A Queue-Based Two-Dimensional Agglomeration Fast Method

Abstract

The present invention relates to a two-dimensional aggregation fast method based on a queue. By correlating the traces obtained after constant false alarm, the method can simultaneously complete range dimension aggregation and Doppler dimension aggregation, and includes the following steps: (a) Number the traces obtained after constant false alarm and clutter map detection and record them in the range-Doppler two-dimensional plane of the corresponding beam; (b) Starting from the first target, determine whether there is a target with a range and Doppler difference of 1 in the plane. If so, continue to correlate with the new target as the center; (c) For the targets correlated in the above process, calculate the center in terms of range and Doppler to obtain the aggregated target information, and mark the correlated targets; (d) Start re-correlating from the next uncorrelated target until all target points are aggregated. The method of the present invention greatly reduces the computational complexity.

Description

Technical Field

[0001] The present invention relates to the field of radar signal processing, and in particular to a point trace condensation method. The present invention specifically relates to a queue-based two-dimensional point trace condensation method, which realizes two-dimensional condensation with lower complexity and has obvious advantages when there are many target point traces, making it possible to implement the two-dimensional condensation method in a resource-constrained processor such as a DSP. Background Art

[0002] The general processing flow of signal processing in linear frequency modulation pulse Doppler radar is AD sampling, digital down-conversion, digital beam forming, pulse compression, moving target detection, constant false alarm detection, clutter map, point trace condensation, target angle measurement, etc. Among them, point trace condensation can remove false alarm points and eliminate redundant targets. Correct and efficient condensation processing can lay a good foundation for subsequent processing procedures.

[0003] At present, the most commonly used one-dimensional condensation algorithm is to divide the condensation process into two steps. The first step is to do distance condensation. The detected points are sorted in the order of beam number, Doppler channel number, and distance unit number. In each Doppler channel of each beam, the points with continuous distance unit numbers are condensed, and the centroid of these condensed points is calculated. The second step is to do Doppler condensation. The targets with the same distance on each channel are adjacent in the storage space, which is convenient for Doppler selection. Although this algorithm has a relatively low complexity, it separates the speed information and distance information of the target, which may cause the target to be discrete, that is, the center of gravity on the range gate is staggered, resulting in the same target condensation on the speed gate is also staggered.

[0004] The agglomeration algorithm based on the parent-child nodes traverses each detected target point multiple times and establishes the following relationship between them: Figure 1 The relationship shown. Then, start traversing from the first point, and each time condense the surrounding points and all the points associated with these points into one point. However, if n targets need to associate parent-child nodes, then the first target needs to make judgments on the left parent, middle parent, right parent, left child, neutron child, and right child with the remaining n-1 targets, and the second target needs to make judgments on the left parent, middle parent, right parent, left child, neutron child, and right child with the remaining n-2 targets... The n-1th target needs to make judgments on the left parent, middle parent, right parent, left child, neutron child, and right child with the last target. Therefore, the number of judgments to be made when associating n targets is n(n-1) / 2*6, which can be approximately regarded as a complexity of O(n^2). When the number of points is large, it cannot meet the processing requirements. Summary of the invention

[0005] The technical problems to be solved by the present invention are:

[0006] To solve the problems of target splitting in traditional one-dimensional clustering algorithms and high computational complexity in two-dimensional clustering algorithms, the present invention provides a fast two-dimensional clustering method based on a queue.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0008] A fast method for two-dimensional point clustering of range-Doppler based on a queue. The method correlates the points obtained through constant false alarm and clutter map detection, and can simultaneously complete range dimension clustering and Doppler dimension clustering. Its characteristics are as follows:

[0009] Step 1: Number the points obtained through constant false alarm and clutter map detection and record them in the range-Doppler two-dimensional plane of the corresponding beam.

[0010] Step 2: Starting from the first target, determine whether there is a target with a range and Doppler difference of 1 in the plane. If so, continue to correlate with the new target as the center.

[0011] Step 3: Calculate the center of the correlated targets in terms of range and Doppler to obtain the information of the clustered target, and mark the correlated targets.

[0012] Step 4: Start re-correlating from the next uncorrelated target until all target points are clustered.

[0013] A further technical solution of the present invention: In the range-Doppler two-dimensional plane established in Step 1, the point numbers of the targets are stored, and the points are not sorted according to a certain feature. It is only necessary to ensure that the point numbers within the same beam are unique.

[0014] A further technical solution of the present invention: In Step 2, determining whether there is a target with a range and Doppler difference of 1 in the plane means determining whether there is a target in the 3*3 plane that can be correlated with the current point as the center.

[0015] A further technical solution of the present invention: In Step 2, the point numbers should be clearly distinguishable from the default numbers filled in the places where there are no targets in the plane. The distinguishing methods include but are not limited to using non-negative integers starting from 0 for the point numbers, and uniformly filling -1 in the places where there are no targets.

