Aerial target formation identification method and device

Abstract

The invention relates to an aerial target formation recognition method and device, and the method comprises the steps: building a multi-mode aerial target detection hardware system which comprises a multi-type visual sensor and a dynamic visual sensor; preprocessing the multi-modal input data, and identifying an aerial target through a target detection algorithm; an air target formation knowledge base is constructed, dynamic association between air targets is modeled through a hypergraph, and possible formations of a target group are compared and recognized; a target formation dynamic sensing model is established, an aerial target formation knowledge base is updated in real time according to formation changes, information such as the position and the speed of a target group is monitored in real time, and the method is used for ground target situation sensing real-time analysis. Therefore, the problems that in the prior art, multi-source data cannot be fused efficiently, high-precision formation recognition cannot be achieved in a complex dynamic environment, and the variable flight mode requirements in a tactical environment cannot be met easily are solved.

Description

Technical Field

[0001] The present application relates to the technical field of aerial target recognition, and in particular to a method and device for recognizing aerial target formations. Background Art

[0002] Traditional aerial target recognition systems primarily rely on a single type of sensor, such as radar or optical cameras, to detect and track aerial targets. However, these methods often struggle to ensure accurate and real-time recognition in complex flight environments, dense formations of multiple targets, rapid dynamic changes, and occlusions. Furthermore, the limited data processing capabilities of a single sensor make it difficult to fully capture the multi-dimensional characteristics of aerial targets, resulting in insufficient robustness in recognition results.

[0003] In recent years, with the development of multi-sensor fusion technology and artificial intelligence algorithms, aerial target formation recognition technology has made considerable progress. Multimodal sensor systems can simultaneously acquire multiple types of data, such as radar, infrared, and optical data, improving target detection performance through data fusion. However, how to efficiently fuse multi-source data, extract meaningful features, and achieve high-precision formation recognition in complex and dynamic environments remains a difficult problem that needs to be solved. In addition, existing formation recognition methods mostly rely on predefined templates or rules, with limited ability to recognize new or unconventional formations. They are unable to adapt to the changing flight pattern requirements in tactical environments, and existing systems lack the ability to update in real time and adapt to dynamic changes, which greatly affects the timeliness and accuracy of command decisions in rapidly changing battlefield environments.

[0004] In summary, existing technologies are difficult to efficiently integrate multi-source data, cannot achieve high-precision formation recognition in complex dynamic environments, and are difficult to adapt to the changing flight mode requirements in tactical environments, which urgently needs to be solved. Summary of the Invention

[0005] The present application provides an aerial target formation recognition method and device to solve the problems that the existing technology is difficult to efficiently fuse multi-source data, cannot achieve high-precision formation recognition in complex dynamic environments, and is difficult to adapt to the changing flight mode requirements in tactical environments.

[0006] A first aspect of the present application provides an aerial target formation recognition method, comprising the following steps: collecting multimodal data corresponding to a target group, wherein the multimodal data includes visible light image data, infrared data, and radar data; preprocessing the multimodal data to obtain corresponding standard data, and identifying each aerial target in the target group based on the standard data and a pre-constructed target detection algorithm to obtain dynamic recognition data corresponding to each aerial target; constructing a current hypergraph structure corresponding to the target group based on the dynamic recognition data and a preset hypergraph modeling strategy, and using the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-constructed aerial target formation knowledge base to determine the current formation type corresponding to the target group.

[0007] Optionally, in one embodiment of the present application, after determining the current formation type corresponding to the target group, it also includes: based on the current formation type and a pre-built target formation dynamic perception model, monitoring the change trend information corresponding to the target group, so as to update the aerial target formation knowledge base according to the change trend information; real-time monitoring of multiple dynamic parameters corresponding to the target group, and determining multiple key performance indicators based on the multiple dynamic parameters, and based on the multiple key performance indicators and a preset anomaly detection algorithm, judging whether there are abnormal aerial targets in the target group that meet the preset abnormal conditions, wherein when the abnormal aerial targets that meet the preset abnormal conditions exist in the target group, executing corresponding early warning and response measures; based on the multiple dynamic parameters, generating corresponding real-time analysis reports and visualization data, and sending the real-time analysis reports and the visualization data to a preset ground control center.

[0008] Optionally, in one embodiment of the present application, the multimodal data corresponding to the target group is collected, wherein the multimodal data includes visible light image data, infrared data and radar data, including: constructing a multimodal aerial target detection hardware system based on preset multi-type visual sensors and dynamic visual sensors; and synchronously collecting the visible light image data, the infrared data and the radar data through the multimodal aerial target detection hardware system.

[0009] Optionally, in one embodiment of the present application, the multimodal data is preprocessed to obtain corresponding standard data, and each aerial target in the target group is identified based on the standard data and a pre-built target detection algorithm to obtain dynamic recognition data corresponding to each aerial target, including: denoising the multimodal data, and performing data format conversion and time synchronization operations on the denoised multimodal data to obtain corresponding standard data; based on the target detection algorithm, features of the standard data are extracted, and the dynamic recognition data corresponding to each aerial target is determined according to a preset multi-task learning strategy and the features, wherein the dynamic recognition data includes the position, speed and direction corresponding to each aerial target.

