Target processing method and device for X-band radar noise region, and storage medium

Abstract

The invention discloses a target processing method and device for an X-band radar noise area and a storage medium. The method comprises the steps that a single-frame grid weight coefficient, a noise area judgment threshold value, a clutter area target quality threshold value and a non-clutter area target quality threshold value are set; constructing a two-dimensional grid chart covering the full detection range of the radar; processing radar single-frame plot data to generate single-frame grid marking information; carrying out weighted fusion on the single-frame grid marking information of the current frame and the historically accumulated grid marking information according to the single-frame grid weight coefficient, and updating the noise area graph; for each tracked target, calculating a comprehensive quality factor based on historical track information of the tracked target; and for each tracked target, querying the noise area map according to the current position of the target, judging whether the target is located in a clutter area, and screening target quality factors. Finally, effective distinguishing of spatial features, balance of time fusion, accurate evaluation of target quality and flexible adaptation to different scenes are realized.

Description

Technical Field

[0001] This application relates to radar signal processing technology, specifically to a target processing method, device, and storage medium for X-band radar noise regions. Background Technology

[0002] X-band radar, with its shorter wavelength, exhibits high spatial resolution and excellent detection capabilities for small targets in scenarios such as maritime surveillance, port security, and low-altitude early warning. However, in practical applications, radar echo signals inevitably contain various interference components, including random noise from receiver thermal noise and circuit interference, as well as clutter caused by ocean waves, nearshore reefs, dock facilities, or weather phenomena. These interferences are often comparable in intensity to the echoes of small targets, especially at close range, leading to the generation of numerous false targets and severely interfering with the accuracy of target tracking and the reliability of situational assessment. Existing technologies mainly rely on threshold suppression, feature extraction, and region segmentation to address noise issues, but these methods have significant limitations. In the spatial dimension, traditional cell-averaged CFAR or ordered statistical CFAR algorithms only set thresholds based on signal strength statistics within a local sliding window, failing to capture the spatial distribution patterns of clutter across the entire radar field of view. When clutter exhibits characteristics of local density and global unevenness, such as the range-dependent changes in ground clutter caused by terrain undulations or sea clutter, the local sliding window is prone to misjudging dense clutter as targets, or causing missed detection of real targets due to sudden changes in clutter. At the same time, existing methods usually only mark the coordinates of the center position of the clutter point, ignoring the spatial clustering characteristics of real targets caused by the radar beam broadening effect, while isolated clutter points are discretely distributed. A single marking method cannot effectively distinguish the spatial extension of the target from the randomness of clutter, weakening the spatial anti-clutter capability. In terms of time dimension, some solutions lack a time accumulation mechanism, and direct screening of single-frame points is easily affected by instantaneous electromagnetic interference or impulse noise, causing false targets to appear instantaneously or real targets to be missed due to signal attenuation. Other solutions, although using moving average fusion with fixed coefficients, do not design adjustable single-frame weights, making it difficult to adapt to the real-time requirements of clutter-prone environments and the stability requirements of stable environments. Furthermore, they do not accumulate and record the clutter positions of consecutive frames, and cannot take advantage of the difference between the fixed or slowly changing clutter regions and the continuous movement of target trajectories, resulting in poor suppression of continuous dense clutter. In the quality assessment stage, traditional methods only set a fixed threshold based on trajectory length, failing to distinguish the contribution differences between matching state trajectory points (consistent with the detection results) and predicted state trajectory points (generated solely through prediction) in historical trajectories. Real targets typically have a high proportion of matching points, while false targets have a high proportion of predicted points; a single-dimensional assessment cannot accurately quantify target credibility. Furthermore, the use of a uniform standard for threshold selection fails to consider the environmental differences between clutter and non-clutter areas. Target quality in clutter areas is easily affected by interference and tends to be lower; a fixed high threshold leads to missed detections of real targets. In non-clutter areas, accidental matching may generate high-quality false targets, while a fixed low threshold increases the false alarm rate. A single threshold mechanism cannot adequately address the detection needs of different scenarios. These shortcomings collectively result in deficiencies in existing technologies regarding spatial feature differentiation, temporal fusion balance, accuracy of quality assessment, and flexibility in scenario adaptation, making it difficult to effectively control the false alarm and missed detection rates of radar target detection in complex environments.

[0003] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0004] The purpose of this application is to provide a target processing method, device and storage medium for X-band radar noise areas, which has the advantages of effectively distinguishing spatial features, balancing time fusion, accurately assessing target quality and flexibly adapting to different scenarios, thereby reducing false alarm rate and false detection rate.

