Radar weak target detection and tracking method and system based on environment self-adaptation

Abstract

The present application relates to the technical field of radar signal processing, in particular to a radar weak and small target detection and tracking method and system based on environment self-adaption, which is applied to a networked radar system and includes the following steps: obtaining original echo data of each radar node in the networked radar, respectively calculating single-radar heterogeneous clutter severity and distance unit atmospheric attenuation severity of each radar node and performing double-domain coupling to obtain a global coupling severity index representing the degree of composite environmental interference; calculating a global regulation weight for regulating and controlling subsequent processing procedures, performing target screening on the radar echo to obtain candidate tracks; performing environment consistency verification on the candidate tracks, performing multi-radar cooperative interpolation repair and motion feature association after the verification to obtain associated tracks; respectively calculating dynamic observation noise covariance matrices of observation noise of three state variables of distance, speed and angle; iteratively updating the associated tracks based on the dynamic observation noise covariance matrices, and outputting target tracking tracks.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing technology, and specifically to a radar weak target detection and tracking method and system based on environment adaptation. Background Technology

[0002] Nearshore networked radar systems consist of shore-based radars, a primary nearshore radar, and a secondary offshore radar network. They are primarily used for all-weather surveillance and detection of small targets with small radar cross-sections and extremely low echo signal-to-noise ratios, such as small vessels and low-altitude, low-speed drones. These systems are widely applied in nearshore security, maritime rights protection, and maritime traffic control. However, nearshore waters experience both heterogeneous sea clutter interference and range-dependent atmospheric attenuation, leading to coupling between these two types of interference. Existing traditional networked radar systems for detecting and tracking small targets suffer from five major shortcomings: Existing technologies only quantify sea clutter intensity or electromagnetic wave atmospheric loss separately, without establishing a dual-domain coupled quantization model of clutter-atmosphere attenuation. This makes it impossible to simultaneously characterize the differences in local clutter floor at radar sites and the differences in atmospheric attenuation at long-range range units. Harsh environment assessments only reflect single interferences, and the quantification results deviate significantly from the actual degree of composite interference. Subsequent adaptive processing lacks accurate environmental basis.

[0003] Traditional multi-radar mutual calibration uses fixed correction coefficients that are not dynamically adjusted according to the severity of the real-time marine environment. Differences in clutter substrates and atmospheric attenuation between shore-based, near-shore, and offshore radars cannot be eliminated in a balanced manner. There are systematic deviations in the echo amplitudes of different radar units at the same distance, resulting in inconsistent benchmarks for the generated point data, which directly reduces the accuracy of multi-radar collaborative detection and correlation matching.

[0004] Traditional constant false alarm detection thresholds and long-term coherent phase compensation iterations are global fixed parameters. In harsh environments such as sea fog, rain, and high sea states, if the fixed threshold is too low, it will generate a large number of clutter false alarms. If the threshold is raised, it will directly overwhelm the echoes of weak targets, resulting in serious missed detections. The number of coherent compensation iterations does not adaptively adjust with the environment. In harsh environments, the target energy accumulation is insufficient, and the signal-to-noise ratio of weak targets cannot be effectively improved.

[0005] Existing track break interpolation relies solely on target motion models to fill in missing states, without introducing an environmental consistency verification mechanism. Under severe clutter, a large number of false points are easily correlated with historical tracks, generating a large number of false tracks without corresponding real targets. At the same time, there is no multi-radar collaborative interpolation mechanism, and after a single radar loses track, it is impossible to use observation data from other networked radars to repair the track, significantly shortening the continuous tracking time.

