Self-adaptive regulation and control method and system suitable for three-dimensional radar dynamic visualization

Abstract

The invention discloses a self-adaptive regulation and control method and system suitable for three-dimensional radar dynamic visualization. The self-adaptive regulation and control method comprises the steps that basic three-dimensional radar data information and operation performance data information are collected in real time; performing dynamic threat prediction on the basic three-dimensional radar data information by using a space-time forgetting mechanism, and obtaining an ADST-TPM dynamic threat index parameter corresponding to each target; generating an importance weight map covering all target sets according to the basic three-dimensional radar data information and the ADST-TPM dynamic threat index parameters corresponding to each target; performing topological structure simplification on the three-dimensional grid model by using the three-dimensional point cloud and the importance weight map corresponding to the target in combination with the operation performance data information to generate a simplified grid; and performing adaptive regulation and control on the performance index of the three-dimensional radar dynamic visualization operation according to the patch number reduction ratio corresponding to the simplified grid. The system comprises modules corresponding to the steps of the method.

Description

Technical Field

[0001] The present invention proposes an adaptive control method and system suitable for three-dimensional radar dynamic visualization, belonging to the technical field of three-dimensional radar visualization control. Background Art

[0002] In the field of three-dimensional radar dynamic visualization technology, with the continuous advancement of science and technology, the application demand for this technology is increasing. It is widely used in many fields such as military monitoring, traffic control, and weather forecasting, and plays a key role in real-time acquisition and analysis of target information in complex environments. Existing technologies for 3D radar dynamic visualization suffer from numerous deficiencies in data processing and performance control. Traditional methods struggle to achieve comprehensive, accurate, and efficient acquisition of basic 3D radar data and real-time performance data, failing to meet the requirements for data timeliness and integrity in complex and ever-changing scenarios. For example, in large-scale traffic or military surveillance scenarios, missing or delayed data can lead to errors in target positioning and tracking. On the other hand, existing technologies for processing basic 3D radar data lack effective dynamic threat prediction mechanisms. Most traditional methods fail to fully consider the temporal and spatial characteristics of the data, making it difficult to accurately predict dynamic threat indicator parameters for targets. This makes it difficult to make accurate judgments and countermeasures in advance when facing fast-moving or potentially threatening targets. For example, in meteorological monitoring, the threat level of targets such as clouds that could potentially develop into severe weather cannot be accurately predicted, affecting the accuracy and timeliness of weather warnings. In addition, when generating importance weight maps for associated targets and point cloud clusters, existing technologies are unable to effectively combine the dynamic threat indicator parameters of the three-dimensional scene and the target, resulting in the generated weight map not accurately reflecting the actual importance and relevance of the target, which in turn affects the subsequent processing of the three-dimensional mesh model. In the process of simplifying the topological structure of the three-dimensional mesh model, existing methods often fail to fully combine the three-dimensional point cloud, importance weight map, and operation performance data information corresponding to the target. The generated simplified mesh cannot effectively optimize the three-dimensional radar dynamic visualization operation performance indicators while ensuring the visualization effect. Finally, when adaptively adjusting the performance indicators of the three-dimensional radar dynamic visualization operation, due to the deficiencies in the above-mentioned various links, the existing technology finds it difficult to accurately and adaptively adjust the performance indicators according to the actual situation. As a result, during the operation of the system, problems such as poor visualization effect, slow processing speed, and excessive resource consumption may occur, seriously affecting the effectiveness of the three-dimensional radar dynamic visualization technology in practical applications. In summary, the existing 3D radar dynamic visualization technology has obvious defects and shortcomings in data acquisition, processing, model simplification, and performance control. There is an urgent need for a new adaptive control method to solve these problems in order to improve the accuracy, efficiency, and stability of 3D radar dynamic visualization technology in various applications. Summary of the Invention

[0003] The present invention provides an adaptive control method and system for three-dimensional radar dynamic visualization to solve the technical problems existing in the above-mentioned prior art. The technical solutions adopted are as follows: An adaptive control method for three-dimensional radar dynamic visualization, the adaptive control method comprising: Real-time collection of basic 3D radar data and operational performance data; Use the spatiotemporal forgetting mechanism to perform ADST-TPM dynamic threat prediction on basic 3D radar data information and obtain the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID; Generate an importance weight map covering all target sets based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, wherein the importance weight map is used to associate a specific target Target_ID with the point cloud cluster corresponding to the target Target_ID; The 3D point cloud and importance weight map corresponding to the target Target_ID are combined with the operation performance data to simplify the topology of the 3D mesh model and generate a simplified mesh. The performance index of the dynamic visualization operation of the three-dimensional radar is adaptively regulated according to the facet number reduction ratio corresponding to the simplified grid.

[0004] Furthermore, basic 3D radar data and operational performance data are collected in real time, including: During the operation of the 3D radar, basic 3D radar data information of the 3D radar is retrieved in real time; wherein the basic 3D radar data information includes raw radar data, target tracking data, geo-fence data, multi-target relationship data, user interaction enhancement data and system load data; Real-time collection of operational performance data during 3D radar dynamic visualization; wherein the operational performance data includes frame rate, GPU occupancy, ADST-TPM latency, and point cloud throughput.

[0005] Furthermore, the spatiotemporal forgetting mechanism is used to perform ADST-TPM dynamic threat prediction on the basic 3D radar data information to obtain the ADST-TPM dynamic threat indicator parameters corresponding to each target, including: Retrieving a radar scanning period Δt from raw radar data included in the basic three-dimensional radar data information, setting a cache window duration for target data according to the radar scanning period Δt, and caching the target data for the radar scanning period Δt; The radial acceleration, stability index and peak frequency shift of the target Target_ID are retrieved from the target tracking data contained in the basic 3D radar data information, and the Doppler dynamic feature vector is generated using the radial acceleration, stability index and peak frequency shift. ;in, A Indicates the radial acceleration of the target Target_ID; stability Indicates the stability index of target Target_ID; f peak Indicates the peak frequency shift of the target Target_ID; Retrieve the geofence list and target operation data corresponding to the target Target_ID, and obtain the spatiotemporal context feature vector corresponding to the target Target_ID based on the geofence list and target operation data ,in, W 01 and W 02 They represent the geographical threat correlation and target interaction threat corresponding to the target Target_ID respectively; Performing weighted fusion on the Doppler dynamics feature vector and the spatiotemporal context feature vector to generate a fused feature vector; Monitor the GPU utilization of the current system in real time and select a trained threat assessment model based on the GPU utilization; wherein the threat assessment model includes a full GRU network model, a lightweight temporal convolution model, and a lightweight temporal convolution model; The fused feature vector is input into the threat assessment model, and the initial threat probability corresponding to the target Target_ID is output. The initial threat probability is forgotten using the spatiotemporal forgetting mechanism to obtain the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID.

[0006] Furthermore, the geo-fence list and target operation data corresponding to the target Target_ID are retrieved, and the spatiotemporal context feature vector corresponding to the target Target_ID is obtained according to the geo-fence list and target operation data, including: Traverse the geofence list and get the minimum distance between the target Target_ID location and the geofence; Retrieving a priority value corresponding to a fence from the geo-fence list; The minimum distance between the target Target_ID location and the geo-fence is used in combination with the priority value corresponding to the fence in the geo-fence list to obtain the geo-threat relevance corresponding to the target Target_ID; Search for neighboring targets within 500 meters of the target Target_ID and select those with an initial threat probability greater than 0.7 as high-threat neighboring targets; The target interaction threat corresponding to the target Target_ID is obtained through the velocity vector of the target Target_ID and the velocity vector of the high-threat neighboring target.

[0007] Furthermore, the GPU utilization of the current system is monitored in real time, and a trained threat assessment model is selected based on the GPU utilization, including: When the GPU utilization is less than 0.7, the complete GRU network model is selected as the threat assessment model; When the GPU utilization is greater than or equal to 0.7 and less than 0.83, the lightweight temporal convolution model is selected as the threat assessment model; When the GPU utilization is greater than or equal to 0.83, the rule engine model is selected as the threat assessment model; The predefined rules are as follows: First check the current GPU load. When the GPU load exceeds 90%, enable the rule engine degradation mode. Define the rule engine function and create a function named rule_engine. This function receives the fused feature vector of the target Target_ID as input and outputs a floating-point value representing the threat level. The execution rules of the rule_engine function are as follows: Rule 1: If the radial acceleration of target_ID (target.a_radial) is greater than 20.0, and the distance from target_ID to the geofence (target.d_geo) is less than 5000.0, then the target is considered to be approaching the no-fly zone at high speed, indicating a high threat situation. The initial threat probability is returned to 0.9. Rule 2: Check the peak frequency shift attribute target.Δf_peak of the target Target_ID. If the peak frequency shift attribute target.Δf_peak is greater than 50, it is determined to be a threat scenario with a drastic change in the Doppler spectrum and the initial threat probability is returned to 0.7. Rule 3: Check the target interaction threat attribute target.neighbor_threat of the target. If the value is greater than 1.5, calculate the initial threat probability through the function min(0.8, 0.3 + target.neighbor_threat * 0.2) and return the calculated result.

