A deviation sum dual-parameter fast estimation constant false alarm rate detection method

Abstract

The application relates to a deviation summation two-parameter fast estimation constant false alarm rate detection method, which realizes the following steps: 1) setting parameters such as a protection window, a reference window and a reference sub-window; 2) receiving logradar pulse echo data; 3) adopting a sliding window mode to sequentially count the background amplitude mean value of each distance sampling unit along the distance direction; 4) calculating the deviation square value of each echo amplitude and the background amplitude mean value; 5) adopting the sliding window mode to sequentially count the background amplitude standard deviation of each distance sampling unit along the distance direction; and 6) calculating a detection threshold according to the background amplitude mean value and the standard deviation, and sequentially performing threshold detection on each distance sampling unit. The application estimates two parameters of the background amplitude mean value and the standard deviation, so that the detector can adaptively cope with a non-uniform clutter environment, a fast parameter estimation means of approximate sorting is adopted, the multiple target environment can be effectively coped with, and the processing speed can be greatly improved.

Description

Technical Field

[0001] This invention belongs to the field of radar target detection technology. Background Technology

[0002] The constant false alarm rate (CFAR) detection method is a technique that adaptively adjusts the detection threshold based on changes in clutter intensity, thereby maximizing the detection probability while maintaining a fixed false alarm probability. To address different background clutter characteristics, CFAR detectors have been developed into mean (ML) CFAR detectors, ordered statistic (OS) CFAR detectors, and two-parameter (DP) CFAR detectors.

[0003] The basic idea of ​​mean-based CFAR detection methods is to estimate the leading-edge clutter power and trailing-edge clutter power using sampled values ​​from reference cells at the leading and trailing edges of the cell to be detected. Then, based on the clutter distribution characteristics, the method performs cell averaging (either by increasing or decreasing the cell averaging) or weighted averaging to determine the background clutter power level of the cell to be detected. The most typical methods are cell averaging (CA-CFAR), large cell averaging (GOCA-CFAR), and small cell averaging (SOCA-CFAR).

[0004] Mean-based constant false alarm rate (CFAR) detection methods are all based on the assumption that clutter amplitude follows a Rayleigh distribution. However, besides the Rayleigh distribution, many probability distribution models, such as the logarithmic distribution or the Weibull distribution, are two-parameter distributions. In these cases, mean-based CFAR detection methods are no longer applicable. The two-parameter constant false alarm rate (DP-CFAR) detection method estimates the mean and standard deviation of the clutter distribution using a reference cell, and the final CFAR detection threshold is determined by both of these parameters. Therefore, the DP-CFAR detection method is applicable to non-uniform clutter environments in the airspace.

[0005] In multi-target scenarios, both mean-based and two-parameter CFAR detection methods tend to overestimate the background mean and standard deviation if targets exist within the reference windows on either side, leading to a decrease in detection probability. Furthermore, the computational complexity of calculating the standard deviation in the two-parameter CFAR method makes it difficult to meet the real-time requirements of the system.

[0006] Ordered Statistical (OS) CFAR detection methods sort the sampled values ​​of the leading and trailing reference cells of the target cell in ascending order, and select the k-th sampled value as the background clutter power level of the target cell. This method is applicable to multi-target environments and non-uniform clutter background environments, but due to the introduction of the sorting algorithm, the processing speed of OS-type detectors is significantly slower. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention proposes a fast dual-parameter estimation method for constant false alarm rate (CFAR) detection using deviation summation. This method estimates two parameters—the mean and standard deviation of the background amplitude—enabling the detector to adaptively handle non-uniform clutter environments. Furthermore, it employs a fast parameter estimation technique with approximate sorting, effectively addressing multi-target environments and significantly improving processing speed.

[0008] The technical solution proposed in this invention is as follows:

[0009] Step 1: Process parameter settings;

[0010] The parameters are set as follows: the length of the single-sided protective window is N. p There are N distance sampling units, and the length of the reference window on one side is N. r There are 1 distance sampling unit, and the number of reference sub-windows divided by the single-sided reference window is M, with a single-sided reference sub-window length of N. s There are distance sampling units, where M can be sampled by N. p Divisible by N s =N p / M.

