A target detection method based on target merging in radar frame

Abstract

The present invention relates to a target detection method for merging targets within a radar frame, and belongs to the technical field of radar target detection. It includes: obtaining radar echo data of a radar scan, and obtaining frequency domain data after down-conversion, pulse compression and MTD processing; determining the maximum peak value P0 of the frequency domain data of a frame of echo and its corresponding frequency channel F0; comparing the maximum peak value P0 with the minimum threshold β when there is a target to determine whether there is a target in this frame of data; if there is a target, determining the merging range [F0-ΔF, F0+ΔF] of the target represented by the main peak; searching for a higher spectrum peak in the frequency domain outside [F0-ΔF, F0+ΔF], that is, the peak value of the suspected target, and comparing it with the moving threshold γ; and finally determining the number of targets in this frame of data. The target detection method can effectively separate the situation where there are multiple targets in the data.

Description

Technical Field

[0001] The invention relates to a target detection method for merging targets within a radar frame, and belongs to the technical field of radar target detection. Background Art

[0002] Target detection method is an important part of radar signal processing. When detecting a target, multiple spectral peaks often appear in the received echo spectrum. The reasons for this phenomenon are: there are two or more targets close to each other in the echo signal; the target is in a complex electromagnetic environment or there are other electromagnetic interferences. Some of these spectral peaks are generated by one target, and some are generated by multiple targets. In order to effectively separate the target, the problem of radar target merging is involved. The spectral peaks related to a target need to be merged and considered as one target, and the spectral peaks of multiple targets cannot be merged and considered as one target. Separating the desired target in such a complex environment with multiple targets and various interferences is a higher standard requirement for radar signal processing.

[0003] The radar signal target detection method is: a constant false alarm detection method with a unified threshold for the radar signal, and all points in the echo with intensity higher than the detection threshold are regarded as targets. However, this will result in one target being reported as multiple targets, and it is difficult for this method to detect multiple targets with large intensity differences using the same threshold, and it is also difficult to detect targets covered by interference. It also has poor adaptability to intensity fluctuations of the target, which makes the radar's target detection capability low.

[0004] After searching the patent databases of the State Intellectual Property Office and Wanfang, the following conclusions were drawn through analysis:

[0005] Through the keyword combination of radar + target detection, target merging, target condensation, point trace merging, and point trace fusion, dozens of patent documents were retrieved. After screening, no documents similar to the patent "intra-frame target merging method for frequency domain widening" were found. Only the following papers mentioned the intra-frame target merging method more, but the methods they proposed are completely different from those of this patent, including:

[0006] (1) A target merging method in ARPA target tracking of a ship navigation radar comprises the following steps: obtaining echo data, obtaining a target distance unit, forming a small connected area, forming a large connected area, obtaining the centroid and amplitude sum of the small connected area in a single azimuth, and obtaining the centroid of the large connected area. The present invention utilizes a target merging method based on connected areas after receiving a circle of radar echo data. The method can accurately merge connected area targets into one target, has a small amount of calculation, and is easy to implement in engineering.

[0007] (1) A target aggregation method based on a ground surveillance radar system, comprising: obtaining spectral data information after radar front-end processing; performing spectral data cancellation processing on the spectral data information to obtain spectral data information after cancellation processing; performing constant false alarm rate processing on the spectral data information after cancellation processing to obtain spectral data information of the target, and forming a two-dimensional image based on the spectral data information of the target; performing target edge extraction on the two-dimensional image to extract spectral data information belonging to the same target; performing ground object target filtering on the spectral data information belonging to the same target to obtain spectral data information belonging to the same target after ground object target filtering. The present invention reduces the impact of fixed strong clutter on low-speed targets, increases the probability of detecting low-speed targets, and provides a reliable basis for subsequent target aggregation and track processing.

