresponding density sensing distance value based on the local density value.

[0032] S102, construct a multi-scale topology based on the density-aware distance value to obtain connected components at different scales.

[0033] In some embodiments, constructing a multi-scale topology based on the density-sensing distance value to obtain connected components at different scales includes: setting a scale parameter, establishing connectivity between scattering points in the point cloud data based on the density-sensing distance value, and forming a topology filtering process as the scale parameter value increases to obtain connected components at different scales.

[0034] In some specific embodiments, the density sensing distance value is:

[0035]

[0036] in, For the point cloud data The eigenvectors of each scattering point; For the point cloud data The eigenvectors of each scattering point; To adjust the parameters; For the first Local density values ​​at each scattering point; For the first Local density values ​​of scattering points; the point cloud data is , This represents the total number of scattering points in the point cloud data. For feature dimensions; , The number of nearest neighbors. It is the set of nearest neighbors.

[0037] In some more specific embodiments, the feature vector includes one or more of distance, azimuth angle, and Doppler velocity.

[0038] S103, Calculate the structural stability metric value based on the connected components.

[0039] In some embodiments, calculating a structural stability metric based on the connected components includes: calculating a first scale based on the connected components; calculating a second scale based on the connected components; and calculating a structural stability metric based on the first scale and the second scale.

[0040] In other embodiments, the structural stability metrics include structural stability margin and scale separation ratio.

[0041] S104, Perform target detection based on the structural stability metric and output the detection result.

[0042] In some embodiments, target detection is performed based on the structural stability metric, and the output detection result includes: threshold discrimination based on the structural stability metric to obtain the corresponding target structure; multidimensional feature values ​​are extracted and classified based on the target structure through principal component analysis, and the detection result is output.

[0043] In other embodiments, the multidimensional feature values ​​include geometric structure feature values, motion consistency feature values, and spatial scale feature values; the detection results include target location, category, and related feature information.

[0044] The advantage of this invention is that by introducing density information into a multi-scale topology analysis framework and combining structural stability and geometric features, the relationship between points is adaptively modulated, thereby suppressing erroneous connectivity while maintaining the topological structure expressiveness, and using structural stability to effectively distinguish between targets and clutter, thus achieving robust detection and classification of targets.

[0045] To facilitate understanding, this embodiment further elaborates on the specific implementation process of the above method in conjunction with a specific application scenario. Taking radar detection of sea-level targets as an example, the specific steps include:

[0046] (1) Radar point cloud data acquisition

[0047] Acquiring point cloud data when radar detects targets at sea level:

[0048]

[0049] in, For the point cloud data The feature vector of each scattering point contains one or more of the following: range, azimuth angle, and Doppler velocity; This represents the total number of scattering points in the point cloud data. For feature dimensions.

[0050] (2) Local density estimation

[0051] Based on the The nearest neighbor set of scattering points Calculate the local density value:

[0052]

[0053] in, For the point cloud data The eigenvectors of each scattering point; This represents the number of nearest neighbors.

[0054] (3) Density-sensing distance construction

[0055] Calculate the corresponding density sensing distance value based on the local density value:

[0056]

[0057] in, The parameters are adjusted to suppress erroneous connections in sparse regions; For the point cloud data The eigenvectors of each scattering point; For the first The local density value of each scattering point.

[0058] (4) Construction of multi-scale topology

[0059] Define scale parameters Connectivity is constructed based on the density-sensing distance values:

[0060]

[0061] along with As the numerical value increases, a topological filtering process is formed, resulting in connected components at different scales.

[0062] (5) Candidate structure extraction

[0063] The connected components formed at any scale are denoted as candidate structures. .

[0064] (6) Calculation of internal connectivity scale (i.e., first scale)

[0065] Define the internal connectivity scale of the structure:

[0066]

[0067] (7) Calculation of external connectivity scale (i.e., second scale)

[0068] Define the external connection dimensions of the structure:

[0069]

[0070] (8) Construction of structural stability metrics

[0071] Define structural stability metrics:

[0072]

[0073] Wherein, S is the structural stability margin, which represents the "margin" by which the target structure can maintain independent existence at multiple scales, reflecting the scale interval between the formation of the structure and its connection to the outside world; R is the scale separation ratio, which represents the relative separation degree between the internal and external scales of the structure and is used to measure the strength of the structure's stability.