[0005] This application provides a target processing method for noise areas in X-band radar, and the technical solution is as follows: The target processing method for X-band radar noise areas includes the following steps: Step 1: System initialization; Set global parameters including single-frame grid weight coefficient K, noise region judgment threshold T, clutter region target quality threshold MN, and non-clutter region target quality threshold MT; Step 2: Grid cell construction; Construct a two-dimensional grid map covering the entire radar detection range to record the cumulative value of the point markers in each grid cell; Step 3: Single-frame point mark processing and spatial weighted labeling; Process radar single-frame point mark data, calculate the center position distance R and azimuth B of each point mark, and use the center grid cell of the point mark as the core to perform weighted labeling on the grid cells within a predetermined range around it to generate single-frame grid labeling information. Step 4: Multi-frame temporal fusion update of the noise region map; The single-frame grid label information of the current frame is weighted and fused with the historical accumulated grid label information according to the single-frame grid weight coefficient K to update the multi-frame accumulated grid label information that reflects the continuous noise distribution, i.e., the noise region map. Step 5: Target quality calculation; For each tracked target, calculate a comprehensive quality factor Q based on its historical trajectory information. The historical trajectory information includes the number of trajectory points in the matching state, the number of trajectory points in the prediction state, and the total trajectory length. Step Six: Differentiated Threshold Filtering and Target Output: For each tracked target, query the noise region map based on its current location. If the cumulative grid marker value at that location is greater than the noise region judgment threshold T, the target is determined to be located within the clutter region, and the target quality factor Q is filtered using the clutter region target quality threshold MN. If the cumulative grid marker value at that location is less than or equal to the noise region judgment threshold T, the target is determined to be located within the non-clutter region, and the target quality factor Q is filtered using the non-clutter region target quality threshold MT. Targets that reach the corresponding quality threshold are output as true targets.

[0006] Furthermore, this application also proposes that the "weighted labeling of the grid cells within a predetermined range around it" in step three specifically involves: using the grid cell at the center of the dot as the core, labeling the grid cells in the surrounding 3×3 area, with the central area having the highest weight and decreasing towards the outer layers.

[0007] Furthermore, this application also proposes that the specific calculation formula for "weighted fusion" in step four is: new cumulative grid label information = historical cumulative grid label information × (1-K) + current frame single-frame grid label information × K.

[0008] Furthermore, this application also proposes that the specific calculation formula for "calculating a comprehensive quality factor Q" in step five is: Q=[(number of matching state trajectory points - number of predicted state trajectory points) / total trajectory length]*P+[(1-P)*total trajectory length / maximum allowed trajectory length], where P is a preset weighting coefficient.

[0009] Furthermore, this application also proposes that the target quality threshold MN in the clutter region is greater than the target quality threshold MT in the non-clutter region, i.e., MN < MT.

[0010] Furthermore, this application also proposes that the dimensions of the "two-dimensional grid map" in step one correspond to the azimuth dimension and range dimension of the radar, with the azimuth dimension ranging from 0 to 4095 and the range dimension covering the effective detection range of the radar.

[0011] Furthermore, this application also proposes a target processing device for X-band radar noise areas, comprising a processor, a memory, and a target processing program for X-band radar noise areas stored in the memory and executable by the processor, wherein when the target processing program for X-band radar noise areas is executed by the processor, the steps of the aforementioned target processing method for X-band radar noise areas are performed.

[0012] Furthermore, this application also proposes a storage medium storing a target processing program for an X-band radar noise region, wherein when the target processing program for the X-band radar noise region is executed by a processor, the steps of the aforementioned target processing method for the X-band radar noise region are performed.

[0013] As can be seen from the above, the target processing method, device and storage medium for X-band radar noise area provided in this application constructs a two-dimensional grid map covering the entire detection range of the radar, processes single-frame point data and performs spatial weighting and labeling, updates the noise area map through multi-frame temporal fusion, calculates the target comprehensive quality factor and performs differentiated threshold screening. It has the advantages of effectively distinguishing spatial features, balancing temporal fusion, accurately assessing target quality, and flexibly adapting to different scenarios, thereby reducing false alarm rate and false detection rate. Attached Figure Description

[0014] Figure 1 This is a schematic flowchart of an embodiment of the target processing method for X-band radar noise region of this application. Figure 2 This is a schematic diagram of the hardware structure of the target processing device in the X-band radar noise region involved in the embodiments of this application. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0016] In X-band radar applications for maritime observation, the intensity of noise and clutter is often comparable to that of small target echoes, easily creating false targets. Existing technologies, in terms of spatial processing, only set thresholds based on signal strength statistics within a local sliding window, ignoring the spatial distribution patterns of clutter across the entire radar field of view and failing to distinguish between the spatial clustering of real target targets and the discrete distribution of isolated clutter points. Furthermore, in the temporal dimension, there is a lack of an adjustable single-frame weighted temporal fusion mechanism, making it unsuitable for the real-time and stability requirements of different scenarios, and clutter positions are not cumulatively recorded. In addition, quality assessment relies solely on trajectory length for a single-dimensional evaluation, failing to consider the difference in the ratio of matched to predicted trajectory points, and using fixed thresholds for filtering, which cannot adapt to the environmental differences between clutter and non-clutter areas. Therefore, in cluttered environments, it is difficult to simultaneously achieve a high detection rate and a low false alarm rate for target detection, affecting target tracking accuracy and situational awareness.

[0017] If the above problems are not solved, the continuous output of false dots will lead to an increase in the cumulative error of the target tracking system, interruption of the real target trajectory, and a significant decrease in situational awareness. In complex clutter environments, the system may not be able to reliably distinguish between real targets and noise, increasing the burden of subsequent data processing, reducing overall detection performance, and making it difficult to meet the real-time and accuracy requirements of key applications such as maritime monitoring and port security.