[0011] Step 2: Receive radar echo data;

[0012] The logarithmic amplitude data of the radar's i-th pulse echo is A i (n), where n is a positive integer and 1≤n≤N, and N is the number of distance sampling units for the i-th pulse;

[0013] Step 3: Estimate the mean background amplitude

[0014] Starting from the first distance sampling unit of the pulse, each distance sampling unit is traversed sequentially along the distance direction using a sliding window method, and the average background amplitude of each distance sampling unit is calculated. The specific steps are as follows:

[0015] Step 3.1: When n=1, the front sub-window is not included in the statistics, and the amplitude and initial value of the back reference sub-window are SumAmp. B (m), the mean of the background amplitude is Where m is a positive integer, and 1≤m≤M.

[0016] Step 3.2: When 2≤n≤N p +N r At this time, the front sub-window is not included in the statistics, and the back sub-window slides along the distance direction and the sum of the amplitudes of the back sub-windows is calculated as SumAmp. B (m), the mean of the background amplitude is

[0017] Step 3.3: When n = N p+N r When +1, the amplitude and initial value of the front reference sub-window are SumAmp. F (m), the rear sub-window slides along the distance direction and the sum of the amplitudes of the rear sub-windows is calculated as SumAmp. B (m), the sum of the amplitudes of the corresponding sub-windows on the front and back sides is taken as MaxSumAmp(m)=max[SumAmp F (m),SumAmp B (m)], the mean background amplitude is

[0018] Step 3.4: When N p +N r +2≤n≤N-(N p +N r When ), the front and rear sub-windows slide simultaneously along the distance direction, and the sum of the amplitudes of the front sub-windows is calculated as SumAmp. F (m), the sum of the amplitudes of the rear sub-windows is SumAmp. B (m), the sum of the amplitudes of the corresponding sub-windows on the front and back sides is taken as MaxSumAmp(m)=max[SumAmp F (m),SumAmp B (m)], the mean background amplitude is

[0019] Step 3.5: When N-(N p +N r When 1 ≤ n ≤ N, the front sub-window slides along the distance direction and the sum of the amplitudes of the front sub-windows is calculated as SumAmp. F (m), the rear sub-window is not included in the statistics, and the mean background amplitude is...

[0020] Step 4: Calculate the squared amplitude deviation value DevSqu;

[0021] Calculate the squared deviation of each echo amplitude from the mean of its background amplitude.

[0022] Step 5: Estimate the standard deviation of the background amplitude σ;

[0023] Starting from the first distance sampling unit of the pulse, each distance sampling unit is traversed sequentially along the distance direction using a sliding window method. The standard deviation of the background amplitude σ(n) of each distance sampling unit is calculated. The specific steps are as follows:

[0024] Step 5.1: When n=1, the front sub-window is not included in the statistics, and the initial value of the sum of squared amplitude deviations of the back reference sub-window is SumDevSqu. B (m), the standard deviation of the background amplitude is Where m is a positive integer, and 1≤m≤M.

[0025] Step 5.2: When 2≤n≤N p +N r At this time, the front sub-window is not included in the statistics, and the back sub-window slides along the distance direction and the sum of squares of the amplitude deviation of the back sub-window is calculated as SumDevSqu B (m), the standard deviation of the background amplitude is

[0026] Step 5.3: When n = N p +N r When +1, the initial value of the sum of squared amplitude deviations of the front reference sub-window is SumDevSqu. F (m), the rear sub-window slides along the distance direction and the sum of squares of the magnitude deviation of the rear sub-window is calculated as SumDevSqu B (m), the sum of squared amplitude deviations of the corresponding sub-windows on the front and back sides is taken as the maximum value MaxDevSqu(m)=max[SumDevSqu F (m),SumDevSqu B (m)], the standard deviation of the background amplitude is

[0027] Step 5.4: When N p +N r +2≤n≤N-(N p +N r When ), the front and rear sub-windows slide simultaneously along the distance direction, and the sum of squares of the amplitude deviation of the front sub-window is calculated as SumDevSqu. F (m), the sum of squared amplitude deviations of the rear sub-window is SumDevSqu B (m), the sum of squared amplitude deviations of the corresponding sub-windows on the front and back sides is taken as the maximum value MaxDevSqu(m)=max[SumDevSqu F (m),SumDevSqu B (m)], the standard deviation of the background amplitude is

[0028] Step 5.5: When N-(N p +N r When 1 ≤ n ≤ N, the front sub-window slides along the distance direction and the sum of squares of the amplitude deviation of the front sub-window is calculated as SumDevSqu. F (m), the rear sub-window is not included in the statistics, and the standard deviation of the background amplitude is...