[0006] Traditional Kalman-type tracking filters use a factory-preset fixed observation noise covariance matrix, which is not dynamically adjusted according to the severity of the real-time environment. In areas with severe clutter and atmospheric attenuation, the reliability of radar measurement data is low, but the filter still accepts the measurement with equal weight, which leads to drift in track position and velocity estimation. For small targets at long distances, the filter is prone to divergence and track loss, making it impossible to achieve long-term stable tracking. Summary of the Invention

[0007] This invention provides a radar weak target detection and tracking method based on environment adaptation. It constructs a complete integrated processing link that combines dual-domain coupled environment quantization, multi-radar dynamic mutual calibration, exponential adaptive detection, environmental constraint-coordinated track repair, and weighted adaptive filtering. The entire process from data source to tracking output reuses unified environment quantization indicators, and simultaneously solves five major technical defects of existing technologies: environment quantization distortion, fixed calibration, rigid detection parameters, easy spurious track breakage, and filter drift. Those skilled in the art can directly reproduce and implement this method based on the following complete steps, formulas, and parameter rules.

[0008] To solve the above-mentioned technical problems, the present invention adopts the following solution: An environment-adaptive radar target detection and tracking method is applied to a networked radar system consisting of shore-based radar, near-shore main radar, and far-sea auxiliary radar. The method includes the following steps: S1: Obtain the raw echo data of each radar node in the network radar, and calculate the single radar heterogeneous clutter severity A and range cell atmospheric attenuation severity B for each radar node. S2: Dual-domain coupling of the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B yields the global coupling severity index C, which characterizes the degree of interference in the complex environment. S3: Based on the global coupling severity index C, calculate the global control weight F for the subsequent processing of the system control, generate the adaptive detection threshold and the number of long-term coherent phase compensation iterations according to the global control weight F, and perform target screening on the radar echo to obtain candidate point traces K. S4: Based on the global control weight F, the candidate point track K is checked for environmental consistency. After passing the check, multi-radar collaborative interpolation repair and motion feature association are performed to obtain the associated track. S5: Based on the global control weight F, the observation noise of the three state variables of distance, velocity and angle is independently and adaptively amplified to obtain the dynamic observation noise covariance matrix; the associated trajectory is iteratively updated using the dynamic observation noise covariance matrix to output the target tracking trajectory.

[0009] Furthermore, in S2, the specific formula for dual-domain coupling of the single-radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B is as follows: C = A + B - A·B; Where A represents the heterogeneous clutter severity of a single radar, B represents the atmospheric attenuation severity of the range unit, and the values ​​of A, B, and C are all in the range of [0,1].

[0010] Furthermore, S2 also includes: using the global coupling severity index of the offshore auxiliary radar as a benchmark, performing channel-by-channel and range-by-range cell normalization calibration on the echo data of the shore-based radar and the near-shore main radar, generating standardized spot traces bound to the corresponding global coupling severity index C tag, as input for target screening in S3.

[0011] Furthermore, the calculation method for the global regulation weight F described in S3 is as follows: F = exp(-λ·C); Where λ is a preset positive constant, C is the global coupling severity index, and F takes values ​​in the range of (0,1]. The adaptive detection threshold is generated by dynamically adjusting the preset basic constant false alarm threshold according to the global control weight F value.

[0012] Furthermore, in S3, determining the number of iterations for long-term coherent phase compensation includes: Based on the magnitude of the global control weight F, multiple non-overlapping numerical intervals are pre-divided, and a fixed set of iterations is pre-configured for each numerical interval. According to the numerical interval where the global control weight F of the current distance unit is located, a corresponding set of fixed iterations is matched as the long-term coherent phase compensation iterations.

[0013] Furthermore, the multiple non-overlapping numerical intervals are divided into three segments: When F ∈ [0, 0.3), configure the first fixed number of iterations; When F ∈ [0.3, 0.7), configure a second fixed number of iterations; When F ∈ [0.7, 1], configure a third fixed number of iterations; The values ​​of the first, second, and third fixed iterations decrease sequentially.

[0014] Furthermore, in S4, the environmental consistency check for candidate point K includes: Extract the first global control weight bound to the current candidate point K to be matched; Extract the global control weights bound to all historical storage frames of the target trajectory, and calculate the arithmetic mean as the second global control weight; Calculate the absolute value of the difference between the first global control weight and the second global control weight; When the absolute value of the weight difference is greater than the preset threshold, the verification is deemed to fail and the candidate point is removed. When the absolute value of the weight difference is less than or equal to the preset threshold, the verification is passed, and multi-radar collaborative interpolation repair and motion feature association are allowed.

