A dense target sorting method, apparatus, and medium

Abstract

The application discloses a dense target sorting method, equipment and medium, and belongs to the field of reconnaissance processing, and comprises the following steps: firstly, buffering the arriving full pulse data, using a pre-sorting algorithm to separate the buffered data in frequency or azimuth, and obtaining a pre-sorting result; then, further performing the resolution of parameter overlapping dense targets on each pre-sorting result. The application provides a dense target sorting method based on long-time statistical fine straight line detection, which can not only separate the frequency or azimuth distinguishable targets, but also resolve the parameter overlapping dense targets.

Description

Technical Field

[0001] This invention relates to the field of reconnaissance processing, and more specifically, to a method, device, and medium for dense target sorting. Background Technology

[0002] Existing sorting algorithms primarily rely on signal parameter variations, such as histogram methods, dynamic correlation algorithms, cumulative difference histogram algorithms, sequence difference histogram algorithms, and PRI transform algorithms. However, with the development of radar technology, radar parameters exhibit increasing signal complexity, diversity, and unpredictability. For example, in the time domain, there are fixed, variable, staggered, jittery, and staggered patterns; in the frequency domain, there are fixed, variable, and single-pulse agility patterns; and even parameters can exhibit no discernible pattern of change. This makes traditional processing algorithms based on electromagnetic parameter variations inadequate for resolving targets with densely overlapping parameters. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a dense target sorting method, device and medium. First, the arriving full pulse data is buffered for a long time. Using the traditional frequency and azimuth histogram pre-sorting algorithm, the separation of targets that can be distinguished by frequency or azimuth can be achieved. However, within the same pre-sorting result, there may be targets with overlapping parameters and dense distribution. Then, for each pre-sorting result, the resolution of targets with overlapping parameters and dense distribution is achieved by coarse fitting of time and azimuth lines, fine line detection and extreme value search.

[0004] The objective of this invention is achieved through the following approach:

[0005] A dense target sorting method includes the following steps:

[0006] First, the arriving full pulse data is buffered, and the buffered data is separated by frequency or azimuth using a pre-sorting algorithm to obtain the pre-sorting result;

[0007] Then, for each pre-sorting result, further resolution of densely overlapping targets with overlapping parameters is performed.

[0008] Furthermore, the pre-sorting algorithm includes a pre-sorting algorithm based on frequency and orientation histogram processing.

[0009] Furthermore, the step of buffering the arriving full-pulse data includes the sub-step of buffering the input full-pulse data for a time of T.

[0010] Furthermore, in the process of using the pre-sorting algorithm to separate targets in the cached data by frequency or orientation, the pre-sorting algorithm is specifically used to separate targets in the cached data by frequency or orientation that can be distinguished; and within the same pre-sorting result, there may be dense targets with overlapping parameters.

[0011] Furthermore, the resolution of densely overlapping targets with overlapping parameters includes the following sub-steps: resolving densely overlapping targets with overlapping parameters through coarse fitting of time-azimuth lines, fine line detection, and extreme value search.

[0012] Furthermore, the process of resolving densely overlapping targets through coarse fitting of time-azimuth lines, fine line detection, and extreme value search specifically includes the following sub-steps:

[0013] Step a: For each pre-sorting result, a linear fitting algorithm is used to obtain a coarse estimate of time-azimuth; the specific steps are as follows: Assume the arrival time of each full-pulse data is TOA. i i = 1, 2, ..., K, with orientation DOA i Let i = 1, 2, ..., K, where N represents the total number of pulses; the coarsely estimated time-azimuth line is represented as: DOA i =A·TOA i +B, where A is the slope of the line and B is the slope of the line; the orientations of all full pulses are combined into a column vector DOA. vec =[DOA1,DOA2,…,DOA] N ] T The arrival times of all full pulses are combined into a column vector TOA. vec =[TOA1,TOA2,…,TOA N ] T ,in[·] T Let A represent the transpose; then, by linear fitting, the estimated values ​​of A and B are obtained as follows:

[0014]

[0015] Where ε is a positive number, and eye(2) represents a diagonal matrix of order 2;