[0029] Step 6: Background adaptive threshold detection;

[0030] Threshold detection is performed on each distance sampling unit sequentially to obtain the detection result CFAR_Out. The specific steps are as follows:

[0031] Step 6.1: Calculate the detection threshold of the distance sampling unit. Here, α is the detection threshold factor, which can be derived from the clutter amplitude probability density distribution model.

[0032] Step 6.2: Compare the echo amplitude of the distance sampling unit with the detection threshold to obtain the detection result.

[0033] Step 7: Receive the next radar pulse echo data, then proceed to step 2.

[0034] Compared to traditional technologies, the advantages of this invention include:

[0035] 1. This invention addresses the problem that traditional GOCA-CFAR detectors struggle to effectively handle non-uniform spatial clutter backgrounds. It proposes a dual-parameter background estimation method that, by estimating the background amplitude and standard deviation within the reference window in real time, can better adapt to non-uniform spatial clutter backgrounds and maintain consistency in the false alarm residue of spatial detection.

[0036] 2. This invention addresses the target occlusion problem faced by traditional two-parameter CFAR detectors in dense target scenes by proposing a fast parameter estimation method based on approximate sorting. By dividing the reference sub-windows and approximating the parameter estimation results of the reference sub-windows, more accurate parameter estimation results can be obtained in dense target scenes, effectively avoiding the influence of target echoes within the reference window.

[0037] 3. This invention addresses the problem of high computational complexity of standard deviation calculation faced by traditional two-parameter CFAR detectors in engineering applications. It proposes a standard deviation estimation method based on multiple traversal solutions, which simplifies the computational complexity of background amplitude standard deviation, significantly improves computational efficiency, and is more suitable for practical engineering applications. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the preferred processing flow of the present invention.

[0039] Figure 2 This is a schematic diagram of the process for estimating the mean background amplitude in this invention.

[0040] Figure 3 This is a schematic diagram of the process for estimating the standard deviation of the background amplitude in this invention.

[0041] Figure 4 This is a schematic diagram of the background adaptive threshold detection process in this invention.

[0042] Figure 5 The image shown is a video image of the radar measured echo data in a specific embodiment of the present invention.

[0043] Figure 6 The image shows the detection results of a conventional GOCA-CFAR detector on radar measured echo data in a specific embodiment of the present invention.

[0044] Figure 7 This is the detection result of the dual-parameter CFAR detector in a specific embodiment of the present invention on the measured radar echo data.

[0045] Figure 8 This is a video image of a target area with dense radar echo data in a specific embodiment of the present invention.

[0046] Figure 9 This is the detection result of a conventional dual-parameter CFAR detector on a target area with dense radar echo data in a specific embodiment of the present invention.

[0047] Figure 10 This is the detection result of a dual-parameter CFAR detector in a specific embodiment of the present invention for a target area with dense radar echo data.

[0048] Figure 11 This is a graph showing the echo amplitude of the 306th pulse of the radar measured echo data in a specific embodiment of the present invention.

[0049] Figure 12 This is a comparison chart showing the background amplitude mean estimation results of the conventional dual-parameter CFAR detector and the dual-parameter CFAR detector involved in this invention for the 306th pulse of the radar measured echo data in a specific embodiment of the present invention.

[0050] Figure 13 This is a comparison chart showing the estimation results of the background amplitude standard deviation of the 306th pulse of the radar measured echo data by a traditional two-parameter CFAR detector and the two-parameter CFAR detector involved in this invention in a specific embodiment of the present invention. Detailed Implementation

[0051] This invention proposes a fast estimation method for constant false alarm rate (CFAR) detection using a dual-parameter approach with deviation summation. A preferred processing flow diagram is shown below. Figure 1 As shown, the specific implementation of the embodiment is as follows:

[0052] Step 1: Process parameter settings;

[0053] The parameters are set as follows: the length of the single-sided protective window is N. p There are N distance sampling units, and the length of the reference window on one side is N. r There are 1 distance sampling unit, and the number of reference sub-windows divided by the single-sided reference window is M, with a single-sided reference sub-window length of N. sThere are distance sampling units, where M can be sampled by N. p Divisible by N s =N p / M.