[0015] Furthermore, in S5, the calculation method for independently adaptively amplifying the observation noise of the three state variables—distance, velocity, and angle—is as follows: R adapt = R0*diag ([ 1+δ1(1-F) , 1+δ2(1-F ) , 1+δ3(1-F )]); Among them, R adapt R0 is the dynamic observation noise covariance matrix, F is the basic observation noise diagonal matrix preset by the radar system, δ1, δ2, and δ3 represent the independent noise amplification adjustment coefficients corresponding to the range state, velocity state, and angle state, respectively.

[0016] Furthermore, in S5, the iterative update of the associated track using the dynamic observation noise covariance matrix is ​​performed according to the following timing sequence: S501 performs real-time updates of global environmental parameters; S502, based on the refreshed global environment parameters, perform the reassignment of the dynamic observation noise covariance matrix; S503 utilizes the reassigned dynamic observation noise covariance matrix to synchronously update the trajectory position and velocity state values.

[0017] An environment-adaptive radar target detection and tracking system is applied to a networked radar system consisting of shore-based radar, near-shore main radar, and far-sea auxiliary radar, including: The multi-source data acquisition module is used to acquire the raw echo data of each radar node and calculate the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B respectively. The dual-domain coupling calibration module is used to couple the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B in a dual-domain manner to obtain the global coupling severity index C. The adaptive detection module is used to calculate the global control weight F based on the global coupled severity index C, and generate the adaptive detection threshold and the number of long-term coherent phase compensation iterations based on the global control weight F, and to screen the radar echoes to obtain candidate points K. The multi-radar collaborative trajectory processing module is used to perform environmental consistency verification on candidate point trajector K based on global control weight F. After verification, multi-radar collaborative interpolation repair and motion feature association are performed to obtain associated trajectories. The adaptive tracking filter module is used to independently and adaptively amplify the observation noise of the three state variables of distance, velocity and angle based on the global control weight F, so as to obtain the dynamic observation noise covariance matrix. The matrix is ​​then used to iteratively update the associated track and output the target tracking track.

[0018] The beneficial effects of this invention are: This invention overcomes the five major shortcomings of the prior art through a complete set of technical means, including dual-domain coupled environment quantization, full-link environment adaptive control, multi-radar collaborative repair, and adaptive noise filtering. In actual tests under harsh nearshore scenarios involving high sea states, heavy fog, and rainfall, compared with traditional fixed-parameter radar detection and tracking schemes, it demonstrates the following quantification technology effects: 1. Simultaneously integrating sea clutter and atmospheric attenuation interference, and uniformly outputting 0-1 range quantitative indicators, the environmental severity assessment error is reduced by 42.6% compared to a single clutter quantization scheme, providing an accurate environmental benchmark for full-process adaptive processing.

[0019] 2. By abandoning fixed correction coefficients and relying on real-time global coupling severity to calibrate echoes channel by channel, the amplitude deviation of echoes from different radar units at the same distance is reduced by 67.1%, and the success rate of multi-radar point matching is significantly improved.

[0020] 3. Automatically raise the detection threshold in harsh areas to suppress clutter, and lower the threshold in good areas to improve the detection of weak targets; segmented adaptive method increases the number of coherent compensation iterations, and the signal-to-noise ratio of weak target echoes is improved by an average of 2.8dB; measured data: the detection rate of weak targets is improved by 27.3%, and the number of clutter false alarms is reduced by 61.5%, which solves the contradiction of the traditional method of missing detection and false alarms.

[0021] 4. Redundant observation by multiple radars compensates for the interruption of flight status, and environmental verification filters out false points due to environmental mismatch; the frequency of measured track breakage decreased by 58.2%, the number of false track generation decreased by 73.4%, and the average duration of continuous target tracking increased by 31.2%.