[0010] Optionally, in one embodiment of the present application, the current hypergraph structure corresponding to the target group is constructed based on the dynamic recognition data and the preset hypergraph modeling strategy, so as to use the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-constructed aerial target formation knowledge base to determine the current formation type corresponding to the target group, including: taking each aerial target as a hypergraph node, and obtaining the correlation index between each aerial target, and determining the corresponding hyperedge according to the correlation index, so as to construct the corresponding current hypergraph structure according to the hypergraph node and the hyperedge; comparing the similarity between the current hypergraph structure and each formation template in a plurality of preset formation templates to obtain the similarity corresponding to each formation template; determining the target formation template with the highest similarity among the plurality of formation templates, and training the pre-constructed hypergraph neural network model based on a preset formation sample training data set, and inputting the target formation template into the trained hypergraph neural network model to output the current formation type corresponding to the target group.

[0011] The second aspect of the present application provides an aerial target formation recognition device, including: an acquisition module for acquiring multimodal data corresponding to a target group, wherein the multimodal data includes visible light image data, infrared data and radar data; a preprocessing module for preprocessing the multimodal data to obtain corresponding standard data, and based on the standard data and a pre-built target detection algorithm, identifying each aerial target in the target group to obtain dynamic recognition data corresponding to each aerial target; a formation recognition module for constructing a current hypergraph structure corresponding to the target group based on the dynamic recognition data and a preset hypergraph modeling strategy, so as to use the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group.

[0012] Optionally, in one embodiment of the present application, it also includes: a monitoring module for monitoring the change trend information corresponding to the target group based on the current formation type and a pre-built target formation dynamic perception model after determining the current formation type corresponding to the target group, so as to update the aerial target formation knowledge base according to the change trend information; a judgment module for monitoring the multiple dynamic parameters corresponding to the target group in real time, and determining a plurality of key performance indicators based on the multiple dynamic parameters, and judging whether there are abnormal aerial targets that meet the preset abnormal conditions in the target group based on the multiple key performance indicators and a preset abnormality detection algorithm, wherein when the abnormal aerial targets that meet the preset abnormal conditions exist in the target group, corresponding early warning and response measures are executed; a generation module for generating corresponding real-time analysis reports and visualization data based on the multiple dynamic parameters, and sending the real-time analysis reports and the visualization data to a preset ground control center.

[0013] Optionally, in one embodiment of the present application, the acquisition module includes: a modeling module for constructing a multimodal aerial target detection hardware system based on preset multi-type visual sensors and dynamic visual sensors; an acquisition module for synchronously acquiring the visible light image data, the infrared data and the radar data through the multimodal aerial target detection hardware system.

[0014] Optionally, in one embodiment of the present application, the preprocessing module includes: a denoising unit, used to denoise the multimodal data, and perform data format conversion and time synchronization operations on the denoised multimodal data to obtain corresponding standard data; an extraction unit, used to extract features of the standard data based on the target detection algorithm, and determine the dynamic recognition data corresponding to each aerial target based on a preset multi-task learning strategy and the features, wherein the dynamic recognition data includes the position, speed and direction corresponding to each aerial target.

[0015] Optionally, in one embodiment of the present application, the formation recognition module includes: a construction unit, used to take each aerial target as a hypergraph node, and obtain the correlation index between each aerial target, and determine the corresponding hyperedge based on the correlation index, so as to construct the corresponding current hypergraph structure based on the hypergraph node and the hyperedge; a comparison unit, used to compare the similarity of the current hypergraph structure with each formation template in a plurality of preset formation templates, so as to obtain the similarity corresponding to each formation template; a training unit, used to determine the target formation template with the highest similarity among the plurality of formation templates, and train a pre-constructed hypergraph neural network model based on a preset formation sample training data set, and input the target formation template into the trained hypergraph neural network model to output the current formation type corresponding to the target group.

[0016] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the aerial target formation recognition method as described in the above embodiment.

[0017] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned method for identifying aerial target formations.

[0018] The fifth aspect of the present application provides a computer program product, including a computer program, which is executed to implement the above-mentioned aerial target formation recognition method.

[0019] Therefore, the embodiments of the present application have the following beneficial effects:

[0020] The embodiments of the present application can collect multimodal data corresponding to a target group, wherein the multimodal data includes visible light image data, infrared data, and radar data; preprocess the multimodal data to obtain corresponding standard data, and identify each aerial target in the target group based on the standard data and a pre-built target detection algorithm to obtain dynamic identification data corresponding to each aerial target; based on the dynamic identification data and a preset hypergraph modeling strategy, construct a current hypergraph structure corresponding to the target group, and use the current hypergraph structure to compare and identify the current formation corresponding to the target group in a pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group. The present application introduces a dynamic knowledge base of aerial target formations and combines hypergraph modeling and analysis technology to achieve situational awareness of the flight of aerial target groups and high-precision, real-time aerial target formation identification, significantly improving the efficiency and accuracy of aerial monitoring and command decision-making. Thus, it solves the problems of the existing technology that it is difficult to efficiently integrate multi-source data, cannot achieve high-precision formation identification in complex dynamic environments, and is difficult to adapt to the changing flight mode requirements in tactical environments.