[0008] Rule 4: Default Initial Threat Probability Return If all of the above rule conditions are not met, it means that the target threat level is low, and the default initial threat probability of 0.3 is returned.

[0009] Furthermore, an importance weight map covering all target sets is generated based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, including: Divide the 3D scene into regular voxel grids of 20m×20m×20m in size, and scan each voxel grid to capture the target Target_ID that appears in each voxel grid; Retrieve the decision weight corresponding to the target Target_ID appearing in each voxel grid and the initial weight corresponding to the voxel grid; The maximum value of the decision weight corresponding to the target Target_ID appearing in each voxel grid and the initial weight corresponding to the voxel grid is used as the calibration weight corresponding to each voxel grid; Performing anisotropic diffusion processing on the calibration weights to obtain weight values ​​after the anisotropic diffusion processing; Normalize the weight values ​​after the anisotropic diffusion process to obtain the final weight value corresponding to each voxel grid; The final weight values ​​corresponding to the voxel grid are encoded into a texture format that can be directly rendered by a GPU to generate an importance weight map for all target sets.

[0010] Furthermore, the 3D point cloud and importance weight map corresponding to the target Target_ID are combined with the operational performance data to simplify the topology of the 3D mesh model and generate a simplified mesh, including: Convert 3D point cloud coordinates into texture coordinates to generate 3D point cloud texture coordinates; Compare the 3D point cloud texture coordinates with the importance weight map to obtain the weight value corresponding to each point cloud; Compare the weight value corresponding to each point cloud with a preset first weight threshold and a second weight threshold, and determine a high-weight target, a medium-weight target, and a low-weight target based on the comparison result; wherein the first weight threshold is set to 0.6 and the second weight threshold is set to 0.3; Perform edge folding and simplification processing on high-weight targets to obtain the geometric details corresponding to the high-weight areas; For medium-weight targets, only those with weight values ​​greater than 0.45 are subjected to edge folding simplification to obtain the key features corresponding to the medium-weight region; Merge vertices for low-weight targets and obtain basic feature nodes corresponding to low-weight areas; The geometric details corresponding to the high-weight areas, the key features corresponding to the medium-weight areas, and the basic features corresponding to the low-weight areas contained in all voxel grids are integrated to generate a simplified grid.

[0011] Furthermore, adaptively regulating the performance index of the three-dimensional radar dynamic visualization operation according to the facet number reduction ratio corresponding to the simplified grid includes: Determine the number of faces of the current simplified mesh based on the simplified mesh, and retrieve the number of faces of the mesh before simplification; Obtain the face number reduction ratio corresponding to the current simplified mesh according to the face number of the current simplified mesh and the face number of the mesh before simplification; Retrieve the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID appearing on the current screen; The frame rate and display bandwidth occupancy of the current 3D radar dynamic visualization operation are adaptively controlled using the facet reduction ratio corresponding to the current simplified mesh and the ADST-TPM dynamic threat index parameters corresponding to each target Target_ID.

[0012] Furthermore, the frame rate and display bandwidth occupancy of the current 3D radar dynamic visualization are adaptively controlled using the facet reduction ratio corresponding to the current simplified mesh and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, including: Retrieve the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID and the final weight value corresponding to the voxel grid where each target Target_ID is located; The threat degree coefficient is obtained by performing weighted averaging processing using the ADST-TPM dynamic threat index parameter corresponding to each target Target_ID and the final weight value corresponding to the corresponding voxel grid; Retrieve the face reduction ratio corresponding to the current simplified mesh; The adjusted frame rate and display bandwidth occupancy are obtained using the facet number reduction ratio and the threat degree coefficient corresponding to the current simplified mesh.

[0013] An adaptive control system for three-dimensional radar dynamic visualization, the adaptive control system comprising: Data acquisition module, used to collect basic 3D radar data and operating performance data in real time; The dynamic threat indicator parameter acquisition module is used to perform ADST-TPM dynamic threat prediction on basic 3D radar data information using the spatiotemporal forgetting mechanism, and obtain the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID; An importance weight map acquisition module is used to generate an importance weight map covering all target sets based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, wherein the importance weight map is used to associate a specific target Target_ID with the point cloud cluster corresponding to the target Target_ID; The simplified mesh acquisition module is used to simplify the topology of the three-dimensional mesh model by using the three-dimensional point cloud and importance weight map corresponding to the target Target_ID in combination with the operation performance data information to generate a simplified mesh; The adaptive control module is used to adaptively control the performance index of the three-dimensional radar dynamic visualization operation according to the facet number reduction ratio corresponding to the simplified grid.

[0014] Beneficial effects of the present invention: The proposed adaptive control method and system for 3D radar dynamic visualization collects two types of data in real time, ensuring the timeliness and comprehensiveness of the information. This lays a solid foundation for accurate processing in subsequent steps and avoids processing errors caused by data lag or loss. The ADST-TPM dynamic threat prediction system, using a spatiotemporal forgetting mechanism, accurately captures the dynamic threat characteristics of a target, effectively improving the accuracy and dynamic adaptability of target threat assessments and ensuring that threat indicator parameters more closely match the target's actual state changes. Furthermore, an importance weight map generated based on the 3D scene and target threat indicators clearly associates targets with point cloud clusters, providing an accurate importance basis for topological simplification of the 3D mesh model. Mesh simplification, combining the target's 3D point cloud, importance weight map, and operational performance data, ensures the visualization accuracy of key targets while achieving reasonable simplification of non-critical areas. The resulting simplified mesh significantly reduces model complexity while maintaining visual quality. Furthermore, adaptive control of operational performance indicators based on the facet reduction ratio of the simplified mesh dynamically balances visualization quality with system operational efficiency. This not only optimizes the system's resource usage and increases the speed of data processing and visualization, but also automatically adjusts based on actual operating performance, ensuring that the 3D radar dynamic visualization system maintains a stable and efficient operating state in different scenarios, significantly improving the system's overall performance and applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Flowchart corresponding to the method of the present invention; Figure 2 This is a system block diagram corresponding to the system described in the present invention. DETAILED DESCRIPTION

[0016] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0017] The embodiment of the present invention proposes an adaptive control method suitable for dynamic visualization of three-dimensional radar, such as Figure 1 As shown, the adaptive control method includes: Real-time collection of basic 3D radar data and operational performance data; Use the spatiotemporal forgetting mechanism to perform ADST-TPM dynamic threat prediction on basic 3D radar data information and obtain the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID; Generate an importance weight map covering all target sets based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, wherein the importance weight map is used to associate a specific target Target_ID with the point cloud cluster corresponding to the target Target_ID; The 3D point cloud and importance weight map corresponding to the target Target_ID are combined with the operation performance data to simplify the topology of the 3D mesh model (3DMesh) and generate a simplified mesh. The performance index of the dynamic visualization operation of the three-dimensional radar is adaptively regulated according to the facet number reduction ratio corresponding to the simplified grid.

[0018] Among them, real-time collection of basic 3D radar data information and operational performance data information includes: During the operation of the 3D radar, basic 3D radar data information of the 3D radar is retrieved in real time; wherein the basic 3D radar data information includes raw radar data, target tracking data, geo-fence data, multi-target relationship data, user interaction enhancement data and system load data; Real-time collection of operational performance data during 3D radar dynamic visualization; wherein the operational performance data includes frame rate, GPU occupancy, ADST-TPM latency, and point cloud throughput.