[0022] 5. Automatically amplifies the observation noise weight in harsh environments, weakening the interference of low-confidence measurements on track estimation; the root mean square error of tracking position of distant weak targets is reduced by 34.7%, with no track drift or filter divergence, and can stably and continuously track weak targets for dozens of frames.

[0023] 6. The entire process of data acquisition and tracking filtering is reused, and the environmental parameters are not recalculated. The single-frame processing latency of the entire algorithm is less than 1ms, which is compatible with the embedded hardware platform for real-time radar signal processing and meets the practical needs of real-time maritime surveillance. Attached Figure Description

[0024] Figure 1This is a flowchart of the radar weak target detection and tracking method based on environment adaptation according to the present invention; Figure 2 This is a flowchart of the encapsulation process for data acquisition and severity calculation in the embodiment; Figure 3 This is a flowchart of the S2 dual-domain coupling calibration and standardized dot generation process in the embodiment; Figure 4 This is the main flowchart for calculating the target detection threshold and weight in the S3 example. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the present invention or its application or use. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of the invention.

[0027] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0028] Furthermore, for clarity and brevity, descriptions of well-known structures, functions, and configurations may have been omitted. Those skilled in the art will recognize that various changes and modifications can be made to the examples described herein without departing from the spirit and scope of this disclosure.

[0029] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0030] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0031] The present invention will now be described in detail with reference to the accompanying drawings and embodiments: Example 1 This invention presents a complete processing method for radar weak target detection and tracking based on environmental adaptation, comprising five consecutive steps: S1 multi-source synchronous data acquisition and dual severity normalization; S2 dual-domain coupled environmental quantization + cross-radar dynamic mutual calibration of the offshore auxiliary radar benchmark; S3 exponential global control weight solution; adaptive detection threshold and three-segment coherent compensation iterative configuration; S4 track break identification based on control weight; multi-radar collaborative interpolation repair; environmental consistency verification; motion feature association; and S5 weight-linked adaptive noise matrix dynamic filtering and track iterative update.

[0032] Specifically, S1 synchronously collects raw echo data, airspace meteorological data, and electromagnetic wave propagation parameters from all channels of shore-based radar, near-shore main radar, and far-sea auxiliary radar; calculates the heterogeneous clutter severity of a single radar and the atmospheric attenuation severity of the range unit, and encapsulates them into raw input datasets. For the network numbered For the radar, extract the maximum clutter floor intensity of all range channels of the radar in the current frame. Minimum value , and the maximum amplitude fluctuation of the full-channel echo signal Minimum value A weighted formula for calculating the severity of heterogeneous clutter in a single radar is constructed, using clutter floor and signal fluctuation as two types of interference as weighting factors: ; In the formula For radar Real-time clutter floor observations For radar Real-time echo fluctuation amplitude observation; , The weighting coefficients corresponding to the clutter floor and signal fluctuations satisfy the following conditions: After calculation, the results are truncated and constrained to the [0,1] interval to obtain the normalized single-radar heterogeneous clutter severity. .

[0033] For radar detection paths numbered as Extract the real-time observation value of electromagnetic wave transmission loss for the distance cell. Extreme loss values ​​within the entire detection range and Simultaneously, four meteorological attenuation weighting factors—rainfall, fog, aerosols, and humidity—are introduced to construct a weighted calculation formula for atmospheric attenuation severity per distance unit: ; In the formula , , , Distance unit Corresponding to normalized meteorological attenuation of rainfall, fog, aerosols, and air humidity; The transmission loss weighting factor is used. , , , The weighting coefficients for the four meteorological categories satisfy the following conditions: The final calculation results are uniformly mapped to the [0,1] interval to obtain the atmospheric attenuation severity of the normalized distance unit. .