[0021] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0023] Figure 1 A flowchart of a method for identifying an aerial target formation according to an embodiment of the present application is provided;

[0024] Figure 2 A schematic diagram of the logical architecture of a method for identifying an aerial target formation provided in accordance with one embodiment of the present application;

[0025] Figure 3 This is an example diagram of an aerial target formation recognition device according to an embodiment of the present application;

[0026] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0027] Among them, 10 is an aerial target formation recognition device; 100 is an acquisition module, 200 is a pre-processing module, 300 is a formation recognition module; 401 is a memory, 402 is a processor, and 403 is a communication interface. DETAILED DESCRIPTION

[0028] The following describes in detail embodiments of the present application, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0029] The following describes an aerial target formation recognition method and device according to an embodiment of the present application with reference to the accompanying drawings. In response to the problems mentioned in the above background technology, the present application provides an aerial target formation recognition method, wherein multimodal data corresponding to a target group is collected, wherein the multimodal data includes visible light image data, infrared data, and radar data; the multimodal data is preprocessed to obtain corresponding standard data, and each aerial target in the target group is identified based on the standard data and a pre-built target detection algorithm to obtain dynamic recognition data corresponding to each aerial target; based on the dynamic recognition data and a preset hypergraph modeling strategy, a current hypergraph structure corresponding to the target group is constructed, and the current hypergraph structure is used to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group. By introducing a dynamic knowledge base of aerial target formations and combining hypergraph modeling and analysis technology, the present application achieves situational awareness of the flight of aerial target groups and high-precision, real-time aerial target formation recognition, significantly improving the efficiency and accuracy of air monitoring and command decision-making. This solves the problems of existing technologies such as the difficulty in efficiently integrating multi-source data, the inability to achieve high-precision formation recognition in complex dynamic environments, and the difficulty in adapting to the changing flight mode requirements in tactical environments.

[0030] Specifically, Figure 1 A flowchart of a method for identifying aerial target formations provided in an embodiment of the present application.

[0031] like Figure 1 As shown, the aerial target formation recognition method includes the following steps:

[0032] In step S101 , multimodal data corresponding to a target group is collected, wherein the multimodal data includes visible light image data, infrared data, and radar data.

[0033] The embodiment of the present application can first use multiple types of visual and dynamic visual sensors to build a multi-modal aerial target detection hardware system, such as Figure 2 As shown, multimodal data such as visible light image data, infrared data, and radar data corresponding to the target group are synchronously collected.

[0034] Optionally, in one embodiment of the present application, multimodal data corresponding to the target group is collected, wherein the multimodal data includes visible light image data, infrared data, and radar data, including: constructing a multimodal aerial target detection hardware system based on preset multi-type visual sensors and dynamic visual sensors; and synchronously collecting visible light image data, infrared data, and radar data through the multimodal aerial target detection hardware system.

[0035] It should be noted that the embodiments of the present application can first select and install multiple types of visual sensors and dynamic visual sensors suitable for aerial target detection, including high-resolution cameras, infrared sensors and radar systems.

[0036] To ensure full coverage of the target monitoring area, sensors should be rationally arranged in key locations and accurately calibrated and synchronized to improve the accuracy and consistency of data collection; in addition, the embodiments of the present application should consider the environmental adaptability and anti-interference ability of the sensors to ensure that they can operate stably under various climate and lighting conditions.

[0037] Secondly, the embodiment of the present application also needs to be configured with a sensor data acquisition module to ensure that various sensors can synchronously collect and transmit multimodal data to the central processing unit. Among them, the data acquisition module should have high-bandwidth and low-latency communication capabilities, and support compatibility with multiple data formats to adapt to the data output of different sensors. At the same time, the embodiment of the present application should realize real-time data synchronization and time stamping functions to ensure the time consistency of different sensor data, thereby providing a reliable foundation for subsequent data fusion and analysis. In addition, in order to ensure the security and stability of data transmission, the embodiment of the present application also needs to be configured with a redundant backup system and protection mechanism to prevent data loss or transmission interruption.

[0038] In step S102, the multimodal data is preprocessed to obtain corresponding standard data, and each aerial target in the target group is identified based on the standard data and a pre-built target detection algorithm to obtain dynamic identification data corresponding to each aerial target.

[0039] Furthermore, the embodiments of the present application also need to pre-process the multimodal input data and, in combination with the target detection algorithm, identify the aerial targets in the target group, thereby obtaining dynamic identification data corresponding to each aerial target.

[0040] Optionally, in one embodiment of the present application, the multimodal data is preprocessed to obtain corresponding standard data, and based on the standard data and a pre-built target detection algorithm, each aerial target in the target group is identified to obtain dynamic identification data corresponding to each aerial target, including: denoising the multimodal data, and performing data format conversion and time synchronization operations on the denoised multimodal data to obtain corresponding standard data; based on the target detection algorithm, extracting features of the standard data, and determining the dynamic identification data corresponding to each aerial target according to a preset multi-task learning strategy and features, wherein the dynamic identification data includes the position, speed and direction corresponding to each aerial target.