[0018] Therefore, in a first aspect, this application provides a target processing method for X-band radar noise regions to solve the above-mentioned problems, which includes the following steps: Step 1: System initialization; Global parameters for single-frame grid weight coefficient K, noise region judgment threshold T, clutter region target quality threshold MN, and non-clutter region target quality threshold MT are set. Step 2: Grid cell construction; A two-dimensional grid map covering the entire radar detection range is constructed to record the cumulative value of the point markers in each grid cell; Step 3: Single-frame point data processing and spatial weighted labeling; The radar single-frame point data is processed, the center position distance R and azimuth B of each point are calculated, and the grid cells around the center of the point are weighted and labeled with different weights, with the center grid cell as the core, and the grid cells around the point within a predetermined range are generated. Step 4: Multi-frame temporal fusion update of the noise region map; The single-frame grid labeling information of the current frame and the historically accumulated grid labeling information are weighted and fused according to the single-frame grid weighting coefficient K, and the multi-frame accumulated grid labeling information reflecting the continuous noise distribution is updated, that is, the noise region map is obtained. Step 5: Target quality calculation; For each tracked target, a comprehensive quality factor Q is calculated based on its historical trajectory information, which includes the number of trajectory points in the matching state, the number of trajectory points in the prediction state, and the total trajectory length. Step Six: Differentiated Threshold Filtering and Target Output: For each tracked target, the noise region map is queried at its current location. If the cumulative grid marker value at that location is greater than the noise region judgment threshold T, the target is determined to be located within the clutter region and its quality factor Q is filtered using the clutter region target quality threshold MN. If the cumulative grid marker value at that location is less than or equal to the noise region judgment threshold T, the target is determined to be located within the non-clutter region and its quality factor Q is filtered using the non-clutter region target quality threshold MT. Targets that reach the corresponding quality threshold are output as true targets.

[0019] This embodiment provides a target processing method for X-band radar noise regions, which achieves efficient and accurate detection of real targets in complex environments through a series of steps.

[0020] First, during system initialization, a series of global parameters need to be set. These parameters affect the performance of subsequent processing flows. For example, the single-frame grid weight coefficient K can be set to a fixed value to balance the weights of current frame information and historical accumulated information during temporal fusion. The noise region judgment threshold T can be set to an empirical value to distinguish clutter regions from non-clutter regions in the noise region map. The clutter region target quality threshold MN and the non-clutter region target quality threshold MT can be set to preset fixed values ​​to filter target quality factors in different regions. These parameters can be set manually or loaded by reading a pre-stored configuration file.

[0021] Next, the system constructs a grid. To manage spatial information across the entire radar detection range in a refined manner, a two-dimensional grid map needs to be built. This grid map can be divided into multiple discrete grid cells, each corresponding to a specific region in the radar detection space. For example, a uniform division can be used, dividing the radar's azimuth and range dimensions into several equal parts, thus forming a rectangular grid. Each grid cell is allocated a storage space to record the accumulated value of its internal point markers. The construction of this grid map can be completed all at once during system initialization or dynamically adjusted according to the radar's detection range.

[0022] Subsequently, single-frame point tracking processing and spatial weighted labeling are performed. The radar system generates a series of point tracking data in each time frame. Processing this single-frame point tracking data first requires calculating the center position of each point, typically represented by distance R and azimuth B. Specifically, the distance information for the center position of each point is calculated using the distance data Rmin of the smallest sampled point and Rmax of the largest sampled point in this batch of point tracking data, as R = (Rmin + Rmax) / 2. Similarly, the azimuth information for the center position of each point is calculated using the minimum azimuth Bmin and the maximum azimuth Bmax in this batch of point tracking data, as B = (Bmin + Bmax) / 2. This generates single-frame grid labeling information reflecting the point tracking distribution of the current frame.

[0023] Furthermore, multi-frame temporal fusion is performed to update the noise region map. To obtain a noise region map reflecting the continuous noise distribution, the single-frame grid marker information of the current frame needs to be fused with the historical accumulated grid marker information of previous frames. For example, a simple accumulation method can be used, directly adding the marker value of the current frame to the historical accumulated value. This fusion process is performed in each time frame, enabling the noise region map to be dynamically updated and reflecting the long-term spatial distribution of noise and clutter.

[0024] Based on this, target quality is calculated. For each target tracked by the radar system, its reliability needs to be evaluated. This is achieved by calculating a comprehensive quality factor Q. The calculation of this quality factor Q is based on the target's historical trajectory information. For example, the quality factor Q can simply be the ratio of the number of trajectory points in a matching state to the total trajectory length. The number of trajectory points in a matching state refers to the number of trajectory points successfully associated with actual radar detection points. The number of trajectory points in a predicted state refers to the number of trajectory points generated by the prediction algorithm when there are no actual detection points to match. The total trajectory length refers to the total number of trajectory points contained in the target from the start of tracking to the current frame.

[0025] Finally, differential threshold filtering and target output are performed. After the target quality factor Q is calculated, differential filtering is required based on the target's environment. First, based on the target's current location, the cumulative grid marker value at that location is obtained from the noise region map. Then, this cumulative value is compared with the noise region judgment threshold T. If the cumulative value is greater than the threshold T, the target is considered to be located within a clutter region, and the clutter region target quality threshold MN is used to judge the target quality factor Q. If the cumulative value is less than or equal to the threshold T, the target is considered to be located within a non-clutter region, and the non-clutter region target quality threshold MT is used to judge the target quality factor Q. For example, a fixed quality threshold can be set, and this fixed threshold is used for filtering regardless of the target's location. Only when the target quality factor Q reaches the quality threshold of the corresponding region is the target recognized as a real target and output.

[0026] The method provided in this embodiment demonstrates significant technological advancements when processing targets in noisy X-band radar regions.