[0008] (2) A target aggregation method for an airport scene surveillance radar system, firstly obtaining target echo information after radar front-end processing, then performing false point trace filtering on the target echo information to obtain filtered target echo information, performing channel selection processing on the filtered target echo information, detecting target echo information belonging to the same target to form a two-dimensional image, estimating target parameters using the two-dimensional image contour and the target echo information within the contour, and obtaining target parameters after aggregation of the same target. The present invention greatly improves the accuracy of target aggregation, reduces the splitting of the target in the distance direction and the azimuth direction, and provides a reliable basis for subsequent track processing.

[0009] (3) A target extraction method based on ground surveillance radar spectrum data, comprising: obtaining spectrum data information after radar front-end processing; performing spectrum data cancellation processing on the spectrum data information to obtain spectrum data information after cancellation processing; performing constant false alarm rate processing on the spectrum data information after cancellation processing to obtain spectrum data information of the target, and forming a two-dimensional image based on the spectrum data information of the target; performing target edge extraction on the two-dimensional image to extract spectrum data information belonging to the same target; performing ground object target filtering on the spectrum data information belonging to the same target to obtain spectrum data information belonging to the same target after ground object target filtering. The present invention reduces the impact of fixed strong clutter on low-speed targets, increases the probability of detecting low-speed targets, and provides a reliable basis for subsequent target aggregation and track processing.

[0010] (4) A radar target point trace condensation method based on the connected labeling rule, the method comprising obtaining an original target point trace sequence; initializing labeling information; connecting domain labeling; target screening and point trace filtering; condensing all points traces with the same label number. The method of the present invention directly connects and labels the original point trace sequence after detection by equivalent preceding adjacent point trace sets, and then condenses all points traces with the same label number by using the amplitude weighting method. The method of the present invention can effectively realize target condensation in complex and dense environments, and substantially solves the problem of target splitting after constant false alarm detection. It has good compatibility with target size and shape, and has small data volume and calculation amount, and is easy to implement.

[0011] (5) A method for detecting and merging weak targets based on inter-frame accumulation, comprising: step S1: acquiring radar echo signals; step S2: storing the acquired echo data as a two-dimensional matrix of distance-azimuth; step S3: performing inter-frame accumulation on the current frame data; step S4: merging connected areas in the azimuth direction; step S5: calculating information of the merged connected areas in the azimuth direction; step S6: repeating steps S4 and S5, merging all connected areas in the azimuth direction and calculating target information; step S7: determining whether the connected area is connected to the connected area in the previous azimuth direction; step S8: calculating information of the merged connected areas in adjacent azimuth directions; step S9: updating information of the merged connected areas; step S10: starting from the second azimuth direction, repeating steps S7, S8 and S9, merging all connected areas in the azimuth direction that can be merged in turn and updating their information.

[0012] This application is based on a frame of target echo data received by the same distance unit radar. When the target and the radar are in relative motion, the spectrum broadening diagram of the target appearing on the frequency axis is used to merge the targets, thereby improving the radar's ability to correctly detect targets in complex environments. Summary of the invention

[0013] The purpose of the present invention is to propose a target detection method for merging targets in radar frames to address the technical defects that targets in radar echoes encounter complex environments and are covered by various interferences or main lobes, resulting in a decrease in target detection capability.

[0014] In order to achieve the above object, the present invention adopts the following technical scheme.

[0015] The target detection method for merging targets in the radar frame comprises the following steps:

[0016] Step 1: Obtain radar echo data of a radar scan, and obtain frequency domain echo after digital down-conversion, pulse compression and MTD processing;

[0017] Step 2: determine the maximum spectrum peak in the frequency domain echo and the frequency peak channel corresponding to the spectrum peak;

[0018] Among them, the maximum spectrum peak is recorded as P0; the frequency peak channel corresponding to the spectrum peak is recorded as F0;

[0019] Step 3: compare the maximum spectrum peak value with the target minimum threshold β. If the maximum spectrum peak value is less than or equal to β, end the method; otherwise, if the maximum spectrum peak value is greater than β, proceed to step 4.