[0016] Step b: Based on the obtained time-azimuth coarse estimate, set the parameter range for fine line detection; assuming the azimuth resolution of the reconnaissance system is δ, obtain the pre-sorting time range in the pre-sorting step as Range. TOA The directional range is Range DOA The slope value for fine line detection is set to... Number of intercepts for fine line detection; pre-sorting azimuth range. DOA The number of [items] is related to the system's azimuth resolution δ, and the formula for calculating its number is as follows: Where round(·) represents rounding; then the intercept value for fine line detection is set to

[0017] Step c: Along the slope and intercept The straight line formed Search for all pulses falling on this line and record their number as Num. m,n Num m,n Form a 3×(N+1) matrix;

[0018] Step d: For Num m,n By padding the edges with zeros, we get

[0019]

[0020] Step e: For The search for the maximum value determines the time-azimuth line containing the target location; the formula for calculating the maximum value is:

[0021] Num m,n >Num m,n-1 &Num m,n >Num m-1,n ...

[0022] Num m,n ≥Num m,n+1 &Num m,n ≥Num m+1,n &Num m,n ≥Thres

[0023] Thres is the threshold that is set by the user;

[0024] Step f: Based on The time-azimuth line corresponding to the maximum value is used for full-pulse screening to obtain the sorting results of densely overlapping targets.

[0025] Furthermore, in step a, ε = 10 -4 Or ε = 10 -5 Or ε = 10 -6 .

[0026] Furthermore, the value range of T includes tens of seconds, and long-term statistical results can be obtained by utilizing the value range of T.

[0027] A dense target sorting device includes a processor and a memory, wherein the memory stores a computer program that, when loaded by the processor, executes the method as described in any of the preceding methods.

[0028] A computer-readable storage medium storing a computer program therein, the computer program being loaded by a processor and executing the method as described in any of the preceding claims.

[0029] The beneficial effects of this invention include:

[0030] This invention provides a dense target sorting method based on long-term statistical fine line detection. First, the arriving full pulse data is buffered for a long time. Using traditional frequency and azimuth histogram pre-sorting algorithms, targets that can be distinguished by frequency or azimuth can be separated. However, within the same pre-sorting result, there may be targets with overlapping parameters and dense distribution. Then, for each pre-sorting result, coarse fitting of time-azimuth lines, fine line detection, and extreme value search are used to distinguish dense targets with overlapping parameters. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 A block diagram of a dense target sorting method based on long-term statistical fine line detection provided in an embodiment of the present invention. Detailed Implementation

[0033] All features disclosed in all embodiments of this specification, or steps in all methods or processes implied in the disclosure, may be combined and / or extended or replaced in any way, except for mutually exclusive features and / or steps.

[0034] This invention provides a method for sorting dense targets based on long-term statistical fine line detection. First, the arriving full-pulse data is buffered for a long time. Using traditional frequency and azimuth histogram pre-sorting algorithms, targets distinguishable by frequency or azimuth can be separated. However, within the same pre-sorting result, there may be densely distributed targets with overlapping parameters. Then, for each pre-sorting result, coarse fitting of time-azimuth lines, fine line detection, and extreme value search are used to distinguish densely distributed targets with overlapping parameters. Figure 1 As shown, the specific steps are as follows:

[0035] Step 1: Buffer the input full-pulse data for a time of T. T usually needs to be on the order of ten seconds to obtain good long-term statistical results;

[0036] Step 2: Perform pre-sorting on the buffered full pulse data based on frequency and azimuth histogram processing to achieve target differentiation by frequency or azimuth. However, within the same pre-sorting result, there may be dense targets with overlapping parameters.

[0037] Step 3: For each pre-sorting result in Step 2, a linear fitting algorithm is used to obtain a coarse estimate of the time-azimuth. The specific steps are as follows: Assume the arrival time of each full-pulse data is TOA. i (i = 1, 2, ..., K), with orientation DOA i (i = 1, 2, ..., K), where N represents the total number of full pulses. A coarse estimate of the time-azimuth line can be expressed as: DOA i =A·TOA i +B, where A is the slope of the line and B is the slope of the line. Form a column vector DOA with the positions of all full pulses. vec =[DOA1,DOA2,…,DOA] N ] T The arrival times of all full pulses are combined into a column vector TOA. vec =[TOA1,TOA2,…,TOA N ] T ,in[·] T This indicates the transpose. Then, by linear fitting, the estimated values ​​of A and B are obtained as follows:

[0038]

[0039] Where ε is a very small positive number, usually taken as ε = 10. -4 Or ε = 10 -5 Or ε = 10 -6 eye(2) represents a diagonal matrix of order 2.