[0027] First, in the spatial dimension, this embodiment constructs a two-dimensional mesh map and weights the mesh cells within a predetermined range surrounding the point trace, achieving a refined description of the spatial distribution of the point trace. For example, in the aforementioned port scenario, traditional methods may only record the center position of the point trace, failing to effectively distinguish between point trace aggregation caused by the widening characteristics of the real target and isolated clutter points. This embodiment, by marking the central mesh cell and its surrounding mesh cells, can more accurately reflect the spatial extensibility of the point trace, thereby enhancing its anti-clutter capability in the spatial dimension.

[0028] Secondly, in terms of time dimension, this embodiment updates the noise region map through multi-frame time fusion, effectively balancing real-time performance and stability. In the example above, by weighted fusion of the single-frame grid marker information of the current frame with historical accumulated information, the system can continuously accumulate the marker values ​​of clutter regions. This differs from traditional methods that rely solely on single-frame point screening or fixed-coefficient moving averages, which suffer from insufficient real-time performance in clutter-prone environments or weak noise resistance in stable environments. The fusion mechanism of this embodiment ensures that persistent fixed clutter (such as dock facilities and reefs) exhibits higher accumulated values ​​on the noise region map, while instantaneous noise or moving targets do not cause their accumulated values ​​to rise rapidly. This effectively distinguishes between persistent clutter and real targets, improving the suppression effect on persistent dense clutter.

[0029] Furthermore, regarding quality assessment and threshold selection, this embodiment introduces a comprehensive quality factor Q calculation based on historical trajectory information, combined with a noise region map for differentiated threshold selection. In the example above, the calculation of the target quality factor Q considers the number of matching state trajectory points, the number of predicted state trajectory points, and the total trajectory length, which is more comprehensive and accurate than the traditional assessment method that relies solely on trajectory length. More importantly, this embodiment dynamically selects different quality thresholds (MN or MT) for selection based on the target's location (cluttered or non-cluttered area). This solves the problem that traditional fixed quality thresholds or fixed clutter thresholds cannot meet the detection needs of different environments. For example, in cluttered areas, even if the target quality factor Q is slightly lower due to clutter interference, it can still be identified as a real target as long as a lower MN threshold is reached, avoiding missed detections; while in non-cluttered areas, a higher MT threshold is used to effectively suppress false targets. This differentiated selection mechanism significantly improves environmental adaptability, enabling a balance between detection rate and false alarm rate in complex cluttered environments.

[0030] In summary, this embodiment addresses the problems of insufficient spatial feature differentiation, poor temporal fusion balance, and lack of specificity in quality assessment and threshold screening in existing technologies through the collaborative design of spatial grid weighted labeling, adaptive temporal fusion, multi-dimensional quality assessment, and differentiated threshold screening. This enables efficient and accurate detection of real targets in complex environments.

[0031] This application further proposes that the "weighted labeling of the grid cells within a predetermined range around it" in step three is specifically as follows: taking the grid cell at the center of the dot as the core, the grid cells in the surrounding 3×3 area are labeled, with the central area having the highest weight and decreasing layer by layer towards the periphery.

[0032] Specifically, "using the center grid cell of the point track as the core" means that when processing radar single-frame point track data, the center position (range R and azimuth B) of each point track is first calculated, and this center position is mapped to a specific grid cell in a pre-constructed two-dimensional grid map. This grid cell is then determined as the spatial marker core of the point track, serving as the reference point for subsequent weighted marking operations. This ensures that the spatial influence of the point track can spread from its most definitive location.

[0033] "Marking the surrounding 3×3 grid cells" refers to selecting eight neighboring grid cells, along with the grid cell at the center of the dot pattern, to form a 3×3 local grid region. A weighted marking operation is then performed on all nine grid cells within this region. This limited local region marking method effectively captures the local diffusion characteristics of the dot pattern in space, avoiding unnecessary propagation of dot pattern information over an excessively large area, thereby improving the accuracy of marking and computational efficiency. Besides a 3×3 region, it can also be a 5×5 region, or a circular region with a specific radius centered on the dot pattern.

[0034] The statement "the central region has the highest weight, decreasing layer by layer outwards" means that within the 3×3 area, the central grid cell is assigned the highest weight value, while the weight value decreases as the grid cell is farther from the center. This weight allocation strategy reflects the spatial decay characteristics of point information; that is, the information at the center of the point is the most reliable and has the greatest influence, while the influence gradually weakens in areas farther from the center. This layer-by-layer decreasing weight allocation allows for a more refined depiction of the spatial distribution characteristics of the points, enabling the generated single-frame grid marker information to more accurately reflect the actual spatial probability or intensity of the points. Besides layer-by-layer decreasing weights, weights can also be allocated according to a Gaussian distribution or an exponential decay function.

[0035] This application's scheme uses the central grid cell of a point as the core in the single-frame point processing and spatial weighted labeling steps, limiting the labeling to grid cells in the surrounding 3×3 area, and employing a strategy of highest weight in the central area and decreasing weight layer by layer outwards. When the radar detects a point, its center position is precisely mapped to a grid cell in the two-dimensional grid map. Subsequently, this grid cell and the other eight grid cells in the surrounding 3×3 area are identified. According to a preset weighting rule, the central grid cell is assigned the highest weight, while its adjacent grid cells are assigned lower weights, and the outermost grid cells have the lowest weights. These weights, combined with point information, are used to update the single-frame grid labeling information of these grid cells. This refined spatial weighted labeling method allows the contribution of each point to be reflected in the grid map in a locally concentrated and gradually decaying manner, thus more accurately reflecting the actual spatial distribution and intensity of the points. This provides a more accurate and reliable input for subsequent multi-frame temporal fusion updates of the noise region map, helping to construct a more stable noise region map that better reflects the actual noise distribution.