[0041] During actual implementation, the embodiments of the present application may first perform pre-processing operations such as denoising, data format conversion, and time synchronization on the received multimodal data to improve data quality and consistency.

[0042] Specifically, during the denoising process, the embodiments of the present application can use advanced filtering techniques such as Kalman filtering and median filtering to remove random noise and interference signals from sensor data. Data format conversion ensures that the data output by different sensors can be unified into a standardized format to facilitate subsequent processing. Time synchronization uses a precise timestamp mechanism to ensure that the data from each sensor is fully aligned in time to obtain corresponding standard data, avoiding recognition errors caused by time deviation. In addition, the embodiments of the present application also include data integrity checks and anomaly detection in the preprocessing process to ensure the reliability and stability of the input data.

[0043] Afterwards, the embodiments of the present application can apply target detection algorithms (such as convolutional neural networks, support vector machines, etc.) to analyze the preprocessed data (i.e., standard data) to identify basic attributes such as the position, speed, and direction of the aerial target (i.e., dynamic recognition data); in the application process of the target detection algorithm, the embodiments of the present application first use convolutional neural networks to extract features, and through multi-layer convolution and pooling operations, extract high-dimensional feature representations of the aerial target; secondly, the embodiments of the present application can use support vector machines or other classification algorithms to classify the extracted features and accurately distinguish different types of aerial targets; in addition, the embodiments of the present application can adopt multi-task learning methods to simultaneously estimate the position, speed, and direction of the target to ensure the coordination and consistency of various attributes.

[0044] It is understandable that in order to improve the performance and adaptability of the algorithm, the embodiments of the present application can also combine transfer learning and data enhancement technology, utilize pre-trained models and expanded training data, and further improve the accuracy and robustness of target recognition.

[0045] Therefore, the embodiments of the present application achieve rapid processing and real-time recognition of large amounts of aerial target data by adopting high-performance computing units and optimized algorithm design, meeting the needs of immediate situational awareness in highly dynamic environments.

[0046] In step S103, based on the dynamic recognition data and the preset hypergraph modeling strategy, a current hypergraph structure corresponding to the target group is constructed, so as to use the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in the pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group.

[0047] Afterwards, the embodiment of the present application can construct an aerial target formation knowledge base and model the dynamic relationship between aerial targets through a hypergraph to compare and identify possible formations of target groups.

[0048] Therefore, the embodiments of the present application can comprehensively capture the multi-dimensional characteristics of aerial targets through the fusion of multimodal data and hypergraph modeling technology, significantly improving the accuracy and reliability of formation recognition.

[0049] Optionally, in one embodiment of the present application, based on dynamic recognition data and a preset hypergraph modeling strategy, a current hypergraph structure corresponding to the target group is constructed, so as to use the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-constructed aerial target formation knowledge base to determine the current formation type corresponding to the target group, including: taking each aerial target as a hypergraph node, and obtaining the correlation index between each aerial target, and determining the corresponding hyperedge based on the correlation index, so as to construct the corresponding current hypergraph structure based on the hypergraph nodes and hyperedges; comparing the similarity between the current hypergraph structure and each formation template in a plurality of preset formation templates to obtain the similarity corresponding to each formation template; determining the target formation template with the highest similarity among the plurality of formation templates, and training the pre-constructed hypergraph neural network model based on the preset formation sample training data set, and inputting the target formation template into the trained hypergraph neural network model to output the current formation type corresponding to the target group.

[0050] It should be noted that the specific steps of obtaining the current formation type corresponding to the target group in the embodiment of the present application are as follows:

[0051] Step 1: Collect and store dynamic data such as the position coordinates, flight speed, heading angle, and altitude information of the identified aerial targets:

[0052] After the target detection is completed, the embodiment of the present application needs to systematically collect the dynamic data of each aerial target, which includes the target's two-dimensional or three-dimensional position coordinates (x, y, z), flight speed v = (v x ,v y ,vz ), heading angle θ, and altitude information h. To ensure the real-time and accuracy of data, embodiments of the present application require the establishment of an efficient data storage mechanism, such as using a time-series database to manage dynamic data. In addition, embodiments of the present application may employ data compression and indexing technologies to optimize storage space and query efficiency, and the data's timestamps and unique identifiers must be accurately recorded to facilitate precise association and analysis in subsequent steps.

[0053] Step 2: Use hypergraph modeling technology to model the spatial and dynamic association relationships between aerial targets, forming a high-order association model that reflects the target group structure:

[0054] Those skilled in the art should understand that a hypergraph is a generalized graph structure that can represent high-order relationships between multiple nodes. In the embodiment of the present application, each aerial target can be considered as a node in the hypergraph. Secondly, the embodiment of the present application can construct hyperedges to represent the relationship between multiple targets based on indicators such as spatial distance, speed similarity, and heading consistency between targets, and can use the following formula to define the weight w of the hyperedge: e :

[0055]

[0056] Among them, d ij represents the distance between target i and target j; δ(v i ,v j ) represents the similarity measure of velocity vectors; cos(θ i -θ j ) represents the consistency of heading angle; α, β, and γ are weight coefficients used to balance the importance of various indicators.