[0019] The working principle of the above technical solution is as follows: It achieves adaptive control of 3D radar dynamic visualization through the coordinated operation of multiple links. First, basic 3D radar data and operational performance data are collected in real time to provide raw data support for subsequent processing. Subsequently, ADST-TPM dynamic threat prediction is performed on the basic 3D radar data using a spatiotemporal forgetting mechanism. Focusing on each target Target_ID, the corresponding ADST-TPM dynamic threat indicator parameters are calculated to quantify the target's threat level. Next, combining the 3D scene and the ADST-TPM dynamic threat indicator parameters for each target, an importance weight map covering all target sets is generated. This weight map establishes the association between specific target Target_IDs and their corresponding point cloud clusters, clarifying the importance hierarchy of different targets in the 3D scene. Furthermore, using the corresponding 3D point clouds, the importance weight map, and operational performance data, the 3D mesh model (3DMesh) is topologically simplified to produce a simplified mesh. Finally, based on the facet reduction ratio corresponding to the simplified mesh, the performance indicators of the 3D radar dynamic visualization are adaptively controlled, forming a complete closed loop from data acquisition, processing, model optimization, to performance adjustment.

[0020] The benefits of this technical solution are as follows: At the data processing level, real-time collection of two types of data ensures the timeliness and comprehensiveness of information, laying a solid foundation for accurate processing in subsequent steps and avoiding processing biases caused by delayed or missing data. ADST-TPM dynamic threat prediction, through the spatiotemporal forgetting mechanism, accurately captures the dynamic threat characteristics of the target, effectively improving the accuracy and dynamic adaptability of target threat assessments, ensuring that threat indicator parameters are more closely aligned with the target's actual state changes.

[0021] In terms of model optimization, an importance weight map generated based on the 3D scene and target threat indicators clearly associates targets with point cloud clusters, providing a precise importance basis for simplifying the topological structure of the 3D mesh model. Mesh simplification, combining the target 3D point cloud, importance weight map, and operational performance data, ensures accurate visualization of key targets while enabling reasonable simplification of non-critical areas. The resulting simplified mesh significantly reduces model complexity while maintaining visual quality.

[0022] In terms of performance control, adaptively adjusting performance indicators based on the number of facets reduced in the simplified mesh dynamically balances visualization effects with system efficiency. This not only optimizes system resource usage and increases data processing and visualization speed, but also automatically adjusts based on actual performance, ensuring the 3D radar dynamic visualization system maintains stable and efficient operation in various scenarios, significantly improving the system's overall performance and applicability.

[0023] In one embodiment of the present invention, a spatiotemporal forgetting mechanism is used to perform ADST-TPM dynamic threat prediction on basic 3D radar data information to obtain ADST-TPM dynamic threat indicator parameters corresponding to each target, including: The radar scan period Δt is retrieved from the raw radar data contained in the basic 3D radar data information, and the target data cache window duration is set based on the radar scan period Δt. The target data is cached for the radar scan period Δt. The cache window duration ranges from 10Δt to 20Δt. The target data refers to the motion and signal-related data of a single identified target (identified by Target_ID) within a certain time window in the radar monitoring scene, and specifically includes three categories: Position sequence (P[t]): The coordinates of the target in three-dimensional space (x, y, z) at different times t, reflecting the motion trajectory.

[0024] Radial velocity sequence (v_radial[t]): Calculated based on the radar Doppler shift, reflecting the change in the target's velocity towards / away from the radar.

[0025] Doppler spectrum matrix (S[t,f]): The power values ​​corresponding to different times t and different frequencies f are used to analyze the target motion characteristics (such as whether there is multi-component motion, whether there is interference, etc.).

[0026] The radial acceleration, stability index and peak frequency shift of the target Target_ID are retrieved from the target tracking data contained in the basic 3D radar data information, and the Doppler dynamic feature vector is generated using the radial acceleration, stability index and peak frequency shift. ;in, A Indicates the radial acceleration of the target Target_ID; stability Indicates the stability index of target Target_ID; f peak Indicates the peak frequency shift of the target Target_ID; The stability index ranges from 0 to 1 and is obtained by the following formula: ; in, stability ( t ) represents the stability index of target Target_ID at time t; P norm ( f , t ) indicates the time t ,frequency f The normalized power spectrum at the moment is calculated by the power spectrum value at that moment. S ( f ,t ) is obtained by dividing the power spectrum of all frequency bins; n represents the total number of frequency bins, that is, the total number of discretized intervals of the spectrum; ε represents a very small number, which is used to avoid P norm ( f , t ) = 0, the logarithmic operation becomes meaningless; Retrieve the geofence list and target operation data corresponding to the target Target_ID, and obtain the spatiotemporal context feature vector corresponding to the target Target_ID based on the geofence list and target operation data ,in, W 01 and W 02 They represent the geographical threat correlation and target interaction threat corresponding to the target Target_ID respectively; Performing weighted fusion on the Doppler dynamics feature vector and the spatiotemporal context feature vector to generate a fused feature vector; Monitor the GPU utilization of the current system in real time and select a trained threat assessment model based on the GPU utilization; wherein the threat assessment model includes a full GRU network model, a lightweight temporal convolution model, and a lightweight temporal convolution model; The fused feature vector is input into the threat assessment model, and the initial threat probability corresponding to the target Target_ID is output. The initial threat probability is forgotten using the spatiotemporal forgetting mechanism to obtain the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID.

[0027] The working principle of the above technical solution is as follows: It is based on obtaining the ADST-TPM dynamic threat indicator parameters corresponding to the target, and realizes dynamic threat prediction through multi-step data processing and analysis. First, the radar scanning period Δt is retrieved from the raw radar data of the basic 3D radar information. This is used to set the buffer window duration between 10Δt and 20Δt. The target data, including the position sequence, radial velocity sequence, and Doppler spectrum matrix, are buffered according to Δt to reserve basic data for subsequent processing. Next, the target's radial acceleration, stability index, and peak frequency shift are extracted from the target tracking data. The stability index is calculated using a formula based on the normalized power spectrum to generate a Doppler dynamics eigenvector. Simultaneously, the target's geofence list and operational data are combined to obtain a spatiotemporal contextual feature vector containing geographic threat correlation and target interaction threat. These two types of feature vectors are weightedly fused to obtain a fused feature vector. Afterwards, the system GPU utilization is monitored in real time. Based on the utilization, an adapted trained threat assessment model (a complete GRU network model or a lightweight temporal convolutional model) is selected. The fused feature vector is input into the selected model to obtain the initial threat probability. Finally, the spatiotemporal forgetting mechanism is used to forget the initial threat probability, and the ADST-TPM dynamic threat indicator parameters corresponding to the target are finally obtained.

[0028] The above technical solution achieves the following benefits: By setting a reasonable buffer window duration and caching target data according to the radar scan cycle, the integrity and timeliness of target motion and signal-related data are ensured, providing a high-quality data foundation for subsequent feature extraction and threat prediction. The quantitative calculation of the stability index accurately reflects the stability of the target signal, enhancing the accuracy of feature description. The weighted fusion of Doppler dynamic features and spatiotemporal context features comprehensively considers the target's inherent motion characteristics and external environmental factors, enabling the fused feature vector to more comprehensively characterize the target's threat-related attributes, improving the comprehensiveness and reliability of threat assessment. The model selection mechanism dynamically selects threat assessment models based on GPU utilization, adapting computing resources to model complexity. This ensures assessment accuracy while effectively optimizing system efficiency and avoiding resource waste or performance issues. Furthermore, the spatiotemporal forgetting mechanism processes the initial threat probability, mitigating the impact of outdated information and highlighting current key information. This ensures that the resulting ADST-TPM dynamic threat indicator parameters are more accurately aligned with the target's real-time state, improving the timeliness and accuracy of dynamic threat prediction and providing a reliable threat assessment basis for subsequent adaptive control of 3D radar dynamic visualization.

[0029] In one embodiment of the present invention, a geofence list and target operation data corresponding to a target Target_ID are retrieved, and a spatiotemporal context feature vector corresponding to the target Target_ID is obtained based on the geofence list and target operation data, including: Traverse the geofence list and get the minimum distance between the target Target_ID location and the geofence; Retrieving a priority value corresponding to a fence from the geo-fence list, wherein the priority value is a manually set data value; The minimum distance between the target Target_ID and the geofence is combined with the priority value of the fence in the geofence list to obtain the geo-threat correlation corresponding to the target Target_ID; wherein the geo-threat correlation reflects the threat level generated by the spatial association between the target Target_ID and the geofence; Furthermore, the geographical threat correlation is obtained by the following formula: ; in,W 01 Indicates the geographical threat correlation corresponding to the target Target_ID; i Indicates the sequence number corresponding to the fence; K Represents a collection of geofences; B i Indicates the i The priority value corresponding to each fence is a value used to indicate the importance or processing order of the fence in a specific system or scenario, and is set using methods known in the art, including but not limited to setting based on fence function, setting based on risk level, and setting based on historical data or user behavior; d i Indicates the target Target_ID to i Minimum distance between fences; L Indicates the geographical threat attenuation length, ranging from 800m to 1200m, preferably 1000m; specifically, the fence priority value B is introduced into the formula i , the threat contribution of fences can be measured differently according to actual scenarios (such as functions, risk levels, etc.) to avoid indiscriminate treatment; the minimum distance d i Combined with the exponential decay exp(−d i / L), accurately simulating the spatial pattern in which threats weaken as distance increases, allowing geographic threat correlation calculations to align with the real-world threat attenuation logic and improve the accuracy of spatial correlation threat quantification.