[0036] This application further proposes a specific calculation formula for the "weighted fusion" in step four: New cumulative grid marker information = historical cumulative grid marker information × (1-K) + current frame single-frame grid marker information × K.

[0037] The core of this calculation formula lies in clarifying the relationship between historical accumulated grid marker information, current frame single-frame grid marker information, and single-frame grid weight coefficient K through mathematical expressions, thereby achieving dynamic updates of the noise region map. Here, "new accumulated grid marker information" refers to the grid marker data obtained after this fusion calculation, used to characterize the noise distribution state at the current moment; it will serve as the "historical accumulated grid marker information" for the next frame processing. "Historical accumulated grid marker information" refers to the grid marker data accumulated through multi-frame temporal fusion before the current frame processing, reflecting the continuous distribution of noise over a past period. "Current frame single-frame grid marker information" refers to the instantaneous information reflecting the noise distribution of the current frame, generated after processing the radar single-frame spot data in step three and performing spatial weighting and marking. "Single-frame grid weight coefficient K" is a preset global parameter, typically ranging from 0 to 1. This coefficient is used to adjust the weight of the current frame single-frame grid marker information during the fusion process. For example, a larger K value indicates that the information from the current frame has a greater impact on the new accumulated grid marking information, making the noise region map more sensitive to instantaneous noise changes; a smaller K value indicates that historical accumulated information dominates, making the noise region map update smoother and more stable. This coefficient can be set according to factors such as the dynamic changes in the radar's environment and the real-time requirements for noise map updates.

[0038] The proposed solution employs a weighted fusion calculation formula, enabling quantification and controllability in the process of updating the noise region map through multi-frame temporal fusion. This formula linearly weights and combines historical accumulated grid marker information with the current frame's single-frame grid marker information, where the single-frame grid weight coefficient K serves as an adjustment factor. Specifically, historical accumulated grid marker information is multiplied by a weight of (1-K), while the current frame's single-frame grid marker information is multiplied by a weight of K. This design allows the new accumulated grid marker information to simultaneously consider the stability of historical data and the real-time nature of current data. When the radar system receives new single-frame point data and generates single-frame grid marker information, this information does not completely overwrite the original historical accumulated grid marker information but is integrated with a weight of K. Simultaneously, the historical accumulated grid marker information is not static but continues to influence the new accumulated results with a weight of (1-K). In this way, the noise region map update process forms a smooth iterative mechanism, avoiding drastic changes in the noise region map caused by excessive fluctuations in single-frame data and preventing the noise region map from being slow to respond to environmental changes. This weighted fusion mechanism ensures that the noise region map can continuously, stably, and effectively reflect the true distribution of noise within the radar detection area, providing a more reliable basis for subsequent target quality calculation and differential threshold screening.

[0039] The above technical solution clarifies the specific calculation method for updating the noise region map through multi-frame temporal fusion, solving the technical problem of balancing the influence of historical data and current observation data during the noise region map update process. This weighted fusion formula ensures that the updated noise region map has both the ability to respond to instantaneous noise changes and maintains the stability of long-term noise distribution. This avoids unstable jumps in the noise region map caused by fluctuations in single-frame data and prevents lag in response to environmental changes, thus enabling the noise region map to more accurately and robustly reflect the continuous noise distribution within the radar detection area. Ultimately, this provides more accurate and reliable noise background information for target quality calculation and differential threshold screening in subsequent steps, significantly improving the accuracy and reliability of target processing by X-band radar in complex noise environments.

[0040] This application further proposes the following specific formula for calculating a comprehensive quality factor (Q) in step five: Q = [(number of matching state trajectory points - number of predicted state trajectory points) / total trajectory length] * P + [(1-P) * total trajectory length / maximum allowed trajectory length], where P is a preset weighting coefficient.

[0041] The comprehensive quality factor Q is a quantitative indicator used to evaluate the reliability and stability of a tracked target's trajectory. Its purpose is to integrate multiple dimensions of information about the target trajectory (such as trajectory continuity, prediction accuracy, trajectory duration, etc.) into a single value for subsequent target screening and classification. Through this factor, the system can prioritize targets and distinguish between real targets and false targets or clutter. The number of matched trajectory points refers to the number of points where the trajectory successfully correlates (matches) with radar markers during target tracking. These points represent reliable evidence that the target has actually been detected by the radar and are important indicators of trajectory stability and authenticity. A higher number of matched points generally means a more stable trajectory and a greater probability of the target's existence. The number of predicted trajectory points refers to the number of points where the trajectory failed to successfully correlate with radar markers during target tracking, but the system maintains its existence through prediction algorithms. These points typically occur when the target is obscured, the signal is lost, or the target is in a weak signal area. An excessive number of predicted points may indicate a decrease in trajectory reliability, suggesting the target may be a false target or about to be lost. The total trajectory length refers to the time or number of frames elapsed from the first time the target was tracked to the current moment. It reflects the duration of a target's presence within the detection area. A longer total trajectory length generally indicates higher target stability, but its reliability may be compromised if it contains too many prediction points. The maximum permissible trajectory length is a preset parameter used to normalize the total trajectory length. By comparing the total trajectory length with the maximum permissible trajectory length, the proportion of the current trajectory's duration to the system's longest acceptable duration can be assessed, thus introducing an upper limit in the quality factor calculation to prevent excessively long trajectories from having an undue impact on the quality factor. The preset weighting coefficient P is a value between 0 and 1 used to balance the relative importance of the two main components in the comprehensive quality factor Q calculation formula (i.e., the term based on the number of matched / predicted points and the term based on the total trajectory length). By adjusting the value of P, the evaluation bias of the quality factor can be flexibly adjusted according to the actual application scenario and the emphasis on target characteristics; for example, placing more emphasis on trajectory stability or trajectory persistence.