[0057] Therefore, in the embodiments of the present application, the hypergraph can effectively capture the complex dynamic association relationships within the target group, form a high-order association model, and reflect the overall structure and behavior pattern of the target group.

[0058] Step 3: Based on the constructed hypergraph model, compare the possible formations of the target group and identify the specific formation type of the current target group through pattern matching or machine learning algorithms:

[0059] After constructing the hypergraph model, the embodiment of the present application can compare the hypergraph of the current target group with the predefined formation template, and can use pattern matching algorithms or machine learning methods, such as hypergraph convolutional network (HGNN) for identification. The specific implementation steps include: formation pattern matching and formation pattern recognition.

[0060] For formation pattern matching, the embodiment of the present application can define multiple pre-set formation templates, each template corresponds to a specific hypergraph structure, and calculate the similarity S(G, G) between the current hypergraph and each template hypergraph. t ), and select the template with the highest similarity as the recognition result of the current formation.

[0061] As an achievable approach, similarity calculation in the embodiments of the present application may employ graph isomorphism detection, subgraph matching, or similarity indicators based on node and edge features, as shown in the following formula:

[0062]

[0063] Among them, E and E t are the current hypergraph G and the template hypergraph G respectively t The hyperedge set of w e and is the weight of each hyperedge.

[0064] Secondly, for formation pattern recognition, the embodiment of the present application can use the constructed hypergraph data to train the hypergraph neural network model, so that it can automatically learn and recognize different formation types. During the training process, the embodiment of the present application can adopt a supervised learning method and use the labeled formation sample data (i.e., formation sample training data set) for training. The hypergraph neural network model outputs the formation category to which the current target group belongs. The prediction process can be expressed as:

[0065]

[0066] Among them, f HGNN represents a hypergraph neural network; G is the input hypergraph; Θ is the model parameter.

[0067] After training, the hypergraph neural network model can efficiently classify the real-time input hypergraph, thereby accurately identifying the current specific formation type of the target group.

[0068] Therefore, the embodiments of the present application can effectively deal with interference and occlusion in complex flight environments through multi-source data fusion, thereby maintaining stable recognition performance in various complex scenarios.

[0069] Optionally, in one embodiment of the present application, after determining the current formation type corresponding to the target group, it also includes: based on the current formation type and a pre-built target formation dynamic perception model, monitoring the change trend information corresponding to the target group, so as to update the aerial target formation knowledge base according to the change trend information; real-time monitoring of multiple dynamic parameters corresponding to the target group, and determining multiple key performance indicators based on the multiple dynamic parameters, and based on the multiple key performance indicators and a preset anomaly detection algorithm, judging whether there are abnormal aerial targets in the target group that meet the preset abnormal conditions, wherein when there are abnormal aerial targets that meet the preset abnormal conditions in the target group, executing corresponding early warning and response measures; based on multiple dynamic parameters, generating corresponding real-time analysis reports and visualization data, and sending the real-time analysis reports and visualization data to a preset ground control center.

[0070] It should be noted that after determining the current formation type corresponding to the target group, the embodiment of the present application further needs to establish a target formation dynamic perception model to update the aerial target formation knowledge base in real time according to the formation changes, and to monitor the position and speed of the target group in real time for real-time analysis of ground target situation awareness. The specific steps are shown in the following formula:

[0071] Step 1: Establish a dynamic perception model of the target formation and continuously monitor the changing trends of the formation (such as the expansion, contraction, rotation and other dynamic behaviors of the formation):

[0072] It is understandable that in order to accurately perceive the dynamic changes in the target formation, the embodiments of the present application need to construct a comprehensive dynamic perception model. First, the embodiments of the present application can timely capture the dynamic behavior of the formation by monitoring the formation changes of the target group in real time, such as expansion, contraction, and rotation. In actual implementation, the embodiments of the present application can use advanced filtering technology and prediction algorithms to track the evolution of the formation to ensure timely response to each changing trend.

[0073] Secondly, embodiments of the present application can utilize machine learning algorithms such as hidden Markov models or long short-term memory networks to identify specific dynamic behavior patterns of target groups. For example, when a formation expansion is detected, the model should be able to automatically identify and classify it as a specific tactical action, thereby providing more accurate situational awareness.

[0074] In addition, the embodiments of the present application can comprehensively analyze formation changes by combining multi-dimensional data such as position, speed, and acceleration, thereby improving the understanding and prediction capabilities of dynamic changes in formations.

[0075] Finally, the embodiment of the present application can establish a real-time feedback mechanism to ensure that the detected formation changes can be transmitted to other functional modules in a timely manner to maintain the coordination and efficient operation of the entire perception system, thereby effectively capturing and analyzing the real-time changes of the target formation and improving the adaptability and response speed to complex dynamic environments.