[0030] Search for neighboring targets within 500 meters of the target Target_ID and select those with an initial threat probability greater than 0.7 as high-threat neighboring targets; The target interaction threat corresponding to the target Target_ID is obtained by comparing the velocity vector of the target Target_ID with the velocity vector of the high-threat neighboring targets. The target interaction threat reflects the threat level Target_ID generated by the interaction between the target and the surrounding high-threat neighboring targets. Furthermore, the target interaction threat is obtained by the following formula: ; in, W 02 Indicates the target interaction threat corresponding to the target Target_ID; v t Represents the velocity vector of the target Target_ID; n Indicates the total number of high-threat neighborhood targets; v j represents the velocity vector of the jth high-threat neighborhood target; d jIndicates the distance between the jth high-threat neighborhood target and the target Target_ID; L n Indicates the preset interactive threat attenuation length, ranging from 120m to 240m, with 200m being the preferred value. Represents the vector norm L2 operation. Specifically, it screens high-threat neighborhood targets with an initial threat probability greater than 0.7, focuses on high-risk interactive objects, and reduces low-threat interference; the velocity vector difference (v t −v j ) norm operation combined with the distance exponential decay exp(−d j / L n ), which not only considers the relative motion trend (velocity vector difference), but also reflects the impact of distance on interactive threats, accurately depicts the threat level generated by the interaction between the target and high-threat neighboring targets, and improves the accuracy of interactive threat quantification.

[0031] The working principle of the above technical solution is as follows: In this embodiment, the above technical solution obtains the spatiotemporal context feature vector corresponding to the target Target_ID, specifically by calculating the geographic threat correlation and the target interaction threat. In terms of calculating the geographic threat correlation, we first traverse the geographic fence list to obtain the minimum distance from the target to each geographic fence; at the same time, we retrieve the manually set priority value of each fence, and then use a specific formula to calculate the minimum distance, priority value and preset geographic threat attenuation length (800m-1200m, preferably 1000m) to obtain the geographic threat correlation W01 that reflects the degree of spatial correlation threat between the target and the geographic fence. To calculate target interaction threat, we first search for neighboring targets within 500 meters of the target and select high-threat neighboring targets with an initial threat probability greater than 0.7. Then, based on the target's velocity vector, the velocity vectors of each high-threat neighboring target, and the distance between the target and each high-threat neighboring target, combined with a preset interaction threat attenuation length (120-240 meters, preferably 200 meters) and vector norm calculation, we calculate the target interaction threat (W02), which reflects the degree of interaction threat between the target and its surrounding high-threat neighboring targets, using a corresponding formula. The geographic threat correlation (W01) and the target interaction threat (W02) together constitute the target's spatiotemporal context feature vector.

[0032] The effect of the above technical solution is: in the calculation of geographic threat correlation, by introducing the priority value of geographic fences, the differentiated impact of different geographic regions on threat assessment is highlighted. Combined with the setting of minimum distance and geographic threat attenuation length, the quantification of geographic threats is more in line with the actual spatial relationship, which can accurately reflect the threat level caused by the spatial correlation between the target and the geographic fence, and improve the accuracy of the threat characteristics of the geographic dimension.

[0033] In the target interaction threat calculation, a range of 500 meters is limited and high-threat neighboring targets with an initial threat probability greater than 0.7 are screened, focusing on objects that may actually pose an interaction threat and avoiding interference from irrelevant targets. At the same time, the velocity vector, distance, and interaction threat attenuation length are incorporated into the calculation, comprehensively considering the motion relationship between targets and the impact of spatial distance on interaction threat. This makes the quantification of target interaction threat more reasonable and effectively captures the dynamic interaction risk between the target and surrounding high-threat objects.

[0034] At the same time, existing technologies often make binary judgments on geo-fence threats, i.e., “threat or not”. This solution dynamically assigns priorities (B i ) + distance attenuation modeling, converting semantic information such as the "functional attributes, risk level" of the geographic fence into mathematically calculable threat weights, thus achieving "semantic quantification" of geographic space threats. For example, the "security area" and "runway area" of the airport fence can differentiate threat logic by priority, breaking through the traditional crude model of "distance determines everything" and deeply binding geographic threats with business semantics. The "vector dynamics" of interactive threats captures traditional target interactive threats that only calculate a single dimension of "distance proximity" or "speed". This embodiment uses the speed vector difference (v t −v j ) norm calculation accurately captures "relative motion trends"—the difference in threat between two vehicles approaching head-on and moving away from each other in the same direction—through the modulus of the vector difference. This dynamic vector interaction modeling upgrades threat assessment from "static distance correlation" to "dynamic motion trend prediction," effectively improving threat assessment accuracy and responsiveness.

[0035] One embodiment of the present invention monitors the GPU utilization of the current system in real time and selects a trained threat assessment model based on the GPU utilization, including: When the GPU utilization rate is less than 0.7, the full GRU network model is selected as the threat assessment model. This model contains three layers of hidden units and is suitable for reasoning when GPU resources are relatively abundant, taking advantage of its strong feature learning ability. When the GPU utilization is greater than or equal to 0.7 and less than 0.83, the lightweight temporal convolution model is selected as the threat assessment model. The lightweight temporal convolution model uses dilated convolution (with a dilated convolution kernel of 3). While ensuring a certain inference effect, it has relatively low GPU resource requirements and is suitable for scenarios with medium GPU load. When the GPU utilization rate is greater than or equal to 0.83, the rule engine model is selected as the threat assessment model. The rule engine model relies on predefined rules and consumes almost no GPU computing power. It is used in scenarios where GPU resources are extremely tight to ensure the basic reasoning process. The predefined rules are as follows: First, check the current GPU load. When the GPU load exceeds 90%, enable the rule engine degradation mode. This allows the predefined rules to quickly calculate key features when GPU resources are extremely tight, ensuring the operation of basic functions. Define the rule engine function and create a function named rule_engine. This function receives the fused feature vector of the target Target_ID as input and outputs a floating-point value representing the threat level. The execution rules of the rule_engine function are as follows: Rule 1: If the radial acceleration of target_ID (target.a_radial) is greater than 20.0 (indicating that the target is accelerating in the radial direction), and the distance from target_ID to the geofence (target.d_geo) is less than 5000.0 (indicating proximity to a geographically sensitive area such as a no-fly zone), the target is considered a high-threat situation approaching the no-fly zone at high speed, and the initial threat probability is returned as 0.9. Rule 2: Check the peak frequency shift attribute target.Δf_peak of the target Target_ID. If the peak frequency shift attribute target.Δf_peak is greater than 50 (indicating a significant change in the peak frequency of the Doppler spectrum, possibly due to target maneuvering or interference), the scenario is determined to be a threat scenario with a drastic Doppler spectrum change and the initial threat probability is returned to 0.7. Rule 3: Check the target's target interaction threat attribute target.neighbor_threat. If the value is greater than 1.5 (indicating the presence of many high-threat neighboring targets), calculate the initial threat probability using the function min(0.8, 0.3 + target.neighbor_threat * 0.2) and return the calculated result.

[0036] Rule 4: Default Initial Threat Probability Return If all of the above rule conditions are not met, it means that the target threat level is low, and the default initial threat probability of 0.3 is returned.

[0037] To summarize, when the GPU load is too high, the rule_engine function is used to determine target threats based on the three rules of "high-speed approach to a no-fly zone, drastic Doppler spectrum changes, and high-threat target clusters." If the corresponding rule is met, the corresponding initial threat probability is returned. If none of the rules are met, the default low initial threat probability is returned.