[0042] This formula constructs a quantitative index that comprehensively reflects the reliability of a target trajectory. Specifically, the first part of the formula, `[(number of matched state trajectory points - number of predicted state trajectory points) / total trajectory length] * P`, focuses on evaluating the "purity" or "stability" of the trajectory. By calculating the ratio of the difference between the number of matched points and the number of predicted points to the total trajectory length, the continuity of the trajectory and the reliability of its actual detection can be effectively measured. The more matched points and the fewer predicted points, the larger this ratio, indicating a more stable trajectory. The second part, `[(1-P) * total trajectory length / maximum allowed trajectory length]`, focuses on evaluating the "persistence" or "lifecycle" of the trajectory. The ratio of the total trajectory length to the maximum allowed trajectory length reflects the length of time the target has been tracked; a longer duration usually means higher accuracy. The preset weighting coefficient P plays a crucial role here, allowing the system to dynamically adjust the weight of stability and persistence in the overall quality assessment according to actual needs. For example, in environments with a lot of clutter, a higher P value may be needed to emphasize the stability of the trajectory; while in scenarios where the target may briefly disappear and reappear, a lower P value may be more appropriate. By using this weighted combination, the formula can overcome the limitations of a single indicator and provide a more comprehensive and robust target quality assessment, thereby more accurately distinguishing real targets from clutter in subsequent differential threshold screening, especially in the X-band radar noise region.

[0043] By employing the aforementioned technical solution, and through a weighted combination of the number of matching state trajectory points, the number of predicted state trajectory points, and the total trajectory length, along with normalization using a maximum permissible trajectory length, the problem of accurately assessing target reliability using simple trajectory information in complex noise environments is effectively solved. Specifically, this formula dynamically balances the stability and persistence of the trajectory, allowing the system to consider not only the reliability of actual target detection but also the duration of target existence when calculating target quality. The introduction of a preset weighting coefficient P further enhances the method's flexibility, allowing adjustments to the emphasis on different trajectory attributes based on varying radar operating environments and target characteristics. This enables more accurate differentiation between real targets and clutter in subsequent differential threshold screening and target output steps, significantly reducing false alarm and false negative rates, especially within the X-band radar noise region, thereby improving the overall performance and reliability of target processing.

[0044] In other embodiments, this application proposes a target processing method for X-band radar noise regions. This method distinguishes between clutter and non-clutter regions where the target is located and filters the target quality factor Q using a target quality threshold MN for clutter regions and a target quality threshold MT for non-clutter regions, respectively. However, in practical applications, if there is no clear relative relationship between these two thresholds, it may lead to a high misclassification rate of real targets in clutter regions or insufficient suppression of false targets in non-clutter regions, thereby affecting the overall target processing performance and reliability of the radar system.

[0045] In this regard, this application further proposes that the target quality threshold MN in the clutter region is greater than the target quality threshold MT in the non-clutter region, i.e., MN > MT.

[0046] The clutter region target quality threshold MN is used to screen targets identified as being located within the clutter region. Its purpose is to improve the "anti-interference capability" of the radar system in the clutter region. In this region, the overall target quality factor Q needs to reach a relatively high level to be considered a genuine target, thus avoiding misjudging genuine targets as false targets due to clutter interference; that is, the MN value should be relatively large. This threshold can be dynamically adjusted based on the radar system's performance under different clutter environments, as well as the required false alarm rate and missed alarm rate. For example, an initial value can be set through historical data analysis or expert experience, and then fine-tuned based on actual performance during system operation.

[0047] To improve the detection sensitivity for target quality screening in non-clutter regions, a target's overall quality factor Q can be considered a true target with a relatively low MN value, i.e., MN > MT. This threshold can also be set according to system performance requirements, environmental characteristics, and other factors. For example, it can be set to a fixed value or adaptively adjusted according to the environmental noise level.

[0048] The setting of MN being greater than MT clarifies the difference in target quality requirements between cluttered and non-cluttered regions. This means that in cluttered regions, target quality screening is more stringent, emphasizing quality selection and improving anti-interference capabilities; while in non-cluttered regions, the requirements for target quality are relatively lower, aiming to improve detection sensitivity. This differentiated approach is based on the different environmental characteristics of cluttered and non-cluttered regions, and aims to optimize target detection performance in different areas.