[0040] Step 4: Based on the obtained time-azimuth coarse estimate, set the parameter range for fine line detection. Assume the azimuth resolution of the reconnaissance system is δ, and the pre-sorting time range in Step 2 is Range. TOA The directional range is Range DOA .

[0041] Step 4.1: Since the azimuth range of the pre-sorting in Step 2 usually doesn't change much, the slope of the fine line also doesn't change much. To reduce the computational load, the slope value for the fine line detection can be selected as [value missing].

[0042] Step 4.2: Number of intercepts for fine line detection; pre-sorting azimuth range. DOA The number of [items] is related to the system's azimuth resolution δ, and the formula for calculating its number is as follows: Where round(·) represents rounding; the intercept value for fine line detection can then be selected as...

[0043] Step 5: Along the slope and intercept The straight line formed Search for all pulses falling on this line and record their number as Num. m,n Num m,n Form a 3×(N+1) matrix.

[0044] Step 6: For Num m,n By padding the edges with zeros, we get

[0045]

[0046] Step 7: [Regarding...] The search for the maximum value determines the time-azimuth line containing the target location. The formula for calculating the maximum value is:

[0047] Num m,n >Num m,n-1 &Num m,n >Num m-1,n ...

[0048] Num m,n ≥Num m,n+1 &Num m,n ≥Num m+1,n &Num m,n ≥Thres

[0049] Thres is the threshold that is set by the user.

[0050] Step 8: Based on The time-azimuth line corresponding to the maximum value is used for full-pulse screening to obtain the sorting results of densely overlapping targets.

[0051] In other embodiments of the present invention, the following steps are specifically included:

[0052] Step 1: Buffer the input full pulse data for 20 seconds;

[0053] Step 2: Perform pre-sorting on the buffered full pulse data based on frequency and azimuth histogram processing to achieve target resolution that can be distinguished by frequency or azimuth.

[0054] Step 3: For each pre-sorting result in Step 2, a linear fitting algorithm is used to obtain a coarse estimate of the time-azimuth. The specific steps are as follows: Assume the arrival time of each full-pulse data is TOA. i (i = 1, 2, ..., K), with orientation DOA i (i = 1, 2, ..., K), where N represents the total number of full pulses. A coarse estimate of the time-azimuth line can be expressed as: DOA i =A·TOAi +B, where A is the slope of the line and B is the slope of the line. Form a column vector DOA with the positions of all full pulses. vec =[DOA1,DOA2,…,DOA] N ] T The arrival times of all full pulses are combined into a column vector TOA. vec =[TOA1,TOA2,…,TOA N ] T ,in[·] T This indicates the transpose. Then, by linear fitting, the estimated values ​​of A and B are obtained as follows:

[0055]

[0056] Here we take ε = 10 -4 eye(2) represents a diagonal matrix of order 2.

[0057] Step 4: Based on the obtained time-azimuth coarse estimate, set the parameter range for fine line detection. Assume the azimuth resolution of the reconnaissance system is 0.5°, the pre-sorting time range in Step 2 is 20 seconds, and the azimuth range is 5°.

[0058] Step 4.1: Since the azimuth range of the pre-sorting in Step 2 usually doesn't change much, the slope of the fine line also doesn't change much. To reduce the computational load, the slope value for the fine line detection can be selected as [value missing].

[0059] Step 4.2: Number of intercepts for fine line detection; pre-sorting azimuth range. DOA The number of values ​​related to the system's azimuth resolution δ is calculated using the formula N = 10, where round(·) represents rounding; therefore, the intercept value for fine line detection can be selected as...

[0060] Step 5: Along the slope and intercept The straight line formed Search for all pulses falling on this line and record their number as Num. m,n Num m,n Form a 3×11 matrix.

[0061] Step 6: For Num m,n By padding the edges with zeros, we get

[0062]

[0063] Step 7: [Regarding...] The search for the maximum value determines the time-azimuth line containing the target location. The formula for calculating the maximum value is:

[0064] Num m,n >Num m,n-1 &Num m,n >Num m-1,n ...