[0074] (9) Target detection

[0075] Based on threshold judgment:

[0076]

[0077] Otherwise, it is classified as sea surface clutter.

[0078] (10) Target Classification

[0079] The detected target structure Extract the following multidimensional feature values ​​and classify them:

[0080] 1. Multidimensional Feature Value Extraction

[0081] (a) Geometric structural eigenvalues

[0082] Principal component analysis (PCA) was performed on the target structure to obtain eigenvalues:

[0083]

[0084] definition:

[0085]

[0086]

[0087]

[0088] in, Characterizing linear structures; Characterizing planar structures; Characterizes discrete or noisy structures;

[0089] (b) Motion consistency characteristic value

[0090] The Doppler at the point is ,calculate:

[0091]

[0092] Used to characterize whether a structure satisfies the rigid body motion characteristics.

[0093] (c) Spatial scale characteristic value

[0094] Define spatial scale eigenvalues:

[0095]

[0096] Used to distinguish between large goals and small goals.

[0097] 2. Classification and identification

[0098] Based on the extracted multidimensional feature values, discrimination rules are constructed, for example:

[0099] when Larger higher and When the size is small, it is identified as a ship target;

[0100] When the structural scale is small and the stability is high, it is judged as a small target (such as a buoy).

[0101] When stability is low or motion is inconsistent, it is judged as sea surface clutter or interference.

[0102] (11) Output results

[0103] Output the target's location, category, and related feature information.

[0104] To address the problems of complex clutter, easily confused structures, and insufficient robustness of traditional methods in sea surface target detection, the core innovation of this invention lies in the following aspects:

[0105] (1) Fusion modeling of density-aware distance and topology

[0106] This invention introduces local density information into the topology analysis process for the first time, by constructing density-aware distance values:

[0107]

[0108] This invention achieves adaptive modulation of relationships between points. In sparse regions, distances are amplified, while in dense regions, the original scale is maintained, effectively suppressing erroneous connections caused by sparse points. Compared to traditional topology analysis methods based on Euclidean distance, this invention avoids the "bridging effect" of sea surface scattering points on the target structure, thus ensuring the correctness and stability of the topology construction.

[0109] (2) Structure extraction mechanism based on multi-scale topological evolution

[0110] This invention constructs scale parameters The topological filtering process analyzes the connectivity of point clouds at multiple scales, without relying on a single scale parameter for clustering or segmentation. By tracking the generation and merging process of connected components during scale changes, the target structure is defined as a stable connected structure within a certain scale range, realizing the transformation from "single-scale clustering" to "multi-scale structure analysis." This mechanism avoids the parameter sensitivity problem of traditional clustering methods and improves the adaptability of the method in complex sea surface environments.

[0111] (3) Structural stability discrimination method based on separation of internal and external scales

[0112] This invention proposes a structural stability modeling method based on internal connectivity scale and external connectivity scale:

[0113]

[0114] in, It is the smallest scale at which the structure is completely interconnected. The scale of the first connection between the structure and the outside.

[0115] Based on the above definition, the target detection problem is transformed into a scale separation problem. That is, real targets correspond to a structure that is "compact internally and dispersed externally," while sea clutter lacks obvious scale separation characteristics. This method is fundamentally different from traditional discrimination methods based on energy or density, and can maintain high detection accuracy under low signal-to-noise ratio and complex sea conditions.

[0116] (4) Target detection mechanism based on topological stability

[0117] This invention proposes using structural stability as the core criterion for target detection. By determining the presence range of the structure at multiple scales, it distinguishes between the target and sea surface clutter. Unlike traditional CFAR methods that rely on amplitude information, this invention utilizes the stability of the structure in scale space as the discrimination criterion, thereby effectively avoiding false alarms caused by strong sea surface clutter, and is particularly suitable for detecting small targets with low observability.

[0118] (5) Joint classification method of topological structure and geometric features

[0119] In the target classification stage, this invention combines topological stability features with geometric structure features, specifically including: structural morphology features (linearity, flatness, and dispersion) based on eigenvalue decomposition, motion consistency features based on Doppler information, and scale features based on spatial distribution. Through multi-dimensional feature fusion, effective differentiation of different types of sea surface targets (such as ships and buoys) is achieved. Compared to methods relying on only a single feature, this invention can more comprehensively characterize target attributes, improving classification accuracy and robustness.