[0038] The above technical solution works as follows: Its core is to dynamically select an appropriate trained threat assessment model based on the system's real-time GPU utilization, balancing threat assessment performance with GPU resource consumption. The system's GPU utilization is monitored in real time, and model selection is based on different thresholds. When GPU utilization is less than 0.7, the full GRU network model with three hidden units is selected. Its strong feature learning capabilities allow for inference when GPU resources are sufficient. When GPU utilization is greater than or equal to 0.7 and less than 0.83, the lightweight temporal convolutional model is selected. This model uses dilated convolution (with a kernel of 3) to reduce GPU resource requirements while maintaining a certain level of inference performance, making it suitable for moderate load scenarios. When GPU utilization is greater than or equal to 0.83, the rule engine model is activated. This model relies on predefined rules and consumes almost no GPU computing power. It is used in extremely resource-constrained scenarios to ensure basic inference. When GPU load exceeds 90%, a degraded mode is activated, and the fused feature vectors are processed by the rule_engine function. This function judges in sequence according to preset rules. Rule 1 returns 0.9 for high-threat situations approaching a no-fly zone at high speed; Rule 2 returns 0.7 for scenarios with drastic changes in the Doppler spectrum; Rule 3 calculates and returns the corresponding value based on the target interaction threat; if none of the above are met, Rule 4 returns the default value of 0.3, thereby quickly completing the threat assessment.

[0039] The above technical solution achieves the following results: In terms of resource adaptation, real-time monitoring of GPU utilization and dynamic model switching ensures a precise match between threat assessment tasks and GPU resources. When resources are abundant, the strong feature learning capabilities of the complete GRU network model are leveraged to improve assessment accuracy. When resources are moderate, the lightweight temporal convolutional model maintains good inference performance while controlling resource consumption. When resources are extremely limited, the rule engine model maintains basic functionality with extremely low resource consumption, avoiding system crashes or assessment interruptions caused by insufficient resources. In terms of assessment effectiveness, the predefined rules of the rule engine model enable rapid judgment of key threat scenarios (such as high-speed approaches to sensitive areas, drastic spectrum fluctuations, and high-threat clusters), ensuring accurate identification of high-threat targets even under resource-constrained conditions. The sequential execution logic of the rules ensures an efficient and orderly assessment process. Furthermore, the layered adaptation of different models fully leverages the advantages of high-performance models while ensuring system robustness through the rule engine model's degradation mechanism. This improves the overall flexibility, efficiency, and reliability of threat assessment, providing stable and resource-adaptive assessment results for subsequent dynamic threat indicator parameter acquisition and 3D radar visualization control.

[0040] In one embodiment of the present invention, the initial threat probability is forgotten using a spatiotemporal forgetting mechanism to obtain the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID, including: Retrieve the initial threat probability and the timestamp corresponding to the target Target_ID determined so far; Generate a threat probability queue using the initial threat probability corresponding to the target Target_ID and the timestamp corresponding to the initial threat probability; Traverse each initial threat probability and its timestamp contained in the threat probability queue, and calculate the time difference between the current moment and the timestamp of the initial threat probability based on the timestamp corresponding to the current moment and the initial threat probability; Setting a time attenuation factor corresponding to each initial threat probability using the time difference between the current moment and the initial threat probability timestamp; The time attenuation factor corresponding to each initial threat probability is used to perform forgetting processing on all the initial threat probabilities currently obtained, and the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID at the current moment is obtained.

[0041] The ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID at the current moment is obtained by the following formula: ; in, J Indicates the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID at the current moment; M Indicates the total number of times the initial threat probability is determined; P s Indicates the s The corresponding probability value of the initial threat probability determined this time; R ( t , t c ) s Indicates the s The timestamp corresponding to the initial threat probability determined t c and the current moment t The corresponding time decay factor, and , λ represents the time decay rate; r s Indicates the s The spatial accelerated forgetting factor corresponding to the initial threat probability determined this time is determined by: ; Among them, d s Indicates the s The actual distance between the target Target_ID in the initial threat profile determined this time and the fence.

[0042] The working principle of the above technical solution is as follows: This technical solution processes the initial threat probability through a spatiotemporal forgetting mechanism to obtain the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID. The core of the technical solution is to weight and integrate the historical threat probabilities by combining time decay and spatial accelerated forgetting factors. First, all the initial threat probabilities of the target to date and their corresponding timestamps are retrieved to generate a threat probability queue. Then, each initial threat probability and timestamp in the queue is traversed, the time difference between the current moment and each timestamp is calculated, and the corresponding time decay factor is set based on this time difference. The time decay factor is determined by the time decay rate λ and the time difference. At the same time, the spatial accelerated forgetting factor is calculated based on the actual distance between the target and the fence when the sth initial threat probability is determined. Finally, each initial threat probability is weighted using the time decay factor and spatial accelerated forgetting factor. All initial threat probabilities are integrated through a formula to obtain the ADST-TPM dynamic threat indicator parameter J corresponding to the target at the current moment.

[0043] The above technical solution achieves the following effects: In the temporal dimension, a time decay factor is used to forget historical initial threat probabilities, gradually reducing the influence of earlier threat information over time and emphasizing the weight of recent threat data. This ensures that the ADST-TPM dynamic threat indicator parameters can promptly reflect the latest changes in the target threat, improving the indicator's time sensitivity and dynamic adaptability. In the spatial dimension, a spatial acceleration forgetting factor is introduced to adjust the threat probability based on the actual distance between the target and the fence. When the target is far from the fence, the influence of historical threat information is accelerated to be weakened, while when the target is close to the fence, the weight of historical threat information is relatively retained. This makes the threat indicator parameters more accurate to the target's spatial location characteristics and enhances its responsiveness to changes in the target's spatial situation. Furthermore, the integrated application of the spatiotemporal forgetting mechanism, through weighted processing in both time and space, enables the ADST-TPM dynamic threat indicator parameters to take into account the accumulation of historical threat information while focusing on key threat factors in the current spatiotemporal state. This effectively improves the accuracy, real-time nature, and pertinence of the threat indicator, providing a more accurate and reliable threat assessment basis for the subsequent adaptive control of 3D radar dynamic visualization. At the same time, the technical solution of this embodiment, through the synergistic effect of the time decay factor and the spatial accelerated forgetting factor, not only solves the problem of "how historical threats decay over time", but also supplements the problem of "how spatial distance affects threat weight". Different from the traditional method of processing threats only by time or only by a single dimension of space, it realizes "dynamic attenuation modeling of threats in both time and space dimensions", effectively improving the accuracy of dynamic threat indicator parameters and their adaptability to the actual situation of "time and space jointly affecting risks" in current real scenarios.

[0044] In one embodiment of the present invention, an importance weight map covering all target sets is generated based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, including: Divide the 3D scene into regular voxel grids of 20m×20m×20m in size, and scan each voxel grid to capture the target Target_ID that appears in each voxel grid; Retrieve the decision weight corresponding to the target Target_ID appearing in each voxel grid and the initial weight corresponding to the voxel grid; The maximum value of the decision weight corresponding to the target Target_ID appearing in each voxel grid and the initial weight corresponding to the voxel grid is used as the calibration weight corresponding to each voxel grid; Perform anisotropic diffusion on the calibration weights to obtain the weight values ​​after anisotropic diffusion. Specifically, create a three-dimensional convolution kernel (3×3×3 structure in this example) with a distribution of values ​​[[[0.05, 0.2, 0.05], [0.20, 0.0, 0.20], [0.05, 0.2, 0.05]]], with a central weight of 0 (retaining the original value) and surrounding weights distributed according to a certain pattern to control the diffusion direction and intensity, achieving an "anisotropic" effect (different diffusion degrees in different directions). Perform a convolution operation: Use the convolve3d function with the calibration weights as input and kernel as the convolution kernel. Use the nearest neighbor mode to perform a three-dimensional convolution to obtain the diffused weight values. This step diffuses the voxel weights according to the weights of the surrounding voxels, simulating spatial weight transfer.

[0045] Normalize the weight values ​​after the anisotropic diffusion process to obtain the final weight value corresponding to each voxel grid; The final weight values ​​corresponding to the voxel grid are encoded into a texture format that can be directly rendered by a GPU to generate an importance weight map for all target sets.