[0034] In S1, the collected raw echo data from all channels completely covers all range channels and all azimuth channels of the radar configuration. Each set of raw echo data completely stores three independent signal values: echo amplitude, Doppler frequency, and range delay. The airspace meteorological data includes four independent meteorological materials: rainfall observation values, fog concentration values, aerosol concentration values, and air humidity values, which are collected synchronously in the detection airspace. The electromagnetic wave propagation parameters are fixed constant values ​​pre-calibrated for electromagnetic wave spatial loss calculation. All data materials are synchronously collected into the original input dataset according to a unified frame time sequence. In S1, the conversion range of the heterogeneous clutter severity of a single radar is jointly limited by the extreme values ​​of the clutter base intensity and the extreme values ​​of the signal fluctuation amplitude of all radars in the same frame period. The conversion range of the atmospheric attenuation severity of the range unit is jointly limited by the extreme values ​​of the electromagnetic wave transmission loss of all range units within the radar's farthest detection range. The lower bound of the normalized conversion range of the heterogeneous clutter severity of a single radar and the atmospheric attenuation severity of the range unit is uniformly set to 0, and the upper bound is uniformly set to 1. All converted numerical data are synchronously entered into the corresponding storage partition inside the original input dataset.

[0035] S2 receives the original input dataset, runs a dual-domain coupled environment quantization algorithm to couple the heterogeneous clutter severity of a single radar and the atmospheric attenuation severity of the range cell, and obtains a global coupled severity index; based on the global coupled severity index, a cross-radar mutual calibration operator is constructed to normalize and calibrate all radar echoes, and generate standardized dots for coupled severity labels. In S2, the mathematical expression corresponding to the dual-domain coupled environment quantization algorithm is: ; in, The device number representing the networked radar; Represents the radar detection range unit number; Representative number Radar, number The global coupling severity index corresponding to the distance unit has a value range of [value missing]. ; Representative number The single-radar heterogeneous clutter severity of the radar is categorized into several ranges. ; Representative number The atmospheric attenuation severity of the distance cell is defined as the distance cell's atmospheric attenuation severity, with a value range of [0, 1]. In S2, when constructing the cross-radar mutual calibration operator, the global coupling severity index of the corresponding location of the offshore auxiliary radar is selected as a unified reference benchmark. The normalization calibration operation is applied to each set of original echo data output by the shore-based radar and the near-shore main radar channel by channel and range unit by range unit. Each set of echo observation data that has completed the normalization calibration process is individually bound to the global coupling severity index that matches the corresponding radar number and the corresponding range unit number. All the observation data with bound labels are uniformly collected to generate a standardized point trace dataset.

[0036] S3: Receive standardized point traces, run an exponential environment adaptive weighting algorithm, and calculate the global control weight using a global coupled severity index; generate an adaptive detection threshold and a long-term coherent phase compensation iteration number based on the global control weight; wherein, the long-term coherent phase compensation iteration number is divided into three non-overlapping numerical intervals according to the numerical boundary of the global control weight; filter out signals exceeding the threshold, and generate candidate point traces with control weight labels; In S3, the mathematical expression corresponding to the exponential environment adaptive weight algorithm is: ; in, The device number representing the networked radar; Represents the radar detection range unit number; Representative number Radar, number The global control weight corresponding to the distance unit has a value range of [value missing]. Represents the severity index of global coupling; Represents the scene adjustment constant, and ; In step S3, the calculation basis of the adaptive detection threshold is the basic constant false alarm rate (CFAR) detection benchmark threshold pre-stored by the system, and each distance unit is matched with a set of independent adaptive detection threshold values. The number of long-term coherent phase compensation iterations is divided into three non-overlapping numerical intervals according to the numerical boundaries of the global control weight. The specific division rules are as follows: the division boundaries of the three global control weight numerical intervals are 0.3 and 0.7, respectively; the global control weight values ​​in the interval of 0 to 0.3 are matched with the first set of fixed iteration values, the global control weight values ​​in the interval of 0.3 to 0.7 are matched with the second set of fixed iteration values, and the global control weight values ​​in the interval of 0.7 to 1 are matched with the third set of fixed iteration values. The three sets of fixed iteration values ​​are set in descending order according to the increasing order of the interval boundary values, and the matching relationship between each set of fixed iteration values ​​and the corresponding interval, as well as the specific size of the iteration values, are pre-stored in the computation storage partition.