[0120] (6) Paradigm shift from "energy discrimination" to "structure discrimination"

[0121] The core concept of this invention lies in:

[0122] Target = Structures that exist stably across multiple scales

[0123] Unlike traditional methods that define "target = high-energy echo," this invention introduces topological analysis, elevating the detection criterion from signal amplitude to structural stability, thus fundamentally shifting the detection paradigm. This approach is particularly suitable for situations at sea where targets and clutter have similar energies but significantly different structural characteristics.

[0124] Compared with the prior art, the advantages of the embodiments of the present invention are as follows:

[0125] (1) Effectively suppresses sea surface clutter interference and improves detection accuracy

[0126] This invention modulates the relationships between points using density-sensing distance, amplifying distance in sparse regions to effectively suppress erroneous connections caused by sea surface scattering points and avoid the generation of false structures due to clutter in traditional methods. Compared to Euclidean distance-based clustering or topology methods, this invention maintains accurate structure partitioning under complex sea conditions, thereby significantly reducing the false alarm rate and improving the accuracy of target detection.

[0127] (2) Eliminate bridging effect and improve structural separation capability

[0128] To address the common "bridging effect" in marine scenarios—where sparse scattering points incorrectly connect different targets or targets with clutter—this invention effectively suppresses the connectivity of low-density regions through a density modulation mechanism, ensuring proper separation between different structures. Therefore, this invention can accurately separate target structures in multi-target scenarios and when targets are surrounded by nearby marine clutter, avoiding the erroneous merging phenomenon that occurs in traditional clustering methods.

[0129] (3) Reduce parameter sensitivity and improve environmental adaptability

[0130] This invention employs a multi-scale topology analysis method, which does not rely on a single-scale parameter for target detection. Instead, it identifies stable structures by analyzing the evolution of structures across different scales. Compared to traditional methods based on fixed thresholds or neighborhood parameters (such as CFAR or DBSCAN), this invention is insensitive to parameter selection and can maintain stable performance under different sea states and signal-to-noise ratios, significantly improving the method's versatility and adaptability.

[0131] (4) Improve weak target detection capability based on structural stability detection mechanism

[0132] This invention transforms the target detection problem into a multi-scale stability discrimination problem by defining structural stability margin and structural stability ratio. Since real targets typically exhibit a "compact internal structure and dispersed external structure" in space, i.e., they have a small internal connectivity scale and a large external connectivity scale, while sea clutter lacks obvious scale separation characteristics, this invention can effectively detect weak targets under low signal-to-noise ratio conditions and prevent them from being overwhelmed by strong clutter.

[0133] (5) Achieve multi-feature fusion to improve target classification accuracy.

[0134] This invention integrates topological stability features, geometric structure features, and motion consistency features in the target classification stage to perform multi-dimensional modeling of targets. Compared with classification methods that rely on only a single feature (such as energy or velocity), this invention can more comprehensively characterize target attributes, thereby achieving accurate classification of different types of sea surface targets (such as ships, buoys, etc.) and improving classification accuracy and robustness.

[0135] (6) It is suitable for complex and dynamic sea surface environments and has good robustness.

[0136] This invention makes judgments based on structural stability rather than instantaneous signal characteristics, thus effectively addressing the impact of dynamic changes in the sea surface. In complex environments such as wave fluctuations and wind speed changes, although sea surface clutter exhibits structure locally, it is unstable in scale space, while the actual target structure is persistent. This invention can utilize this difference to achieve stable detection, significantly improving the robustness of the system.

[0137] (7) Realize the transformation of the detection paradigm from "energy discrimination" to "structure discrimination".

[0138] Traditional sea surface target detection methods primarily rely on echo energy or statistical characteristics for discrimination. This invention, however, introduces topological analysis, elevating the detection basis to the structural level, thus shifting from "energy-based detection" to "structural stability-based detection." This transformation enables the invention to effectively distinguish between targets and clutter even when their energies are similar, giving it greater theoretical advantages and engineering application value.