[0046] The decision weight corresponding to each target Target_ID is obtained as follows: Retrieve the ADST-TPM dynamic threat indicator parameters corresponding to the current target Target_ID; Performing nonlinear amplification processing on the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID to obtain the ADST-TPM dynamic threat indicator parameter after the nonlinear amplification processing; Real-time monitoring of the distance between the position of the target Target_ID on the screen and the user's eye gaze coordinates; Determine the current user's attention to the target Target_ID by the distance between the position of the target Target_ID on the screen and the user's eye gaze coordinates; The user's attention to the target Target_ID is obtained by the following formula: ; Among them, G represents the user's attention to the target Target_ID; d g Indicates the distance between the position of the target Target_ID on the screen and the user's eye gaze coordinates; σ represents the Gaussian attenuation parameter, which is 50 pixels; The decision weight of the target Target_ID is obtained by combining the user's attention to the target Target_ID with the ADST-TPM dynamic threat indicator parameters of the target Target_ID.

[0047] The above technical solution works as follows: It is based on generating an importance weight map covering all target sets, which is achieved through a multi-step process of spatial partitioning, weight calculation, and processing. First, the 3D scene is divided into a regular voxel grid of 20m×20m×20m. Each grid is scanned to capture the target Target_ID. To calculate the weights, the decision-making weight and initial weight corresponding to the target within each voxel grid are retrieved, and the maximum value of the two is used as the calibration weight for that grid. Obtaining the decision-making weight requires a series of processing: The target's ADST-TPM dynamic threat indicator parameters are retrieved and nonlinearly amplified. The distance between the target's on-screen position and the user's gaze coordinates is simultaneously monitored in real time. The user's attention is calculated using the Gaussian decay formula. This is then combined with the amplified threat indicator parameters to obtain the decision-making weight. Subsequently, the calibration weights are anisotropically diffused. A three-dimensional convolution kernel with a 3×3×3 structure is used to perform a three-dimensional convolution operation in nearest mode to achieve anisotropic diffusion adjustment of the weights. The diffused weight values ​​are then normalized to obtain the final weight value of each voxel grid. Finally, they are encoded into a texture format that can be directly rendered by the GPU to generate an importance weight map.

[0048] The effect of the above technical solution is: in terms of spatial division and weight basis construction, the segmentation of regular voxel grids enables the weight calculation of the three-dimensional scene to have clear spatial units. The calibration weight is determined by combining the decision weight of the target and the maximum value of the initial weight of the grid. This not only highlights the importance of the target's own threat and user attention, but also retains the basic weight attributes inherent in the grid, providing a reasonable starting point for subsequent weight processing.

[0049] The calculation of decision weights comprehensively considers the target's dynamic threat indicators (differences enhanced through nonlinear amplification) and user attention (precisely quantified based on eye movement data), so that the weights can simultaneously reflect the target's objective threat level and the user's subjective focus, enhancing the comprehensiveness and pertinence of the weights.

[0050] The heterogeneous diffusion processing realizes the spatial transfer of weights through a specific three-dimensional convolution kernel, so that the weights of adjacent voxels influence each other and the degree of diffusion in different directions varies, which is more in line with the spatial characteristics of target association in the three-dimensional scene and avoids the isolation of weights; the normalization processing unifies the weight scale to ensure that the weights of different voxels are comparable.

[0051] The resulting weight map, in a texture format that can be directly rendered by the GPU, can efficiently support the rendering process of 3D radar dynamic visualization, enabling the visualization effect to accurately reflect the importance distribution of the target set, improving the efficiency and accuracy of visualization information transmission, and providing a reliable importance basis for subsequent 3D mesh model simplification and performance regulation.

[0052] In one embodiment of the present invention, a three-dimensional point cloud and an importance weight map corresponding to a target Target_ID are combined with operational performance data information to perform topological simplification on a three-dimensional mesh model (3DMesh) to generate a simplified mesh, including: Convert 3D point cloud coordinates into texture coordinates to generate 3D point cloud texture coordinates; Compare the 3D point cloud texture coordinates with the importance weight map to obtain the weight value corresponding to each point cloud; The weight value corresponding to each point cloud is compared with the preset first weight threshold and second weight threshold, and a high-weight target, a medium-weight target and a low-weight target are determined according to the comparison result; wherein the first weight threshold is set to 0.6 and the second weight threshold is set to 0.3; specifically: point clouds with weight values ​​exceeding the first weight threshold are selected as high-weight targets; point clouds with weight values ​​not exceeding the first weight threshold but not lower than the second weight threshold are selected as medium-weight targets; point clouds with weight values ​​lower than the second weight threshold are selected as low-weight targets; Perform edge collapse simplification processing on high-weight targets to obtain geometric details corresponding to high-weight areas; wherein the geometric details include but are not limited to the target's outline, key features, etc.; For medium-weight targets, only those with weights greater than 0.45 are subjected to edge collapse simplification to obtain key features corresponding to the medium-weight region; wherein the key features include but are not limited to the general shape and movement trend of the target; Vertex merging is performed on low-weight targets to obtain basic feature nodes corresponding to low-weight areas; wherein the basic features include but are not limited to the approximate location and existence of the targets.

[0053] The geometric details corresponding to the high-weight areas, the key features corresponding to the medium-weight areas, and the basic features corresponding to the low-weight areas contained in all voxel grids are integrated to generate a simplified grid.

[0054] The working principle of the above technical solution is: this technical solution combines the target's three-dimensional point cloud, importance weight map and operation performance data information to perform hierarchical topological simplification on the three-dimensional grid model and generate a simplified grid. First, the 3D point cloud coordinates are converted to texture coordinates and then compared with the importance weight map to obtain the weight value corresponding to each point cloud. Then, based on the preset first weight threshold (0.6) and second weight threshold (0.3), the point cloud is divided into high-weight targets (weight>0.6), medium-weight targets (0.3≤weight≤0.6), and low-weight targets (weight<0.3). Different simplification strategies are employed for objects of varying weights: high-weight objects are simplified by edge-collapse, preserving their geometric details; medium-weight objects are simplified by edge-collapse only for those with weights exceeding 0.45, preserving key features; and low-weight objects are simplified by vertex merging, preserving essential features. Finally, the features corresponding to regions of varying weights across all voxel meshes are combined to generate a simplified mesh.

[0055] The effect of the above technical solution is: based on the hierarchical simplification strategy of weight values, the geometric details of high-weight targets are retained, the key features of medium-weight targets are maintained, and only the basic features of low-weight targets are retained, realizing differentiated processing of grid simplification, which not only ensures the integrity of important target information, but also reasonably simplifies secondary targets, avoiding the loss of important information or excessive redundant information caused by indiscriminate simplification.

[0056] In terms of model optimization, the precise application of various simplification methods, such as edge collapsing and vertex merging, effectively reduces the complexity of the 3D mesh model while maintaining overall model accuracy. Preserving geometric details in high-weight areas ensures the visualization of key targets, while retaining key features in medium-weight areas maintains the basic shape and motion of the target. Preserving basic features in low-weight areas reduces the model data volume with minimal information loss.

[0057] The simplified mesh generated by the above technical solution can adapt to the requirements of operational performance data information on the basis of balancing model accuracy and complexity, providing an optimized model basis for the performance regulation of subsequent three-dimensional radar dynamic visualization, improving the operational efficiency of the visualization system, and at the same time ensuring the effective transmission of key information, enhancing the adaptability and practicality of the system in different scenarios.

[0058] In one embodiment of the present invention, adaptively controlling a performance index of a three-dimensional radar dynamic visualization operation based on a facet number reduction ratio corresponding to the simplified mesh includes: Determine the number of faces of the current simplified mesh based on the simplified mesh, and retrieve the number of faces of the mesh before simplification; Obtain the face number reduction ratio corresponding to the current simplified mesh according to the face number of the current simplified mesh and the face number of the mesh before simplification; Retrieve the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID appearing on the current screen; The frame rate and display bandwidth occupancy of the current 3D radar dynamic visualization operation are adaptively controlled using the facet reduction ratio corresponding to the current simplified mesh and the ADST-TPM dynamic threat index parameters corresponding to each target Target_ID.

[0059] The frame rate and display bandwidth occupancy of the current 3D radar dynamic visualization are adaptively controlled using the facet reduction ratio corresponding to the current simplified mesh and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, including: Retrieve the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID and the final weight value corresponding to the voxel grid where each target Target_ID is located; The threat degree coefficient is obtained by performing weighted averaging processing using the ADST-TPM dynamic threat index parameter corresponding to each target Target_ID and the final weight value corresponding to the corresponding voxel grid; Retrieve the face reduction ratio corresponding to the current simplified mesh; The adjusted frame rate and display bandwidth occupancy are obtained using the facet number reduction ratio and the threat degree coefficient corresponding to the current simplified mesh.