[0054] Step 2: Receive radar echo data;

[0055] The logarithmic amplitude data of the radar's i-th pulse echo is A i (n), where n is a positive integer and 1≤n≤N, and N is the number of distance sampling units for the i-th pulse;

[0056] Step 3: Estimate the mean background amplitude

[0057] Starting from the first distance sampling unit of the pulse, each distance sampling unit is traversed sequentially along the distance direction using a sliding window method, and the average background amplitude of each distance sampling unit is calculated. The specific steps are as follows:

[0058] Step 3.1: When n=1, the front sub-window does not participate in the statistics, and the back reference sub-window is assigned an initial value.

[0059] Step 3.1.1: The amplitude and initial value of the rear reference sub-window are: Where m is a positive integer, and 1≤m≤M, n s =N p +(m-1)*N s n e =N p +1+m*N s .

[0060] Step 3.1.2: Calculate the mean background amplitude.

[0061] Step 3.2: When 2≤n≤N p +N r When the data is in the range of 'time', the front sub-window is not included in the statistics, and the back sub-window slides along the distance direction.

[0062] Step 3.2.1: The sum of the amplitudes of the rear sub-windows is

[0063] SumAmp B (m)=SumAmp B (m)-A i ((m-1)·N s +n+N p )

[0064] Step 3.2.2: Calculate the mean background amplitude.

[0065] Step 3.3: When n = Np +N r When +1 is applied, the front reference sub-window is initialized, and the rear sub-window slides along the distance direction.

[0066] Step 3.3.1: The amplitude and initial value of the front reference sub-window are: Where, n s =N r +1-m*N s n e =N r -(m-1)*N s .

[0067] Step 3.3.2: The sum of the amplitudes of the rear sub-windows is

[0068] SumAmp B (m)=SumAmp B (m)-A i ((m-1)·N s +n+N p )+A i ((m-1)·N s +n+N p +N s )

[0069] Step 3.3.3: Take the larger value of the amplitude sum of the corresponding sub-windows on the front and back sides.

[0070] MaxSumAmp(m) = max[SumAmp F (m),SumAmp B (m)]

[0071] Step 3.3.4: Calculate the mean background amplitude.

[0072]

[0073] Step 3.4: When N p +N r +2≤n≤N-(N p +N r When ), the front and rear side sub-windows slide simultaneously along the distance direction.

[0074] Step 3.4.1: The sum of the amplitudes of the front sub-windows is

[0075] SumAmp F (m)=SumAmp F (m)-A i (nN p -m·N s -1)+A i (nN p -(m-1)·Ns -1)

[0076] Step 3.4.2: The sum of the amplitudes of the rear sub-windows is

[0077] SumAmp B (m)=SumAmp B (m)-A i ((m-1)·N s +n+N p )+A i ((m-1)·N s +n+N p +N s )

[0078] Step 3.4.3: Take the larger value of the amplitude sum of the corresponding sub-windows on the front and back sides.

[0079] MaxSumAmp(m) = max[SumAmp F (m),SumAmp B (m)]

[0080] Step 3.4.4: Calculate the mean background amplitude.

[0081]

[0082] Step 3.5: When N-(N p +N r When 1 ≤ n ≤ N, the front sub-window slides along the distance direction, and the back sub-window does not participate in the statistics.

[0083] Step 3.5.1: The sum of the amplitudes of the front sub-windows is

[0084] SumAmp F (m)=SumAmp F (m)-A i (nN p -m·N s -1)+A i (nN p -(m-1)·N s -1)

[0085] Step 3.5.2: Calculate the mean background amplitude.

[0086] Step 4: Calculate the squared amplitude deviation value DevSqu;

[0087] Calculate the squared deviation of each echo amplitude from the mean of its background amplitude.