[0037] S4 receives candidate points and uses global control weights to identify track breaks, perform multi-radar collaborative interpolation repair, and verify environmental consistency; it then performs motion feature association matching on the verified points to obtain associated tracks and corresponding control weights. In S4, the mathematical expression corresponding to the weighted adaptive noise matrix algorithm is: ; in, The device number representing the networked radar; Represents the radar detection range unit number; Representative number Radar, number The dynamic observation noise covariance matrix corresponding to the isolated unit; This represents the basic observation noise diagonal matrix preset by the radar system; Represents the overall regulatory weight; These represent the noise amplification adjustment constants corresponding to the distance state quantity, velocity state quantity, and angle state quantity, respectively, and all three are non-negative real numbers; This means constructing a diagonal matrix from the vector elements; In S4, the determination criterion for track breakage identification is that the initialized stable track in the system cannot match the candidate track material for a fixed number of consecutive frames; the multi-radar collaborative interpolation repair operation retrieves all candidate track materials output by other network radars in the same airspace range and the same distance interval to participate in the interpolation numerical calculation, and the interpolation calculation output data fills in the missing position, velocity and angle values ​​inside the track. In S4, the operation process of environmental consistency verification includes three fixed steps: the first step is to extract the global control weight value bound to a single candidate point trace to be matched; the second step is to extract the global control weight values ​​bound to all historical storage frames of the target trajectory and calculate the arithmetic mean; the third step is to calculate the absolute value of the difference between the two sets of values. A preset fixed threshold is stored separately in the verification operation partition. Candidate points with an absolute value of the difference greater than the preset fixed threshold are directly eliminated. Only candidate points with an absolute value of the difference less than or equal to the preset fixed threshold enter the motion feature association matching process.

[0038] S5 receives the associated trajectory and corresponding control weights, runs the weight-linked adaptive noise matrix algorithm, calculates the dynamic observation noise covariance matrix using the corresponding control weights, iteratively updates the trajectory using the dynamic observation noise covariance matrix, and outputs the target tracking trajectory. In S5, the dynamic observation noise covariance matrix completely replaces the fixed numerical observation noise matrix in the storage partition and participates in each round of track iteration update calculation. Before each round of iteration calculation starts, the current frame's global coupled severity index and global control weight are read synchronously. The two sets of numerical materials are input synchronously into the weight linkage adaptive noise matrix algorithm to complete the reassignment of the dynamic observation noise covariance matrix.

[0039] In S5, each round of track iteration update in the tracking filtering stage synchronously executes three sequentially arranged operations. The first operation is to refresh the real-time values ​​of the global environmental parameters, the second operation is to reassign the values ​​of the dynamic observation noise covariance matrix, and the third operation is to synchronously iterate and calculate the track position status values ​​and track velocity status values. After the three operations are completed, the complete target tracking track material of the current frame is output.

[0040] In harsh nearshore environments characterized by high sea states, heavy fog, and rainfall, the method of this invention was used to conduct continuous 300-frame tracking tests. Compared with traditional fixed-parameter schemes, the detection rate of weak targets increased by 27.3%, the number of false alarms due to clutter decreased by 61.5%, the frequency of track breakage decreased by 58.2%, the number of false tracks decreased by 73.4%, the root mean square error of target tracking position decreased by 34.7%, there was no track drift or filter divergence, and the continuous and stable target tracking time exceeded 300 frames.

[0041] Example 2 This embodiment proposes a radar weak target detection and tracking signal processing system for implementing the method of the present invention. It consists of a multi-source data acquisition module, a dual-domain coupling calibration module, an adaptive detection module, a multi-radar collaborative trajectory processing module, an adaptive tracking filtering module, and a parameter storage partition collaborative system. The data of each module can be interconnected at high speed, realizing full-process environmental adaptive weak target detection and tracking.