[0060] The adjusted frame rate and display bandwidth occupancy are obtained by the following formula: ; in, FPS Indicates the adjusted frame rate; FPS old Indicates the frame rate before regulation; H Indicates the face reduction ratio corresponding to the current simplified mesh; Y Indicates the threat level coefficient; ; in, B us Indicates the display bandwidth usage after regulation; B oldIndicates the display bandwidth usage before regulation; H Indicates the face reduction ratio corresponding to the current simplified mesh; Y Indicates the threat level coefficient.

[0061] The working principle of the above technical solution is: this technical solution aims to adaptively control the performance indicators of the dynamic visualization operation of the three-dimensional radar based on the reduction ratio of the number of faces in the simplified grid. The core is to achieve dynamic adjustment of the frame rate and display bandwidth occupancy through quantitative analysis and formula calculation. First, the number of facets in the current simplified mesh is determined, and the number of facets in the mesh before simplification is retrieved. The ratio of the two is calculated to obtain the facet reduction ratio. Next, the ADST-TPM dynamic threat indicator parameter for each target on the current screen and the final weight value of the voxel grid in which it resides are retrieved. These two parameters are weighted averaged to generate the threat level coefficient. Subsequently, the patch reduction ratio and threat level coefficient are substituted into a specific formula to calculate the pre-control frame rate and display bandwidth utilization, resulting in the post-control frame rate (FPS) and display bandwidth utilization (Bus). The frame rate control formula is based on the pre-control frame rate and is adjusted in conjunction with the patch reduction ratio and threat level coefficient. The display bandwidth utilization control formula is also based on the pre-control value and dynamically adapts to the patch reduction ratio and threat level coefficient.

[0062] The above technical solution achieves precise performance control by quantifying the degree of mesh simplification through the facet reduction ratio. Combined with the threat level coefficient (which integrates the target's dynamic threat and voxel weight), the system adjusts the frame rate and display bandwidth usage to closely match the current mesh complexity and target threat status. When the mesh simplification level is high (large reduction ratio) and the threat level is low, the system can appropriately reduce the frame rate and bandwidth usage to avoid wasted resources. However, when the threat level is high, even with mesh simplification, the coefficient adjustment ensures performance support in critical scenarios, achieving a precise match between performance and demand. In terms of system resource optimization, a control mechanism based on quantitative parameters avoids blind adjustments to performance indicators, enabling a dynamic balance between frame rate and bandwidth utilization based on actual scenarios. While ensuring visualization of high-threat targets, it effectively reduces resource consumption in low-threat scenarios, improving the efficiency and stability of system operation. Furthermore, a unified, formulaic control logic ensures consistent and predictable performance adjustments, reduces the uncertainty of human intervention, and enhances the adaptability and overall operational efficiency of the 3D radar dynamic visualization system in complex environments.

[0063] The embodiment of the present invention proposes an adaptive control system suitable for dynamic visualization of three-dimensional radar, such as Figure 2 As shown, the adaptive control system includes: Data acquisition module, used to collect basic 3D radar data and operating performance data in real time; The dynamic threat indicator parameter acquisition module is used to perform ADST-TPM dynamic threat prediction on basic 3D radar data information using the spatiotemporal forgetting mechanism, and obtain the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID; An importance weight map acquisition module is used to generate an importance weight map covering all target sets based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, wherein the importance weight map is used to associate a specific target Target_ID with the point cloud cluster corresponding to the target Target_ID; The simplified mesh acquisition module is used to simplify the topology of the three-dimensional mesh model (3DMesh) by using the three-dimensional point cloud and importance weight map corresponding to the target Target_ID in combination with the operation performance data information to generate a simplified mesh; The adaptive control module is used to adaptively control the performance index of the three-dimensional radar dynamic visualization operation according to the facet number reduction ratio corresponding to the simplified grid.

[0064] The working principle of the above technical solution is as follows: It achieves adaptive control of 3D radar dynamic visualization through the coordinated operation of multiple links. First, basic 3D radar data and operational performance data are collected in real time to provide raw data support for subsequent processing. Subsequently, ADST-TPM dynamic threat prediction is performed on the basic 3D radar data using a spatiotemporal forgetting mechanism. Focusing on each target Target_ID, the corresponding ADST-TPM dynamic threat indicator parameters are calculated to quantify the target's threat level. Next, combining the 3D scene and the ADST-TPM dynamic threat indicator parameters for each target, an importance weight map covering all target sets is generated. This weight map establishes the association between specific target Target_IDs and their corresponding point cloud clusters, clarifying the importance hierarchy of different targets in the 3D scene. Furthermore, using the corresponding 3D point clouds, the importance weight map, and operational performance data, the 3D mesh model (3DMesh) is topologically simplified to produce a simplified mesh. Finally, based on the facet reduction ratio corresponding to the simplified mesh, the performance indicators of the 3D radar dynamic visualization are adaptively controlled, forming a complete closed loop from data acquisition, processing, model optimization, to performance adjustment.

[0065] The benefits of this technical solution are as follows: At the data processing level, real-time collection of two types of data ensures the timeliness and comprehensiveness of information, laying a solid foundation for accurate processing in subsequent steps and avoiding processing biases caused by delayed or missing data. ADST-TPM dynamic threat prediction, through the spatiotemporal forgetting mechanism, accurately captures the dynamic threat characteristics of the target, effectively improving the accuracy and dynamic adaptability of target threat assessments, ensuring that threat indicator parameters are more closely aligned with the target's actual state changes.

[0066] In terms of model optimization, an importance weight map generated based on the 3D scene and target threat indicators clearly associates targets with point cloud clusters, providing a precise importance basis for simplifying the topological structure of the 3D mesh model. Mesh simplification, combining the target 3D point cloud, importance weight map, and operational performance data, ensures accurate visualization of key targets while enabling reasonable simplification of non-critical areas. The resulting simplified mesh significantly reduces model complexity while maintaining visual quality.

[0067] In terms of performance control, adaptively adjusting performance indicators based on the number of facets reduced in the simplified mesh dynamically balances visualization effects with system efficiency. This not only optimizes system resource usage and increases data processing and visualization speed, but also automatically adjusts based on actual performance, ensuring the 3D radar dynamic visualization system maintains stable and efficient operation in various scenarios, significantly improving the system's overall performance and applicability.

Claims

1. An adaptive control method for three-dimensional radar dynamic visualization, characterized in that: The adaptive control method comprises: Real-time collection of basic 3D radar data and operational performance data; Use the spatiotemporal forgetting mechanism to perform ADST-TPM dynamic threat prediction on basic 3D radar data information and obtain the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID; Generate an importance weight map covering all target sets based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, wherein the importance weight map is used to associate a specific target Target_ID with the point cloud cluster corresponding to the target Target_ID; The 3D point cloud and importance weight map corresponding to the target Target_ID are combined with the operation performance data to simplify the topology of the 3D mesh model and generate a simplified mesh. The performance index of the dynamic visualization operation of the three-dimensional radar is adaptively regulated according to the facet number reduction ratio corresponding to the simplified grid.

2. The adaptive control method for three-dimensional radar dynamic visualization according to claim 1, characterized in that: Real-time collection of basic 3D radar data and operational performance information, including: During the operation of the 3D radar, basic 3D radar data information of the 3D radar is retrieved in real time; wherein the basic 3D radar data information includes raw radar data, target tracking data, geo-fence data, multi-target relationship data, user interaction enhancement data and system load data; Real-time collection of operational performance data during 3D radar dynamic visualization; wherein the operational performance data includes frame rate, GPU occupancy, ADST-TPM latency, and point cloud throughput.