[0049] In the target processing method for X-band radar noise regions, this application first updates the noise region map through multi-frame time fusion and, combined with a noise region judgment threshold T, determines whether the region where each tracked target is located is cluttered or non-cluttered. Subsequently, the comprehensive quality factor Q is calculated for each target. Based on this, this application further clarifies the relationship between the target quality threshold MN in cluttered regions and the target quality threshold MT in non-cluttered regions as MN < MT. This means that when a target is determined to be located in a cluttered region, the minimum requirement (MN) for its comprehensive quality factor Q to pass screening is lower than the minimum requirement (MT) required when the target is located in a non-cluttered region. This setting is based on a deep understanding of the radar detection environment: in cluttered regions, due to complex environmental interference, the traces of real targets may be unstable, resulting in a relatively low comprehensive quality factor Q. If the same or higher screening criteria are still used in this region as in non-cluttered regions, a large number of real targets may be misjudged as false targets and filtered out. By reducing the quality threshold MN in cluttered regions, this method can more effectively retain real targets in complex cluttered backgrounds. In non-clutter regions, the environment is relatively clean, and target trace quality is typically high. Using a higher quality threshold (MT) in this case allows for more rigorous screening of high-quality real targets, while effectively suppressing occasional false traces or low-quality tracking trajectories. Therefore, by setting MN < MT, this method achieves adaptive adjustment of the target screening strategy for different regions. This improves the detection rate of real targets in clutter regions while maintaining high target confidence in non-clutter regions, thereby enhancing the overall accuracy and reliability of radar target processing.

[0050] The above technical solution clarifies that the target quality threshold MN in cluttered areas is greater than the target quality threshold MT in non-cluttered areas. This setting allows for a more relaxed requirement for target quality in areas with severe clutter interference, effectively reducing the risk of real targets being misidentified as false targets due to clutter and improving the detection capability of real targets against clutter backgrounds. Simultaneously, in relatively clear non-cluttered areas, the requirement for target quality is more stringent, more effectively suppressing false targets and low-quality trajectories, ensuring the confidence level of the output target. This differentiated threshold screening strategy enables the radar system to adaptively adjust the screening criteria according to the actual environment in which the target is located, thereby significantly improving the target processing performance of X-band radar in complex noise areas, enhancing the accuracy of target identification, and improving the overall reliability of the system.

[0051] This application further proposes that in step one above, the dimensions of the two-dimensional grid map correspond to the azimuth dimension and range dimension of the radar, with the azimuth dimension ranging from 0 to 4095 and the range dimension covering the effective detection range of the radar.

[0052] Specifically, the dimensions of the two-dimensional grid map correspond to the azimuth and range dimensions of the radar, meaning that the two axes of the grid map are directly mapped to the basic measurement dimensions of the radar system, namely azimuth and range. This correspondence ensures that each cell in the grid accurately represents a specific physical region in the radar detection space, thus providing an accurate spatial basis for subsequent point marking and noise region map construction. For example, rows of the grid can correspond to different range cells, while columns can correspond to different azimuth sectors. The azimuth dimension ranges from 0 to 4095, which clearly defines the degree of discretization and coverage in the azimuth dimension. The azimuth dimension typically refers to the horizontal angle, and setting its range to 0 to 4095 indicates that the entire azimuth detection range of the radar (e.g., 360 degrees) is divided into 4096 discrete azimuth cells. This ensures fine and comprehensive coverage of the radar detection space in the azimuth dimension, avoiding information loss or ambiguity due to insufficient azimuth resolution. For example, these 4096 values ​​can correspond to the digital values ​​output by the radar azimuth encoder, with each value representing a fixed angle increment. The range dimension covers the radar's effective detection range, thus defining the coverage area in the range dimension. The range dimension refers to the radial distance from the radar antenna to the target. The effective detection range is the limit at which the radar can reliably detect a target. By covering this effective detection range in the range dimension, it ensures that the grid map includes all targets and noise sources that the radar can potentially detect, avoiding the omission of important information due to insufficient range coverage. For example, the range dimension can be divided into a series of equally spaced range gates, extending from the radar's minimum detection range to its maximum effective detection range.

[0053] The proposed solution precisely maps the dimensions of the two-dimensional mesh map to the azimuth and range dimensions of the radar, explicitly defining the azimuth dimension as ranging from 0 to 4095 and the range dimension as covering the radar's effective detection range. This ensures the constructed mesh map closely matches the radar's actual detection characteristics. This precise dimensional definition guarantees that the mesh cells accurately represent physical regions in the radar's detection space. Consequently, in step three, when spatially weighting and labeling single-frame points, the point information can be accurately mapped to the corresponding mesh cells and their surrounding areas. Subsequently, in step four, when updating the noise region map through multi-frame temporal fusion, the accumulated mesh labeling information, based on this precisely constructed mesh, can more realistically and precisely reflect the continuous noise distribution. This optimized mesh construction method provides a solid foundation for subsequent steps five (target quality calculation) and six (differential threshold screening), enabling the system to more accurately determine the target's environment and apply more appropriate screening strategies.

[0054] Through the aforementioned technical solution, the dimensions of the two-dimensional mesh map are precisely defined, ensuring a high degree of consistency with the radar's actual detection capabilities and data structure. This significantly improves the efficiency and accuracy of mesh map construction, avoiding wasted computational resources or distorted detection information caused by improper mesh partitioning. This precise mesh foundation makes subsequent point marking, multi-frame fusion, and noise region map updates more accurate and reliable. Ultimately, in the target selection stage, the system can more accurately determine the target's environment based on a more precise noise region map, thereby applying more refined thresholds for target selection, effectively improving the accuracy and reliability of X-band radar target processing in complex noise environments.

[0055] Secondly, embodiments of this application provide a target processing device for X-band radar noise areas based on an optimal transmission algorithm. The target processing device for X-band radar noise areas based on the optimal transmission algorithm can be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0056] Reference Figure 2 , Figure 2 This is a schematic diagram of the hardware structure of a target processing device for X-band radar noise regions based on an optimal transmission algorithm, as described in an embodiment of this application. In this embodiment, the target processing device for X-band radar noise regions based on an optimal transmission algorithm may include a processor, a memory, a communication interface, and a communication bus.