[0042] Specifically, the environment-adaptive radar weak target detection and tracking system is applied to a networked radar system consisting of shore-based radar, near-shore main radar, and far-sea auxiliary radar, including: The multi-source data acquisition module is used to acquire the raw echo data of each radar node and calculate the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B respectively. The dual-domain coupling calibration module is used to couple the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B in a dual-domain manner to obtain the global coupling severity index C. The adaptive detection module is used to calculate the global control weight F based on the global coupled severity index C, and generate the adaptive detection threshold and the number of long-term coherent phase compensation iterations based on the global control weight F, and to screen the radar echoes to obtain candidate points K. The multi-radar collaborative trajectory processing module is used to perform environmental consistency verification on candidate point trajector K based on global control weight F. After verification, multi-radar collaborative interpolation repair and motion feature association are performed to obtain associated trajectories. The adaptive tracking filter module is used to independently and adaptively amplify the observation noise of the three state variables of distance, velocity and angle based on the global control weight F, so as to obtain the dynamic observation noise covariance matrix. The matrix is ​​then used to iteratively update the associated track and output the target tracking track.

[0043] This invention is the first to achieve dual-domain coupled quantization of sea clutter and atmospheric attenuation composite interference, and constructs an integrated environmental adaptive control link that runs through data acquisition to tracking output. It solves the industry pain points of near-shore network radar such as missed detection of weak targets, false alarms, track breaks, and tracking drift, and has extremely high engineering application value in marine security, sea area monitoring, and low-speed small target early warning scenarios.

[0044] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Based on the technical essence of the present invention, any simple modifications, equivalent substitutions, and improvements made to the above embodiments within the spirit and principles of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A radar weak target detection and tracking method based on environment adaptation, applied to a networked radar system consisting of shore-based radar, near-shore main radar, and far-sea auxiliary radar, characterized in that... Includes the following steps: S1: Obtain the raw echo data of each radar node in the network radar, and calculate the single radar heterogeneous clutter severity A and range cell atmospheric attenuation severity B for each radar node. S2: Dual-domain coupling of the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B yields the global coupling severity index C, which characterizes the degree of interference in the complex environment. S3: Based on the global coupling severity index C, calculate the global control weight F for the subsequent processing of the system control, generate the adaptive detection threshold and the number of long-term coherent phase compensation iterations according to the global control weight F, and perform target screening on the radar echo to obtain candidate point traces K. S4: Based on the global control weight F, the candidate point track K is checked for environmental consistency. After passing the check, multi-radar collaborative interpolation repair and motion feature association are performed to obtain the associated track. S5: Based on the global control weight F, the observation noise of the three state variables, distance, velocity and angle, is independently and adaptively amplified to obtain the dynamic observation noise covariance matrix. The associated track is updated iteratively using the dynamic observation noise covariance matrix to output the target tracking track.

2. The radar weak target detection and tracking method based on environment adaptation according to claim 1, characterized in that, In S2, the specific formula for dual-domain coupling of the heterogeneous clutter severity A of a single radar and the atmospheric attenuation severity B of the range unit is as follows: C = A + B - A·B; Where A represents the heterogeneous clutter severity of a single radar, B represents the atmospheric attenuation severity of the range unit, and the values ​​of A, B, and C are all in the range of [0,1].

3. The radar weak target detection and tracking method based on environment adaptation according to claim 1, characterized in that, S2 also includes: using the global coupling severity index of the offshore auxiliary radar as a benchmark, performing channel-by-channel and range-by-range cell normalization calibration on the echo data of the shore-based radar and the near-shore main radar, generating standardized spot traces bound to the corresponding global coupling severity index C tag, which serve as input for target screening in S3.

4. The radar weak target detection and tracking method based on environment adaptation according to claim 1, characterized in that, The calculation method for the global regulation weight F mentioned in S3 is as follows: F = exp(-λ·C); Where λ is a preset positive constant, C is the global coupling severity index, and F takes values ​​in the range of (0,1]. The adaptive detection threshold is generated by dynamically adjusting the preset basic constant false alarm threshold according to the global control weight F value.