[0088] Step 5: Estimate the standard deviation of the background amplitude σ;

[0089] Starting from the first distance sampling unit of the pulse, each distance sampling unit is traversed sequentially along the distance direction using a sliding window method. The standard deviation of the background amplitude σ(n) of each distance sampling unit is calculated. The specific steps are as follows:

[0090] Step 5.1: When n=1, the front sub-window does not participate in the statistics, and the back reference sub-window is assigned an initial value.

[0091] Step 5.1.1: The initial value of the sum of squares of the amplitude deviation of the rear reference sub-window is: Where, n s =N p +(m-1)*N s n e =N p +1+m*N s .

[0092] Step 5.1.2: Calculate the standard deviation of the background amplitude.

[0093] Step 5.2: When 2≤n≤N p +N r When the data is in the range of 'time', the front sub-window is not included in the statistics, and the back sub-window slides along the distance direction.

[0094] Step 5.2.1: The sum of squares of the amplitude deviation of the rear sub-window is

[0095] SumDevSqu B (m)=SumDevSqu B (m)-DevSqu((m-1)·N s +n+N p )

[0096] +DevSqu((m-1)·N s +n+N p +N s )

[0097] Step 5.2.2: Calculate the standard deviation of the background amplitude.

[0098] Step 5.3: When n = N p +N r When +1 is applied, the front reference sub-window is initialized, and the rear sub-window slides along the distance direction.

[0099] Step 5.3.1: The initial value of the sum of squares of the amplitude deviation of the front reference sub-window is: Where, n s =N r +1-m*N s ne =N r -(m-1)*N s .

[0100] Step 5.3.2: The sum of squares of the amplitude deviation of the rear sub-window is

[0101] SumDevSqu B (m)=SumDevSqu B (m)-DevSqu((m-1)·N s +n+N p )

[0102] +DevSqu((m-1)·N s +n+N p +N s )

[0103] Step 5.3.3: Take the larger of the sum of squared amplitude deviations of corresponding sub-windows on the front and back sides.

[0104] MaxDevSqu(m) = max[SumDevSqu F (m),SumDevSqu B (m)]

[0105] Step 5.3.4: Calculate the standard deviation of the background amplitude.

[0106]

[0107] Step 5.4: When N p +N r +2≤n≤N-(N p +N r When ), the front and rear side sub-windows slide simultaneously along the distance direction.

[0108] Step 5.4.1: The sum of squares of the amplitude deviation of the front sub-window is

[0109] SumDevSqu F (m)=SumDevSqu F (m)-DevSqu(nN p -m·N s -1)

[0110] +DevSqu(nN p -(m-1)·N s -1)

[0111] Step 5.4.2: The sum of squares of the amplitude deviation of the rear sub-window is

[0112] SumDevSqu B(m)=SumDevSqu B (m)-DevSqu((m-1)·N s +n+N p )

[0113] +DevSqu((m-1)·N s +n+N p +N s )

[0114] Step 5.4.3: Take the larger of the sum of squared amplitude deviations of corresponding sub-windows on the front and back sides.

[0115] MaxDevSqu(m) = max[SumDevSqu F (m),SumDevSqu B (m)]

[0116] Step 5.4.4: Calculate the standard deviation of the background amplitude.

[0117]

[0118] Step 5.5: When N-(N p +N r When 1 ≤ n ≤ N, the front sub-window slides along the distance direction, and the back sub-window does not participate in the statistics.

[0119] Step 5.5.1: The sum of squares of the amplitude deviation of the front sub-window is

[0120] SumDevSqu F (m)=SumDevSqu F (m)-DevSqu(nN p -m·N s -1)

[0121] +DevSqu(nN p -(m-1)·N s -1)

[0122] Step 5.5.2: Calculate the standard deviation of the background amplitude.

[0123]

[0124] Step 6: Background adaptive threshold detection;

[0125] Threshold detection is performed on each distance sampling unit sequentially to obtain the detection result CFAR_Out. The specific steps are as follows:

[0126] Step 6.1: Calculate the detection threshold of the distance sampling unit. Here, α is the detection threshold factor, which can be derived from the clutter amplitude probability density distribution model.