[0076] Step 2: Based on the formation changes monitored in real time, the aerial target formation knowledge base is dynamically updated to ensure that the formation model in the knowledge base is consistent with the actual situation:

[0077] It is understandable that in order to ensure that the knowledge base always reflects the current actual formation situation, the embodiment of the present application needs to dynamically update it. First, the embodiment of the present application can synchronize the formation change data monitored by the dynamic perception model with the existing knowledge base, integrate the new and old data through data fusion technology, and maintain the real-time and accuracy of the knowledge base; secondly, the embodiment of the present application can use the pattern recognition algorithm to match the new formation pattern monitored with the formation model in the knowledge base. If a new formation type is detected, the embodiment of the present application can automatically generate a new model and incorporate it into the knowledge base, so that the content of the knowledge base can be continuously expanded and improved; in addition, the formation model in the knowledge base is version controlled, and the content and reason of each update are recorded to ensure that it can be backtracked and reviewed when necessary, thereby maintaining the reliability and traceability of the knowledge base. Finally, the embodiment of the present application can also introduce automated management tools to regularly scan and update the knowledge base, reduce manual intervention, and improve update efficiency and accuracy. This automated management not only increases the speed of updates, but also ensures the continuous optimization and consistency of the knowledge base content.

[0078] Therefore, the knowledge base in the embodiment of the present application can timely reflect the latest dynamics of the target formation, so that data identification and analysis can always be based on the most accurate and latest data.

[0079] Step 3: Monitor the target group's location, speed, and other key information in real time, and feed the monitoring results back to the ground control center to support ground target situation awareness and real-time analysis, and assist in command decision-making:

[0080] In order to achieve comprehensive situational awareness and efficient command decision support, real-time monitoring and feedback of key information of the target group is required. First, the embodiment of the present application can continuously collect and analyze the location information, speed and other dynamic parameters of the target group to timely reflect the current situation; secondly, the embodiment of the present application can set key performance indicators such as the target's speed change rate, acceleration, and heading angle change, and monitor these key performance indicators in real time to see if they exceed preset thresholds, thereby helping to promptly detect anomalies or potential threats and ensure a clear understanding of the dynamic changes of the target group.

[0081] In addition, the embodiments of the present application can use anomaly detection algorithms to identify abnormal behaviors or emergency situations that may exist in a group. Once an anomaly is detected, the embodiments of the present application can immediately issue an alarm and take corresponding countermeasures to ensure that emergencies can be responded to and handled quickly. At the same time, the embodiments of the present application can display real-time monitoring data to the ground control center through visualization tools, including electronic maps, dynamic trajectory maps, etc., to help commanders intuitively understand the current situation. At the same time, the embodiments of the present application can also generate real-time reports and analysis results to assist command decisions and make the decision-making process more scientific and efficient.

[0082] Finally, the embodiments of the present application can also establish an efficient two-way communication mechanism to ensure that the ground control center can receive monitoring data in a timely manner and issue instructions or request detailed information to the air system as needed. This efficient communication mechanism effectively guarantees the timely transmission of information and the rapid execution of instructions, improves the response speed and accuracy of overall operations or monitoring, thereby enabling comprehensive and real-time monitoring of target groups, and effectively supports the situational awareness and decision-making of the ground control center, thereby improving the response speed and accuracy of overall operations or monitoring.

[0083] Therefore, the embodiments of the present application can update the knowledge base in real time through the target formation dynamic perception model, thereby adapting to the changes of new or unconventional formations, and improving the flexibility and adaptability of aerial target formation recognition; in addition, the embodiments of the present application can also provide reliable decision-making support for the ground command center by outputting high-precision formation recognition results in real time, thereby improving the efficiency and accuracy of overall air monitoring and command decision-making.

[0084] According to the aerial target formation recognition method proposed in the embodiment of the present application, multimodal data corresponding to the target group is collected, wherein the multimodal data includes visible light image data, infrared data and radar data; the multimodal data is preprocessed to obtain corresponding standard data, and based on the standard data and a pre-built target detection algorithm, each aerial target in the target group is identified to obtain dynamic recognition data corresponding to each aerial target; based on the dynamic recognition data and a preset hypergraph modeling strategy, a current hypergraph structure corresponding to the target group is constructed, so as to use the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group. This application realizes situational awareness of the flight of aerial target groups and high-precision, real-time aerial target formation recognition by introducing an aerial target formation dynamic knowledge base and combining hypergraph modeling and analysis technology, significantly improving the efficiency and accuracy of aerial monitoring and command decision-making.

[0085] Next, an aerial target formation recognition device proposed according to an embodiment of the present application will be described with reference to the accompanying drawings.

[0086] Figure 3 It is a block diagram of an aerial target formation recognition device according to an embodiment of the present application.

[0087] like Figure 3 As shown, the aerial target formation recognition device 10 includes: a collection module 100 , a pre-processing module 200 and a formation recognition module 300 .

[0088] The acquisition module 100 is used to acquire multimodal data corresponding to the target group, wherein the multimodal data includes visible light image data, infrared data and radar data.

[0089] The preprocessing module 200 is used to preprocess the multimodal data to obtain corresponding standard data, and identify each aerial target in the target group based on the standard data and a pre-built target detection algorithm to obtain dynamic identification data corresponding to each aerial target.

[0090] The formation recognition module 300 is used to construct a current hypergraph structure corresponding to the target group based on dynamic recognition data and a preset hypergraph modeling strategy, so as to use the current hypergraph structure to compare and recognize the current formation corresponding to the target group in a pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group.