5. The radar weak target detection and tracking method based on environment adaptation according to claim 1, characterized in that, In S3, determining the number of iterations for long-term coherent phase compensation includes: Based on the magnitude of the global control weight F, multiple non-overlapping numerical intervals are pre-divided, and a fixed set of iterations is pre-configured for each numerical interval. According to the numerical interval where the global control weight F of the current distance unit is located, a corresponding set of fixed iterations is matched as the long-term coherent phase compensation iterations.

6. The radar weak target detection and tracking method based on environment adaptation according to claim 5, characterized in that, Multiple non-overlapping numerical intervals are divided into three segments: When F ∈ [0, 0.3), configure the first fixed number of iterations; When F ∈ [0.3, 0.7), configure a second fixed number of iterations; When F ∈ [0.7, 1], configure a third fixed number of iterations; The values ​​of the first, second, and third fixed iterations decrease sequentially.

7. The radar weak target detection and tracking method based on environment adaptation according to claim 1, characterized in that, In S4, the environmental consistency check for candidate point K includes: Extract the first global control weight bound to the current candidate point K to be matched; Extract the global control weights bound to all historical storage frames of the target trajectory, and calculate the arithmetic mean as the second global control weight; Calculate the absolute value of the difference between the first global control weight and the second global control weight; When the absolute value of the weight difference is greater than the preset threshold, the verification is deemed to fail and the candidate point is removed. When the absolute value of the weight difference is less than or equal to the preset threshold, the verification is passed, and multi-radar collaborative interpolation repair and motion feature association are allowed.

8. The radar weak target detection and tracking method based on environment adaptation according to claim 1, characterized in that, In S5, the calculation method for independently adaptively amplifying the observation noise of the three state variables—distance, velocity, and angle—is as follows: R adapt = R 0* diag ([ 1+δ1(1-F) , 1+δ2(1-F ) , 1+δ3(1-F )]); Among them, R adapt R0 is the dynamic observation noise covariance matrix, F is the basic observation noise diagonal matrix preset by the radar system, δ1, δ2, and δ3 represent the independent noise amplification adjustment coefficients corresponding to the range state, velocity state, and angle state, respectively.

9. The radar weak target detection and tracking method based on environment adaptation according to claim 1, characterized in that, In S5, the iterative update of the associated track using the dynamic observation noise covariance matrix is ​​performed according to the following timing sequence: S501 performs real-time updates of global environmental parameters; S502, based on the refreshed global environment parameters, perform the reassignment of the dynamic observation noise covariance matrix; S503 utilizes the reassigned dynamic observation noise covariance matrix to synchronously update the trajectory position and velocity state values.

10. An environment-adaptive radar target detection and tracking system, applied to a networked radar system consisting of shore-based radar, near-shore main radar, and far-sea auxiliary radar, characterized in that... include: The multi-source data acquisition module is used to acquire the raw echo data of each radar node and calculate the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B respectively. The dual-domain coupling calibration module is used to couple the single radar heterogeneous clutter severity A and the range unit atmospheric attenuation severity B in a dual-domain manner to obtain the global coupling severity index C. The adaptive detection module is used to calculate the global control weight F based on the global coupled severity index C, and generate the adaptive detection threshold and the number of long-term coherent phase compensation iterations based on the global control weight F, and to screen the radar echoes to obtain candidate points K. The multi-radar collaborative trajectory processing module is used to perform environmental consistency verification on candidate point trajector K based on global control weight F. After verification, multi-radar collaborative interpolation repair and motion feature association are performed to obtain associated trajectories. The adaptive tracking filter module is used to independently and adaptively amplify the observation noise of the three state variables of distance, velocity and angle based on the global control weight F, so as to obtain the dynamic observation noise covariance matrix. The module then uses the dynamic observation noise covariance matrix to iteratively update the associated track and output the target tracking track.

Images