[0065] Num m,n ≥Num m,n+1 &Num m,n ≥Num m+1,n &Num m,n ≥100

[0066] Step 8: Based on The time-azimuth line corresponding to the maximum value is used for full-pulse screening to obtain the sorting results of densely overlapping targets.

[0067] It should be noted that, within the scope of protection defined in the claims of this invention, the following embodiments can be combined and / or extended or replaced in any logical manner from the above specific embodiments, such as the disclosed technical principles, disclosed technical features or implicitly disclosed technical features.

[0068] Example 1

[0069] A dense target sorting method includes the following steps:

[0070] First, the arriving full pulse data is buffered, and the buffered data is separated by frequency or azimuth using a pre-sorting algorithm to obtain the pre-sorting result;

[0071] Then, for each pre-sorting result, further resolution of densely overlapping targets with overlapping parameters is performed.

[0072] Example 2

[0073] Based on Example 1, the pre-sorting algorithm includes a pre-sorting algorithm based on frequency and orientation histogram processing.

[0074] Example 3

[0075] Based on Example 1, the step of buffering the arriving full pulse data includes the sub-step of buffering the input full pulse data for a time of T.

[0076] Example 4

[0077] Based on Example 1, the process of using the pre-sorting algorithm to separate targets in the cached data by frequency or orientation specifically involves using the pre-sorting algorithm to separate targets in the cached data that can be distinguished by frequency or orientation; and, within the same pre-sorting result, there may be dense targets with overlapping parameters.

[0078] Example 5

[0079] Based on Example 1, the resolution of densely overlapping targets with overlapping parameters includes the following sub-steps: resolving densely overlapping targets with overlapping parameters through coarse fitting of time-azimuth lines, fine line detection, and extreme value search.

[0080] Example 6

[0081] Based on Example 5, the method of resolving densely overlapping targets through coarse fitting of time-azimuth lines, fine line detection, and extreme value search specifically includes the following sub-steps:

[0082] Step a: For each pre-sorting result, a linear fitting algorithm is used to obtain a coarse estimate of time-azimuth; the specific steps are as follows: Assume the arrival time of each full-pulse data is TOA. i i = 1, 2, ..., K, with orientation DOA i Let i = 1, 2, ..., K, where N represents the total number of pulses; the coarsely estimated time-azimuth line is represented as: DOA i =A·TOA i +B, where A is the slope of the line and B is the slope of the line; the orientations of all full pulses are combined into a column vector DOA. vec =[DOA1,DOA2,…,DOA] N ] T The arrival times of all full pulses are combined into a column vector TOA. vec =[TOA1,TOA2,…,TOA N ] T ,in[·] T Let A represent the transpose; then, by linear fitting, the estimated values ​​of A and B are obtained as follows:

[0083]

[0084] Where ε is a positive number, and eye(2) represents a diagonal matrix of order 2;

[0085] Step b: Based on the obtained time-azimuth coarse estimate, set the parameter range for fine line detection; assuming the azimuth resolution of the reconnaissance system is δ, obtain the pre-sorting time range in the pre-sorting step as Range. TOA The directional range is Range DOAThe slope value for fine line detection is set to... Number of intercepts for fine line detection; pre-sorting azimuth range. DOA The number of [items] is related to the system's azimuth resolution δ, and the formula for calculating its number is as follows: Where round(·) represents rounding; then the intercept value for fine line detection is set to

[0086] Step c: Along the slope and intercept The straight line formed Search for all pulses falling on this line and record their number as Num. m,n Num m,n Form a 3×(N+1) matrix;

[0087] Step d: For Num m,n By padding the edges with zeros, we get

[0088]

[0089] Step e: For The search for the maximum value determines the time-azimuth line containing the target location; the formula for calculating the maximum value is:

[0090] Num m,n >Num m,n-1 &Num m,n >Num m-1,n ...

[0091] Num m,n ≥Num m,n+1 &Num m,n ≥Num m+1,n &Num m,n ≥Thres

[0092] Thres is the threshold that is set by the user;

[0093] Step f: Based on The time-azimuth line corresponding to the maximum value is used for full-pulse screening to obtain the sorting results of densely overlapping targets.