3. The adaptive control method for three-dimensional radar dynamic visualization according to claim 1, characterized in that: The spatiotemporal forgetting mechanism is used to perform ADST-TPM dynamic threat prediction on basic 3D radar data information, and the ADST-TPM dynamic threat indicator parameters corresponding to each target are obtained, including: Retrieving a radar scanning period Δt from raw radar data included in the basic three-dimensional radar data information, setting a cache window duration for target data according to the radar scanning period Δt, and caching the target data for the radar scanning period Δt; The radial acceleration, stability index and peak frequency shift of the target Target_ID are retrieved from the target tracking data contained in the basic 3D radar data information, and the Doppler dynamic feature vector is generated using the radial acceleration, stability index and peak frequency shift. ;in, A Indicates the radial acceleration of the target Target_ID; stability Indicates the stability index of target Target_ID; f peak Indicates the peak frequency shift of the target Target_ID; Retrieve the geofence list and target operation data corresponding to the target Target_ID, and obtain the spatiotemporal context feature vector corresponding to the target Target_ID based on the geofence list and target operation data ,in, W 01 and W 02 They represent the geographical threat correlation and target interaction threat corresponding to the target Target_ID respectively; Performing weighted fusion on the Doppler dynamics feature vector and the spatiotemporal context feature vector to generate a fused feature vector; Monitor the GPU utilization of the current system in real time and select a trained threat assessment model based on the GPU utilization; wherein the threat assessment model includes a full GRU network model, a lightweight temporal convolution model, and a lightweight temporal convolution model; The fused feature vector is input into the threat assessment model, and the initial threat probability corresponding to the target Target_ID is output. The initial threat probability is forgotten using the spatiotemporal forgetting mechanism to obtain the ADST-TPM dynamic threat indicator parameter corresponding to the target Target_ID.

4. The adaptive control method for three-dimensional radar dynamic visualization according to claim 3, characterized in that: Retrieve the geofence list and target operation data corresponding to the target Target_ID, and obtain the spatiotemporal context feature vector corresponding to the target Target_ID based on the geofence list and target operation data, including: Traverse the geofence list and get the minimum distance between the target Target_ID location and the geofence; Retrieving a priority value corresponding to a fence from the geo-fence list; The minimum distance between the target Target_ID location and the geo-fence is used in combination with the priority value corresponding to the fence in the geo-fence list to obtain the geo-threat relevance corresponding to the target Target_ID; Search for neighboring targets within 500 meters of the target Target_ID and select those with an initial threat probability greater than 0.7 as high-threat neighboring targets; The target interaction threat corresponding to the target Target_ID is obtained through the velocity vector of the target Target_ID and the velocity vector of the high-threat neighboring target.

5. The adaptive control method for three-dimensional radar dynamic visualization according to claim 3, characterized in that: Monitor the GPU utilization of the current system in real time and select a trained threat assessment model based on the GPU utilization, including: When the GPU utilization is less than 0.7, the complete GRU network model is selected as the threat assessment model; When the GPU utilization is greater than or equal to 0.7 and less than 0.83, the lightweight temporal convolution model is selected as the threat assessment model; When the GPU utilization is greater than or equal to 0.83, the rule engine model is selected as the threat assessment model; The predefined rules corresponding to the rule engine model are as follows: First check the current GPU load. When the GPU load exceeds 90%, enable the rule engine degradation mode. Define the rule engine function and create a function named rule_engine. This function receives the fused feature vector of the target Target_ID as input and outputs a floating-point value representing the threat level. The execution rules of the rule_engine function are as follows: Rule 1: If the radial acceleration of target_ID (target.a_radial) is greater than 20.0, and the distance from target_ID to the geofence (target.d_geo) is less than 5000.0, then the target is considered to be approaching the no-fly zone at high speed, indicating a high threat situation. The initial threat probability is returned to 0.

9. Rule 2: Check the peak frequency shift attribute target.Δf_peak of the target Target_ID. If the peak frequency shift attribute target.Δf_peak is greater than 50, it is determined to be a threat scenario with a drastic change in the Doppler spectrum and the initial threat probability is returned to 0.

7. Rule 3: Check the target's target interaction threat attribute target.neighbor_threat. If the value is greater than 1.5, calculate the initial threat probability using the function min(0.8, 0.3 + target.neighbor_threat * 0.2) and return the calculated result. Rule 4: The default initial threat probability is returned. If all the above rule conditions are not met, it means that the target threat level is low, and the default initial threat probability of 0.3 is returned.

6. The adaptive control method for three-dimensional radar dynamic visualization according to claim 1, characterized in that: Generate an importance weight map covering all target sets based on the 3D scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, including: Divide the 3D scene into regular voxel grids of 20m×20m×20m in size, and scan each voxel grid to capture the target Target_ID that appears in each voxel grid; Retrieve the decision weight corresponding to the target Target_ID appearing in each voxel grid and the initial weight corresponding to the voxel grid; The maximum value of the decision weight corresponding to the target Target_ID appearing in each voxel grid and the initial weight corresponding to the voxel grid is used as the calibration weight corresponding to each voxel grid; Performing anisotropic diffusion processing on the calibration weights to obtain weight values ​​after the anisotropic diffusion processing; Normalize the weight values ​​after the anisotropic diffusion process to obtain the final weight value corresponding to each voxel grid; The final weight values ​​corresponding to the voxel grid are encoded into a texture format that can be directly rendered by a GPU to generate an importance weight map for all target sets.

7. The adaptive control method for three-dimensional radar dynamic visualization according to claim 1, characterized in that: The 3D point cloud and importance weight map corresponding to the target Target_ID are combined with the operational performance data to simplify the topology of the 3D mesh model and generate a simplified mesh, including: Convert 3D point cloud coordinates into texture coordinates to generate 3D point cloud texture coordinates; Compare the 3D point cloud texture coordinates with the importance weight map to obtain the weight value corresponding to each point cloud; Compare the weight value corresponding to each point cloud with a preset first weight threshold and a second weight threshold, and determine a high-weight target, a medium-weight target, and a low-weight target based on the comparison result; wherein the first weight threshold is set to 0.6 and the second weight threshold is set to 0.3; Perform edge folding and simplification processing on high-weight targets to obtain the geometric details corresponding to the high-weight areas; For medium-weight targets, only those with weight values ​​greater than 0.45 are subjected to edge folding simplification to obtain the key features corresponding to the medium-weight region; Vertex merging is performed on low-weight targets to obtain the basic feature nodes corresponding to the low-weight areas; the geometric details corresponding to the high-weight areas, the key features corresponding to the medium-weight areas, and the basic features corresponding to the low-weight areas contained in all voxel grids are integrated to generate a simplified mesh.

8. The adaptive control method for three-dimensional radar dynamic visualization according to claim 1, characterized in that: Adaptively regulating the performance indicators of the three-dimensional radar dynamic visualization operation according to the facet number reduction ratio corresponding to the simplified grid, including: Determine the number of faces of the current simplified mesh based on the simplified mesh, and retrieve the number of faces of the mesh before simplification; Obtain the face number reduction ratio corresponding to the current simplified mesh according to the face number of the current simplified mesh and the face number of the mesh before simplification; Retrieve the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID appearing on the current screen; The frame rate and display bandwidth occupancy of the current 3D radar dynamic visualization operation are adaptively controlled using the facet reduction ratio corresponding to the current simplified mesh and the ADST-TPM dynamic threat index parameters corresponding to each target Target_ID.

9. The adaptive control method for three-dimensional radar dynamic visualization according to claim 8, characterized in that: The frame rate and display bandwidth occupancy of the current 3D radar dynamic visualization are adaptively controlled using the facet reduction ratio corresponding to the current simplified mesh and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, including: Retrieve the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID and the final weight value corresponding to the voxel grid where each target Target_ID is located; The threat degree coefficient is obtained by performing weighted averaging processing using the ADST-TPM dynamic threat index parameter corresponding to each target Target_ID and the final weight value corresponding to the corresponding voxel grid; Get the face reduction ratio corresponding to the current simplified mesh; The adjusted frame rate and display bandwidth occupancy are obtained using the facet number reduction ratio and the threat degree coefficient corresponding to the current simplified mesh.

10. An adaptive control system for three-dimensional radar dynamic visualization, characterized in that: The adaptive control system comprises: Data acquisition module, used to collect basic 3D radar data and operating performance data in real time; The dynamic threat indicator parameter acquisition module is used to perform ADST-TPM dynamic threat prediction on basic 3D radar data information using the spatiotemporal forgetting mechanism, and obtain the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID; An importance weight map acquisition module is used to generate an importance weight map covering all target sets based on the three-dimensional scene and the ADST-TPM dynamic threat indicator parameters corresponding to each target Target_ID, wherein the importance weight map is used to associate a specific target Target_ID with the point cloud cluster corresponding to the target Target_ID; The simplified mesh acquisition module is used to simplify the topology of the three-dimensional mesh model by using the three-dimensional point cloud and importance weight map corresponding to the target Target_ID in combination with the operation performance data information to generate a simplified mesh; The adaptive control module is used to adaptively control the performance index of the three-dimensional radar dynamic visualization operation according to the facet number reduction ratio corresponding to the simplified grid.

Images