[0020] Among them, the maximum spectrum peak value P0 is less than or equal to β, indicating that there is no target information in this frame of data; the maximum spectrum peak value P0 is greater than β, indicating that there is target information in this frame of data;

[0021] The calculation formula of the minimum threshold β is as follows (1):

[0022]

[0023] Among them, a i is the amplitude of the ith frequency channel; N is the total number of frequency channels of the frequency domain echo;

[0024] Step 4: Taking the frequency peak channel determined in step 2 as the center, set the main peak target frequency domain merging area [F0-ΔF, F0+ΔF];

[0025] Among them, [F0-ΔF, F0+ΔF] represents the target merging area of ​​the main peak, and the range is selected as the entire frequency domain channel [0, F max ] is 15% to 25%, that is, the value range of 2ΔF is 0.15*F max To 0.25*F max ;

[0026] Step 5: Find a higher spectrum peak in the frequency domain outside the main peak target frequency domain merging area;

[0027] Among them, the higher spectral peak is denoted as P j , j = 1, 2, 3, ...;

[0028] Step 6: Determine the size of the higher spectrum peak and the dynamic threshold γ. If the higher spectrum peak is greater than or equal to γ, record the frequency domain position corresponding to the peak. After traversing all the higher spectrum peaks, the target detected in the current frame is obtained, specifically:

[0029] The jth higher spectral peak P j Compared with the dynamic threshold γ, if P j Greater than or equal to γ, record the peak value P j The corresponding frequency domain position F j , after traversing all higher peaks, the targets detected in the current frame can be obtained;

[0030] Among them, Pj Greater than or equal to γ, indicating the current peak value P j is a goal; j Less than γ, indicating the current peak value P j It is not a goal;

[0031] Among them, the dynamic threshold γ is set according to the maximum spectrum peak value P0;

[0032] γ=P0 / m, where m ranges from 1.5 to 3. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 The present invention is a flowchart of an implementation of a target detection method for merging targets within a radar frame. Implementation

[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the protection scope of the present invention.

[0035] Example 1

[0036] This embodiment describes the specific implementation of the method of the present invention. Figure 1 As shown, it is an implementation flow chart of an implementation flow of a target detection method for merging targets within a radar frame of the present invention.

[0037] Step A, receiving radar target echo data scanned by the radar once;

[0038] Step B, performing digital down-conversion, pulse compression and MTD processing on the radar target echo data of the first scan received in step A, and then performing modulus calculation to obtain a frequency domain echo of the first scan;

[0039] Step C, determining the position F0 of the maximum peak value in the frequency domain echo of a scan on the frequency axis, and recording the peak value P0;

[0040] Step D, determining whether the maximum spectrum peak P0 in the frequency domain echo of a scan is higher than the minimum threshold β when there is a target, if it is higher than the β threshold, executing step E; otherwise, executing step I;

[0041] Step E, select the position F0 corresponding to the peak value P0 as the center, set the target frequency domain merging area [F0-ΔF, F0+ΔF], the selected range 2ΔF is 23% of the entire frequency domain range, and the peak value P0 of the target after merging is used as the judgment condition for whether other targets exist, and N1=1;

[0042] Step F: According to step D, P0 is obtained, and the decision threshold γ=P0 / m of whether there are other targets is obtained. In the specific implementation, m=2. Select all the relatively high spectrum peaks P1, P2, P3, etc. in the remaining 77% of the frequency domain range, and determine whether they are greater than or equal to γ. If they are greater than or equal to γ, execute step G; otherwise, execute step H;

[0043] Step G: If P1, P2, P3, ... are greater than or equal to γ, it is considered that there are other targets at their peaks. If there are n spectrum peaks greater than or equal to γ ​​in the remaining 77% frequency domain, it is considered that there are n targets, N2 = n, and step J is executed;

[0044] Step H, if P1, P2, P3, ... are less than γ, it is considered to be clutter or other interference rather than the desired target N = 0, and step J is executed;

[0045] Step I: No target, no target merging is performed. N1=0, N2=0, and step J is executed;

[0046] Step J, counting the number of multiple targets N=N1+N2 of target echoes in a single radar frame.