[0094] Example 7

[0095] Based on Example 6, in step a, ε = 10 -4 Or ε = 10 -5 Or ε = 10 -6 .

[0096] Example 8

[0097] Based on Example 3, the value range of T includes the ten-second range, and long-term statistical results are obtained by utilizing the value range of T.

[0098] Example 9

[0099] A dense target sorting device includes a processor and a memory, wherein the memory stores a computer program that, when loaded by the processor, executes the method described in any one of Embodiments 1 to 8.

[0100] Example 10

[0101] A computer-readable storage medium storing a computer program, the computer program being loaded by a processor and executed as described in any one of Examples 1 to 8.

[0102] The units described in the embodiments of the present invention can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0103] According to one aspect of the present invention, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the methods provided in the various optional implementations described above.

[0104] In another aspect, embodiments of the present invention also provide a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the methods described in the above embodiments.

Claims

1. A dense target sorting method, characterized in that, Includes the following steps: First, the arriving full pulse data is buffered, and the buffered data is separated by frequency or azimuth using a pre-sorting algorithm to obtain the pre-sorting result; Then, for each pre-sorting result, further resolution of densely overlapping targets with parameters is performed; The resolution of densely overlapping targets with overlapping parameters includes the following sub-steps: achieving the resolution of densely overlapping targets with overlapping parameters through coarse fitting of time-azimuth lines, fine line detection, and extreme value search. The method of resolving densely overlapping targets by coarse fitting of time-azimuth lines, fine line detection, and extreme value search specifically includes the following sub-steps: Step a: For each pre-sorting result, a linear fitting algorithm is used to obtain a coarse estimate of the time-azimuth; the specific steps are as follows: assuming the arrival time of each full-pulse data is... , The direction is , Where N represents the total number of full pulses; the coarsely estimated time-azimuth line is expressed as: Where A is the slope of the line and B is the intercept of the line; the orientations of all full pulses are grouped into a column vector. The arrival times of all full pulses are arranged into a column vector. ,in Let A represent the transpose; then, by linear fitting, the estimated values ​​of A and B are obtained as follows: in, It is a positive number. This represents a diagonal matrix of order 2; Step b: Based on the obtained time-azimuth coarse estimate, set the parameter range for fine line detection; assume the azimuth resolution of the reconnaissance system is... The time range for pre-sorting in the pre-sorting step is... The directional range is The slope value for fine line detection is set to... Number of intercepts for fine line detection; pre-sorting azimuth range and system orientation resolution The formula for calculating the number of items is as follows: ,in Indicates rounding; therefore, the intercept value for fine line detection is set to... ; Step c: Along the slope and intercept The straight line formed Search for the complete pulses that fall on the straight line and record their number. , constitute Matrix; Step d: For By padding the edges with zeros, we get , ； Step e: For The search for the maximum value determines the time-azimuth line containing the target location; the formula for calculating the maximum value is: in It is a threshold set by humans; Step f: Based on The time-azimuth line corresponding to the maximum value is used for full-pulse screening to obtain the sorting results of densely overlapping targets.

2. The dense target sorting method according to claim 1, characterized in that, The pre-sorting algorithm includes a pre-sorting algorithm based on frequency and orientation histogram processing.

3. The dense target sorting method according to claim 1, characterized in that, The step of buffering the arriving full pulse data includes the sub-step of buffering the input full pulse data for a time of T.

4. The dense target sorting method according to claim 1, characterized in that, The process of using a pre-sorting algorithm to separate targets in the cached data by frequency or orientation specifically involves using a pre-sorting algorithm to separate targets in the cached data that can be distinguished by frequency or orientation; and within the same pre-sorting result, there may be dense targets with overlapping parameters.

5. The dense target sorting method according to claim 1, characterized in that, In step a, or or .

6. The dense target sorting method according to claim 3, characterized in that, The value range of T includes tens of seconds, and long-term statistical results can be obtained by using the range of T values.

7. A dense target sorting device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that, when loaded by the processor, executes the method as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that, A computer program is stored in a readable storage medium, the computer program being loaded by a processor and executing the method as described in any one of claims 1 to 6.

Images