[0016] A further technical solution of the present invention: The judgment condition in Step 2 can be modified to: A range or Doppler difference of 2 can also be regarded as the same target, which is equivalent to expanding to determining whether there is a target in the 4*4 plane that can be correlated with it.

[0017] A further technical solution of the present invention: After clustering the correlated targets in Step 3, the range information dis_a and velocity information dop_a can be obtained by the center calculation method from the following formula:

[0018]

[0019]

[0020] Where M is the number of target traces associated with the current stage, dis(j) is the distance cell number of the j-th trace, dop(j) is the Doppler cell number of the j-th trace, and amp(j) is the amplitude of the j-th trace.

[0021] A computer system, characterized in that it includes: one or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the above method.

[0022] A computer-readable storage medium, characterized in that it stores computer-executable instructions, and the instructions are used to implement the above method when executed.

[0023] The beneficial effects of the present invention are as follows:

[0024] A fast method for two-dimensional trace condensation of range-Doppler based on a queue provided by the present invention. The method can complete range dimension condensation and Doppler dimension condensation simultaneously by associating the traces obtained after constant false alarm, and includes the following steps: (a) Number the traces obtained after constant false alarm and clutter map detection and record them in the range-Doppler two-dimensional plane of the corresponding beam; (b) Starting from the first target, judge whether there is a target with a range and Doppler difference of 1 in the plane. If so, continue to associate with the new target as the center; (c) Calculate the center of the associated targets in terms of range and Doppler to obtain the condensed target information, and mark the associated targets; (d) Start re-associating from the next unassociated target until all target traces are condensed.

[0025] Since the method of the present invention uses the information of target numbers in the range-Doppler plane during the process of associating targets, it is not necessary to repeatedly traverse all target traces to obtain association information. At most, only n*8 judgments need to be made for the number of targets, and the time complexity can be approximately regarded as O(n). Compared with the trace condensation method based on parent-child node traversal in the prior art, it is faster and more efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings are only for the purpose of showing specific embodiments, and are not considered to be a limitation of the present invention. Throughout the drawings, the same reference signs represent the same components.

[0027] Figure 1 It is a schematic structural diagram based on parent-child node traversal.

[0028] Figure 2 It is a flowchart of the steps of the method of the present invention.

[0029] Figure 3 It is an example of the plot distribution diagram of the range-Doppler plane. Detailed implementation manners

[0030] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0031] As Figure 2 shown, a two-dimensional condensation fast method based on a queue provided by the present invention includes the following steps:

[0032] S101: Number the plots obtained through constant false alarm and clutter map detection and record them in the range-Doppler two-dimensional plane of the corresponding beam;

[0033] Furthermore, the plot numbers of the targets are stored in the established range-Doppler two-dimensional plane. Since the filling order does not affect the result, it is not required to sort the plots according to a certain feature. It is only necessary to ensure that the plot numbers within the same beam are unique.

[0034] S102: Starting from the first target, judge whether there is a target with a range and Doppler difference of 1 in the plane. If so, continue to associate with the new target as the center;

[0035] Furthermore, judging whether there is a target with a range and Doppler difference of 1 in the plane is equivalent to judging whether there is a target in the 3*3 plane that can be associated with the current point as the center;

[0036] Furthermore, the plot numbers should be clearly distinguishable from the default numbers filled in the places without targets in the plane. The distinguishing methods include but are not limited to using non-negative integers starting from 0 for the plot numbers, and uniformly filling -1 in the places without targets;

[0037] Furthermore, the judgment condition can be appropriately modified. In some cases, a range or Doppler difference of 2 can also be regarded as the same target, which is equivalent to expanding to judging whether there is a target in the 4*4 plane that can be associated with it;

[0038] S103: Calculate the center of the associated targets in terms of range and Doppler to obtain the condensed target information, and mark the associated targets;

[0039] Further, after the associated targets are aggregated, the distance information and velocity information can be obtained by the centroid method as follows:

[0040]

[0041]

[0042] where M is the number of target traces associated in the current stage, dis(j) is the distance cell number of the j-th trace, dop(j) is the Doppler cell number of the j-th trace, and amp(j) is the amplitude of the j-th trace.

[0043] S104: Start re-associating from the next unassociated target until all target points are aggregated.

[0044] Example:

[0045] In step S101, different arrays need to be established for sub-beams and chips in the range-Doppler two-dimensional plane. When initializing, default values distinguishable from the trace numbers should be filled. If the trace number starts from 1, the default value can be filled with 0, and 0 should also be filled when marking that the target has been aggregated. If the trace number starts from 0, the default value can be filled with -1. Since the filling order does not affect the result, it is not required to sort the traces according to a certain characteristic.