[0057] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interfaces. The communication interfaces include input / output (I / O) interfaces, physical interfaces, and logical interfaces. These interfaces are used for interconnecting devices within the target processing equipment operating in X-band radar noise areas based on optimal transmission algorithms, as well as for interconnecting the target processing equipment operating in X-band radar noise areas with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0058] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0059] The processor can be a general-purpose processor, which can call a target processing program for X-band radar noise regions based on an optimal transmission algorithm stored in memory and execute the target processing method for X-band radar noise regions provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the target processing program for X-band radar noise regions based on the optimal transmission algorithm is called can refer to the various embodiments of the target processing method for X-band radar noise regions in this application, and will not be repeated here.

[0060] Fourthly, embodiments of this application also provide a storage medium.

[0061] This application stores a target processing program for X-band radar noise areas based on an optimal transmission algorithm on its storage medium, wherein... When the target processing program for the X-band radar noise region based on the optimal transmission algorithm is executed by the processor, it implements the steps of the target processing method for the X-band radar noise region as described above.

[0062] The method implemented when the target processing program for the X-band radar noise region based on the optimal transmission algorithm is executed can be referred to in various embodiments of the target processing method for the X-band radar noise region of this application, and will not be repeated here.

[0063] It should be noted that the sequence numbers of the embodiments in this application are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not represent a sequential order, nor do they limit "first," "second," and "third" to different types.

[0064] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0065] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0066] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of operation 5 is only used to distinguish the different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0067] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0068] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A target processing method for X-band radar noise regions, characterized in that, It includes the following steps: Step 1: System initialization; Set global parameters including single-frame grid weight coefficient (K), noise region judgment threshold (T), clutter region target quality threshold (MN), and non-clutter region target quality threshold (MT); Step 2: Grid cell construction; Construct a two-dimensional grid map covering the entire radar detection range to record the cumulative value of the point markers in each grid cell; Step 3: Single-frame point data processing and spatial weighted labeling; Process radar single-frame point data, calculate the center position (distance R and azimuth B) of each point, and use the center grid cell of the point as the core to perform weighted labeling on the grid cells within a predetermined range around it to generate single-frame grid labeling information. Step 4: Multi-frame temporal fusion update of the noise region map; The single-frame grid labeling information of the current frame and the historically accumulated grid labeling information are weighted and fused according to the single-frame grid weighting coefficient (K) to update the multi-frame accumulated grid labeling information that reflects the continuous noise distribution, i.e., the noise region map; Step 5: Target quality calculation; For each tracked target, calculate a comprehensive quality factor (Q) based on its historical trajectory information, which includes the number of trajectory points in the matching state, the number of trajectory points in the prediction state, and the total trajectory length. Step Six: Differentiated Threshold Filtering and Target Output: For each tracked target, query the noise region map based on its current location. If the cumulative grid marker value at that location is greater than the noise region judgment threshold (T), the target is determined to be located within the clutter region, and the target quality factor (Q) is filtered using the clutter region target quality threshold (MN). If the cumulative grid marker value at that location is less than or equal to the noise region judgment threshold (T), the target is determined to be located within the non-clutter region, and the target quality factor (Q) is filtered using the non-clutter region target quality threshold (MT). Targets that reach the corresponding quality threshold are output as true targets.

2. The target processing method for X-band radar noise region according to claim 1, characterized in that, The step three, "weighting the grid cells within a predetermined range around it with different weights", specifically means: taking the grid cell at the center of the dot as the core, marking the grid cells in the surrounding 3×3 area, with the central area having the highest weight and decreasing layer by layer towards the outside.

3. The target processing method for X-band radar noise areas as described in claim 1, characterized in that, The specific calculation formula for "weighted fusion" in step four is: new cumulative grid marker information = historical cumulative grid marker information × (1-K) + current frame single-frame grid marker information × K.

4. The target processing method for X-band radar noise areas as described in claim 1, characterized in that, The specific calculation formula for "calculating a comprehensive quality factor (Q)" in step five is: Q=[(number of matching state trajectory points - number of predicted state trajectory points) / total trajectory length]*P+[(1-P)*total trajectory length / maximum allowed trajectory length], where P is a preset weighting coefficient.

5. The target processing method for X-band radar noise areas as described in claim 1, characterized in that, The target quality threshold (MN) in the clutter region is less than the target quality threshold (MT) in the non-clutter region, i.e., MN < MT.

6. The target processing method for X-band radar noise areas as described in claim 1, characterized in that, In step one, the dimensions of the "two-dimensional grid map" correspond to the azimuth and range dimensions of the radar. The azimuth dimension ranges from 0 to 4095, and the range dimension covers the effective detection range of the radar.

7. A target processing device for X-band radar noise areas, characterized in that, The target processing device for the X-band radar noise region includes a processor, a memory, and a target processing program for the X-band radar noise region stored in the memory and executable by the processor, wherein when the target processing program for the X-band radar noise region is executed by the processor, it implements the steps of the target processing method for the X-band radar noise region as described in any one of claims 1-6.

8. A storage medium, characterized in that, The storage medium stores a target processing program for the X-band radar noise region, wherein when the X-band radar noise region target processing program is executed by the processor, the steps of the target processing method for the X-band radar noise region as described in claims 1 to 6 are implemented.

Images