[0091] Optionally, in one embodiment of the present application, the aerial target formation recognition device 10 of the embodiment of the present application further includes: a monitoring module, a judgment module and a generation module.

[0092] Among them, the monitoring module is used to monitor the change trend information corresponding to the target group based on the current formation type and a pre-built target formation dynamic perception model after determining the current formation type corresponding to the target group, so as to update the aerial target formation knowledge base according to the change trend information.

[0093] The judgment module is used to monitor multiple dynamic parameters corresponding to the target group in real time, determine multiple key performance indicators based on the multiple dynamic parameters, and judge whether there are abnormal aerial targets that meet the preset abnormal conditions in the target group based on the multiple key performance indicators and the preset abnormality detection algorithm. When there are abnormal aerial targets that meet the preset abnormal conditions in the target group, corresponding early warning and response measures are executed.

[0094] The generation module is used to generate corresponding real-time analysis reports and visualization data based on multiple dynamic parameters, and send the real-time analysis reports and visualization data to a preset ground control center.

[0095] Optionally, in one embodiment of the present application, the acquisition module 100 includes: a modeling module and an acquisition module.

[0096] Among them, the modeling module is used to build a multimodal aerial target detection hardware system based on preset multi-type visual sensors and dynamic visual sensors.

[0097] The acquisition module is used to synchronously collect visible light image data, infrared data and radar data through a multi-modal aerial target detection hardware system.

[0098] Optionally, in one embodiment of the present application, the pre-processing module 200 includes: a denoising unit and an extraction unit.

[0099] The denoising unit is used to perform denoising on the multimodal data and perform data format conversion and time synchronization operations on the denoised multimodal data to obtain corresponding standard data.

[0100] The extraction unit is used to extract features of the standard data based on the target detection algorithm, and determine the dynamic recognition data corresponding to each aerial target according to the preset multi-task learning strategy and features, wherein the dynamic recognition data includes the position, speed and direction corresponding to each aerial target.

[0101] Optionally, in one embodiment of the present application, the formation recognition module 300 includes: a construction unit, a comparison unit, and a training unit.

[0102] Among them, the construction unit is used to take each aerial target as a hypergraph node, obtain the correlation index between each aerial target, and determine the corresponding hyperedge according to the correlation index, so as to construct the corresponding current hypergraph structure according to the hypergraph nodes and hyperedges.

[0103] The comparison unit is used to compare the similarity between the current hypergraph structure and each of the plurality of preset formation templates to obtain the similarity corresponding to each formation template.

[0104] The training unit is used to determine the target formation template with the highest similarity among multiple formation templates, and train a pre-built hypergraph neural network model based on a preset formation sample training data set, and input the target formation template into the trained hypergraph neural network model to output the current formation type corresponding to the target group.

[0105] It should be noted that the above explanations of the embodiment of the aerial target formation recognition method are also applicable to the aerial target formation recognition device of this embodiment, and will not be repeated here.

[0106] According to the embodiment of the present application, the aerial target formation recognition device proposed includes an acquisition module 100 for collecting multimodal data corresponding to the target group, wherein the multimodal data includes visible light image data, infrared data and radar data; a preprocessing module 200 for preprocessing the multimodal data to obtain corresponding standard data, and based on the standard data and a pre-built target detection algorithm, identifying each aerial target in the target group to obtain dynamic recognition data corresponding to each aerial target; a formation recognition module 300 for constructing a current hypergraph structure corresponding to the target group based on the dynamic recognition data and a preset hypergraph modeling strategy, so as to use the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group. This application introduces a dynamic knowledge base of aerial target formations and combines hypergraph modeling and analysis technology to achieve situational awareness of the flight of aerial target groups and high-precision, real-time aerial target formation recognition, significantly improving the efficiency and accuracy of air monitoring and command decision-making.

[0107] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0108] Memory 401 , processor 402 , and computer programs stored in the memory 401 and executable on the processor 402 .

[0109] When the processor 402 executes the program, the aerial target formation recognition method provided in the above embodiment is implemented.

[0110] Furthermore, the electronic device further includes:

[0111] The communication interface 403 is used for communication between the memory 401 and the processor 402 .

[0112] The memory 401 is used to store computer programs that can be run on the processor 402 .

[0113] The memory 401 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0114] If the memory 401, the processor 402, and the communication interface 403 are implemented independently, the communication interface 403, the memory 401, and the processor 402 can be connected to each other via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0115] Optionally, in a specific implementation, if the memory 401 , the processor 402 and the communication interface 403 are integrated on a chip, the memory 401 , the processor 402 and the communication interface 403 can communicate with each other through an internal interface.

[0116] The processor 402 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0117] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for identifying aerial target formations.

[0118] An embodiment of the present application further provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned aerial target formation recognition method.

[0119] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0120] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "N" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0121] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing a custom logical function or process step, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed in a different order than shown or discussed, including performing functions in a substantially simultaneous manner or in a reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application pertain.

[0122] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or N wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program can be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing it in other suitable ways as necessary, and then storing it in a computer memory.