[0139] In summary, this invention achieves high-precision detection and classification of targets in complex sea environments through density-sensing topology analysis, effectively solving problems such as high false alarm rate, strong parameter sensitivity, and difficulty in detecting weak targets in traditional methods, and has significant engineering application value.

[0140] like Figure 2As shown, based on the above method, the present invention provides a target detection device based on density-sensing topology analysis, comprising: an acquisition unit 201 for acquiring radar point cloud data and calculating the corresponding density-sensing distance value; a construction unit 202 for constructing a multi-scale topology based on the density-sensing distance value to obtain connected components at different scales; a calculation unit 203 for calculating a structural stability metric value based on the connected components; and a detection unit 204 for performing target detection based on the structural stability metric value and outputting the detection result.

[0141] 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.

[0142] 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.

[0143] 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.

[0144] The memory 302 can be a first storage unit of the electronic device 300, such as a hard disk or memory of the electronic device 300. The memory 302 can also be a second storage device of the electronic device 300, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 300. Furthermore, the memory 302 can include both the first and second 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.

[0145] 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.

[0146] In addition to the above-described structure, those skilled in the art will understand that Figure 3 This 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.

[0147] Those skilled in the art will clearly 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 above functions can be assigned to different functional units and modules as needed, that is, the first 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.

[0148] 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.

[0149] 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)).

[0150] The descriptions of the processes or structures corresponding to the above-mentioned 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.

[0151] 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 description and drawings of the present invention, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A target detection method based on density-aware topology analysis, characterized in that, Including the following steps: Acquire radar point cloud data and calculate the corresponding density sensing distance value; A multi-scale topology is constructed based on the density-aware distance value to obtain connected components at different scales; Calculate the structural stability metric value based on the connected components; Target detection is performed based on the structural stability metric, and the detection results are output. The density sensing distance value is: in, For the point cloud data The feature vector of each scattering point; For the point cloud data The feature vector of each scattering point; To adjust the parameters; For the first Local density values ​​at each scattering point; For the first Local density values ​​at each scattering point; The point cloud data is , This represents the total number of scattering points in the point cloud data. For feature dimensions; , The number of nearest neighbors. For the set of nearest neighbors; The feature vector includes one or more of the following: distance, azimuth angle, and Doppler velocity.

2. The method according to claim 1, characterized in that, Based on the density-aware distance values, a multi-scale topology is constructed, resulting in connected components at different scales, including: By setting a scale parameter, the connectivity between scattering points in the point cloud data is established based on the density sensing distance value. As the value of the scale parameter increases, a topological filtering process is formed to obtain connected components at different scales.

3. The method according to claim 1, characterized in that, Acquiring radar point cloud data and calculating the corresponding density sensing distance values ​​includes: Acquire radar point cloud data and perform local density estimation to obtain local density values; The corresponding density sensing distance value is calculated based on the local density value.

4. The method according to claim 1, characterized in that, Calculating the structural stability metric based on the connected components includes: Calculate the first scale based on the connected components; Calculate the second scale based on the connected components; Calculate the structural stability metric value based on the first and second scales; The structural stability metrics include structural stability margin and scale separation ratio; Wherein, the first scale is the internal connectivity scale, the second scale is the external connectivity scale, and the structural stability margin is the difference between the external connectivity scale and the internal connectivity scale.

5. The method according to any one of claims 1-4, characterized in that, Target detection is performed based on the structural stability metric, and the output detection results include: The target structure is obtained by thresholding based on the structural stability metric. Based on the target structure, multidimensional feature values ​​are extracted and classified using principal component analysis, and the detection results are output. And / or the multidimensional feature values ​​include geometric structure feature values, motion consistency feature values, and spatial scale feature values; The detection results include the target location, category, and related feature information.

6. A target detection device based on density-aware topology analysis, used in the method of any one of claims 1-4, characterized in that, include: The acquisition unit is used to acquire radar point cloud data and calculate the corresponding density sensing distance value; A construction unit is used to construct a multi-scale topology based on the density-aware distance value to obtain connected components at different scales. A calculation unit is used to calculate a structural stability metric value based on the connected components; The detection unit is used to perform target detection based on the structural stability metric and output the detection result.

7. 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 1-4.

8. A readable storage medium storing a program, characterized in that, When the program is executed, it implements the method of any one of claims 1-4.

9. 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 1-4.

Images