[0047] The method was used to conduct experiments, using the data of wheeled vehicles and tracked vehicles collected by a certain type of reconnaissance radar, and the frequency domain echo of a single scan was obtained by processing. The number of targets was detected using the method of the present invention and the method without merging targets. Since the acquisition was a uniform movement of one target away from or approaching the radar, theoretically only one target would be detected, but due to various interferences, the number of detected targets would be greater than 1. The results of Table 1 and Table 2 can be obtained by statistically analyzing the detection methods of the two methods:

[0048] Table 1 Detection accuracy of wheeled vehicles

[0049]

[0050] Table 2 Detection accuracy of tracked vehicles

[0051]

[0052] From the comparison results of Table 1 and Table 2, the detection accuracy of this method is improved compared with the traditional detection method. When the radar signal is subject to greater interference, such as the data of wheeled vehicle No. 2, 3 and tracked vehicle No. 5, 6, the accuracy can be improved by 10% to 18% after using this method. When the radar signal is subject to less interference, the traditional detection method can have an accuracy of 90%, and the accuracy can be improved by 5% to 8% after using this method.

[0053] The above is only a preferred embodiment of the present invention, and the present invention should not be limited to the contents disclosed in the embodiment and the drawings. Any equivalent or modification completed without departing from the spirit disclosed in the present invention shall fall within the scope of protection of the present invention.

Claims

1. A target detection method for target merging within a radar frame, characterized in that: The steps include: Step 1: Obtain radar echo data of a radar scan, and obtain frequency domain echo after digital down-conversion, pulse compression and MTD processing; Step 2: determine the maximum spectrum peak value P0 in the frequency domain echo and the frequency peak channel F0 corresponding to the spectrum peak value; Step 3: compare the maximum spectrum peak value with the target minimum threshold β. If the maximum spectrum peak value is less than or equal to β, end the method; otherwise, if the maximum spectrum peak value is greater than β, proceed to step 4. Step 4: With the frequency peak channel F0 determined in step 2 as the center, set the main peak target frequency domain merging area, the merging area is recorded as [F0-ΔF, F0+ΔF], and [F0-ΔF, F0+ΔF] represents the target merging area of ​​the main peak, and the range is selected as the entire frequency domain [0, F max ] is 15% to 25%, that is, the value range of 2ΔF is 0.15*F max To 0.25*F max ; Step 5: Find a higher spectrum peak in the frequency domain outside the main peak target frequency domain merging area; Step 6. Determine the size of the higher spectral peak and the dynamic threshold γ. If the higher spectral peak is greater than or equal to γ, record the frequency domain position corresponding to the peak. After traversing all the higher spectral peaks, obtain the target detected in the current frame; the dynamic threshold γ is set according to the maximum spectral peak P0, specifically: γ = P0 / m, and the value of m is between 1.5 and 3.

2. The target detection method for merging targets within a radar frame according to claim 1, characterized in that: In step 3, if the maximum spectrum peak value P0 is less than or equal to β, it means that there is no target information in this frame of data; if the maximum spectrum peak value P0 is greater than β, it means that there is target information in this frame of data.

3. The target detection method for merging targets within a radar frame according to claim 1, characterized in that: The calculation formula of the minimum threshold β in step 3 is as follows (1): Among them, a i is the amplitude of the i-th frequency channel; N is the total number of frequency channels of the frequency domain echo.

4. The target detection method for merging targets within a radar frame according to claim 1, characterized in that: In step 5, the higher spectral peak is denoted as P j , j = 1, 2, 3 ...; Step 6, specifically: take the jth higher spectral peak value P j Compared with the dynamic threshold γ, if P j Greater than or equal to γ, record the peak value P j The corresponding frequency domain position F j , after traversing all higher peaks, the targets detected in the current frame can be obtained.

5. The target detection method for merging targets within a radar frame according to claim 4, characterized in that: In step 6, P j Greater than or equal to γ, indicating the current peak value P j is a goal; j Less than γ, indicating the current peak value P j Not the goal.

Images