[0123] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0124] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0125] In addition, the functional units in the various embodiments of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into a module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0126] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A method for identifying aerial target formations, characterized in that: The following steps are involved: Collecting multimodal data corresponding to the target group, wherein the multimodal data includes visible light image data, infrared data, and radar data; Preprocessing the multimodal data to obtain corresponding standard data, and identifying each aerial target in the target group based on the standard data and a pre-built target detection algorithm to obtain dynamic identification data corresponding to each aerial target; Based on the dynamic recognition data and the preset hypergraph modeling strategy, a current hypergraph structure corresponding to the target group is constructed, so as to utilize the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-constructed aerial target formation knowledge base to determine the current formation type corresponding to the target group.

2. The method according to claim 1, characterized in that After determining the current formation type corresponding to the target group, the method further includes: Based on the current formation type and a pre-built target formation dynamic perception model, monitoring change trend information corresponding to the target group, so as to update the aerial target formation knowledge base according to the change trend information; monitoring a plurality of dynamic parameters corresponding to the target group in real time, determining a plurality of key performance indicators based on the plurality of dynamic parameters, and determining whether an abnormal aerial target meeting a preset abnormal condition exists in the target group based on the plurality of key performance indicators and a preset abnormality detection algorithm, wherein when an abnormal aerial target meeting the preset abnormal condition exists in the target group, executing corresponding early warning and response measures; Based on the multiple dynamic parameters, corresponding real-time analysis reports and visualization data are generated, and the real-time analysis reports and visualization data are sent to a preset ground control center.

3. The method according to claim 1, characterized in that The multimodal data corresponding to the target group is collected, wherein the multimodal data includes visible light image data, infrared data, and radar data, including: Build a multimodal aerial target detection hardware system based on preset multi-type visual sensors and dynamic visual sensors; The visible light image data, the infrared data and the radar data are synchronously collected by the multimodal aerial target detection hardware system.

4. The method according to claim 3, characterized in that The preprocessing of the multimodal data to obtain corresponding standard data, and identifying each aerial target in the target group based on the standard data and a pre-built target detection algorithm to obtain dynamic identification data corresponding to each aerial target, includes: Performing denoising on the multimodal data, and performing data format conversion and time synchronization operations on the denoised multimodal data to obtain corresponding standard data; Based on the target detection algorithm, the features of the standard data are extracted, and the dynamic recognition data corresponding to each aerial target is determined according to a preset multi-task learning strategy and the features, wherein the dynamic recognition data includes the position, speed and direction corresponding to each aerial target.

5. The method according to claim 4, characterized in that The constructing of a current hypergraph structure corresponding to the target group based on the dynamic recognition data and a preset hypergraph modeling strategy, and using the current hypergraph structure to perform a comparison and recognition operation on the current formation corresponding to the target group in a pre-constructed aerial target formation knowledge base to determine the current formation type corresponding to the target group, includes: Taking each of the aerial targets as a hypergraph node, obtaining a correlation index between each of the aerial targets, and determining a corresponding hyperedge according to the correlation index, so as to construct a corresponding current hypergraph structure according to the hypergraph nodes and the hyperedges; Comparing the current hypergraph structure with each of the plurality of preset formation templates for similarity to obtain a similarity corresponding to each formation template; Determine the target formation template with the highest similarity among the multiple formation templates, and train a pre-constructed hypergraph neural network model based on a preset formation sample training data set, and input the target formation template into the trained hypergraph neural network model to output the current formation type corresponding to the target group.

6. An aerial target formation recognition device, characterized in that: include: An acquisition module, configured to acquire multimodal data corresponding to a target group, wherein the multimodal data includes visible light image data, infrared data, and radar data; a preprocessing module, configured to preprocess the multimodal data to obtain corresponding standard data, and identify each aerial target in the target group based on the standard data and a pre-built target detection algorithm to obtain dynamic identification data corresponding to each aerial target; A formation recognition module is used to construct a current hypergraph structure corresponding to the target group based on the dynamic recognition data and a preset hypergraph modeling strategy, so as to use the current hypergraph structure to compare and recognize the current formation corresponding to the target group in a pre-built aerial target formation knowledge base to determine the current formation type corresponding to the target group.

7. The device according to claim 6, characterized in that Also includes: a monitoring module configured to, after determining a current formation type corresponding to the target group, monitor change trend information corresponding to the target group based on the current formation type and a pre-built target formation dynamic perception model, so as to update the aerial target formation knowledge base according to the change trend information; a judgment module, configured to monitor multiple dynamic parameters corresponding to the target group in real time, determine multiple key performance indicators based on the multiple dynamic parameters, and judge whether there is an abnormal aerial target in the target group that meets preset abnormal conditions based on the multiple key performance indicators and a preset abnormality detection algorithm, wherein when the abnormal aerial target that meets the preset abnormal conditions exists in the target group, corresponding early warning and response measures are executed; A generation module is used to generate corresponding real-time analysis reports and visualization data based on the multiple dynamic parameters, and send the real-time analysis reports and the visualization data to a preset ground control center.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the aerial target formation recognition method according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the aerial target formation recognition method according to any one of claims 1 to 5.

10. A computer program product comprising a computer program, characterized in that The computer program is executed to implement the aerial target formation recognition method according to any one of claims 1 to 5.

Images