[0046] In step S102, starting from the first target, determine whether there are surrounding targets that can be associated with it by means of the range-Doppler two-dimensional plane established in step S101. If so, continue to associate with the new target as the center.

[0047] In Figure 3 , P is an example node (assuming the Doppler channel number is i and the range cell number is j); the coordinates of its surrounding 8 points are respectively (i - 1, j - 1), (i, j - 1), (i + 1, j - 1), (i - 1, j), (i + 1, j), (i - 1, j + 1), (i, j + 1), (i + 1, j + 1). The association process is as follows:

[0048] (a) Enqueue the P node number, determine that the queue is not empty, then dequeue P, set the number at the corresponding position of the P node on the plane to -1, and record its range cell number dis(1), Doppler cell number dop(1), and amplitude amp(1);

[0049] (b) Perform an association judgment on the P node, that is, whether the trace numbers of the surrounding 8 nodes are greater than or equal to 0. The association order is to judge clockwise starting from the upper left corner. The Q, S, and T nodes can be associated and enqueued in the first association, and Q is the head element of the queue and T is the tail;

[0050] (c) Determine whether the queue is empty. If it is not empty, dequeue Q, reset the number, and record its distance, Doppler, and amplitude information. After associating the Q node, R can be enqueued. At this time, S is the head element of the queue and R is the tail element;

[0051] (d) Continue to determine whether the queue is empty. If it is not empty, dequeue S, reset the number, and record its distance, Doppler, and amplitude information. After associating the S node, no new node can be enqueued. At this time, only the target number of the R node remains in the queue;

[0052] (e) Continue to determine whether the queue is empty. If it is not empty, dequeue R, reset the number, and record its distance, Doppler, and amplitude information. After associating the R node, no new node can be enqueued. At this time, the queue is empty;

[0053] (f) Determine that the queue is empty, and go to step S103 to calculate the distance and Doppler information of the target after aggregation.

[0054] For Figure 3 After processing the example target according to this algorithm flow, the amplitude amp_t, the accurate Doppler unit number dop_t, and the accurate distance unit number dis_t of the aggregated target are respectively:

[0055]

[0056] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should be covered within the protection scope of the present invention.

Claims

1. A fast method for distance-Doppler two-dimensional point track condensation based on a queue. The point tracks obtained after constant false alarm and clutter map detection are associated, and distance dimension condensation and Doppler dimension condensation can be completed simultaneously. It is characterized in that the steps are as follows: Step 1: Number the point tracks obtained after constant false alarm and clutter map detection and record them in the distance-Doppler two-dimensional plane of the corresponding beam. In the distance-Doppler two-dimensional plane established in Step 1, the point track numbers of the targets are stored, and the point tracks are not sorted according to a certain feature. It is only necessary to ensure that the point track numbers within the same beam are unique. Step 2: Starting from the first target, judge whether there is a target with a distance and Doppler difference of 1 in the plane. If so, continue to associate with the new target as the center. Step 3: Calculate the center of the associated targets in terms of distance and Doppler to obtain the target information after condensation, and mark the associated targets. The distance information dis_a and velocity information dop_a after condensation of the associated targets can be obtained by the centering method as follows: where M is the number of point tracks of the targets associated in the current stage, dis(j) is the distance cell number of the j-th point track, dop(j) is the Doppler cell number of the j-th point track, and amp(j) is the amplitude of the j-th point track. Step 4: Start re-associating from the next unassociated target until all target points are condensed.

2. The fast method for distance-Doppler two-dimensional point track condensation based on a queue according to claim 1, it is characterized in that: In Step 2, judging whether there is a target with a distance and Doppler difference of 1 in the plane means judging whether there is a target in the 3*3 plane centered on the current point that can be associated with it.

3. The fast method for distance-Doppler two-dimensional point track condensation based on a queue according to claim 2, it is characterized in that: The point track numbers in Step 2 should be clearly distinguishable from the default numbers filled in the places where there are no targets in the plane. The distinguishing methods include but are not limited to using non-negative integers starting from 0 for the point track numbers, and uniformly filling -1 in the places where there are no targets.

4. The fast method for distance-Doppler two-dimensional point track condensation based on a queue according to claim 3, it is characterized in that: The judgment condition in Step 2 can be modified to: a distance or Doppler difference of 2 can also be regarded as the same target, which is equivalent to expanding to judging whether there is a target in the 4*4 plane that can be associated with it.

5. A computer system, it is characterized in that it includes: One or more processors, a computer-readable storage medium for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to claim 1.

6. A computer-readable storage medium, it is characterized in that it stores computer-executable instructions, and the instructions are used to implement the method according to claim 1 when executed.

Images