[0127] Step 6.2: Compare the echo amplitude of the distance sampling unit with the detection threshold to obtain the detection result CFAR_Out, as follows:

[0128]

[0129] Step 7: Receive the next radar pulse echo data, then proceed to step 2. Specific implementation examples:

[0131] Simulation analysis is performed using radar data as an example. In the simulation: the total number of pulses in the measured radar echo data is 900, and the number of range sampling units in each pulse is N = 2000. The parameters for the dual-parameter CFAR detector involved in this invention are set as follows: the length of the single-sided protective window is N. p =5 distance sampling units, with a single-sided reference window length of N r = 20 distance sampling units, the number of reference sub-windows divided by a single-sided reference window is M = 4, and the length of a single-sided reference sub-window is N. s =5 distance sampling units. For comparison, the single-sided guard window length and single-sided reference window length of the conventional GOCA-CFAR detector and the conventional two-parameter CFAR detector are the same as above. Furthermore, the standard deviation of the noise background amplitude is known to be σ. n =1.5, the background mean estimated by the traditional GOCA-CFAR detector is The detection threshold of the traditional GOCA-CFAR detector is The background mean estimated by a traditional two-parameter CFAR detector is The estimated background mean is σ c The detection threshold of a traditional two-parameter CFAR detector is The background mean estimated by the dual-parameter CFAR detector involved in this invention is The estimated background mean is σ d The detection threshold of the dual-parameter CFAR detector involved in this invention is: The detection threshold factor for all three methods is α = 4.

[0132] Figure 5 This is a video image of the radar's measured echo data. Figure 6 This refers to the detection results of a traditional GOCA-CFAR detector on radar measured echo data. Figure 7 The detection results of the dual-parameter CFAR detector of this invention on radar measured echo data are provided by... Figure 6 and Figure 7It is understood that, compared with the traditional GOCA-CFAR detector, the dual-parameter CFAR detector of this invention can better maintain the spatial consistency of residual false alarms in the context of spatially non-uniform clutter.

[0133] Figure 8 This is a video image of a target area with dense radar echo data. Figure 9 This represents the detection results of a traditional two-parameter CFAR detector for target areas with dense radar echo data. Figure 10 The present invention relates to the detection results of a dual-parameter CFAR detector on a target area with dense radar echo data, by means of... Figure 9 and Figure 10 The comparison shows that the dual-parameter CFAR detector involved in this invention can better avoid target occlusion effects in dense target scenes compared with the traditional dual-parameter CFAR detector.

[0134] Figure 11 This is a graph showing the echo amplitude of the 306th pulse from the radar's measured echo data. Figure 12 This image shows a comparison of the background amplitude mean estimation results of the 306th pulse of the radar measured echo data using a traditional two-parameter CFAR detector and the two-parameter CFAR detector of the present invention. Figure 13 This is a comparison chart showing the estimation results of the background amplitude standard deviation of the 306th pulse of the radar measured echo data by a traditional two-parameter CFAR detector and the two-parameter CFAR detector of this invention. Figure 12 and Figure 13 It can be seen that the dual-parameter CFAR detector involved in this invention has basically the same background mean and standard deviation estimation results as the traditional dual-parameter CFAR detector, and can better avoid the influence of target echo on the background mean and standard deviation estimation.

Claims

1. A fast estimation method for constant false alarm rate (CFAR) detection using a dual-parameter approach with deviation summation, characterized in that: Step 1: Processing parameter settings; set the parameters as follows: the length of the single-sided protective window is N. p There are N distance sampling units, and the length of the reference window on one side is N. r There are 1 distance sampling unit, and the number of reference sub-windows divided by the single-sided reference window is M, with a single-sided reference sub-window length of N. s There are distance sampling units, where M can be sampled by N. p Divisible by N s =N p / M; Step 2: Receive radar echo data; the logarithmic amplitude data of the i-th radar pulse echo is A. i (n), where n is a positive integer and 1≤n≤N, and N is the number of distance sampling units for the i-th pulse; Step 3: Estimate the mean background amplitude Starting from the first distance sampling unit of the pulse, each distance sampling unit is traversed sequentially along the distance direction using a sliding window method, and the average background amplitude of each distance sampling unit is calculated. This includes: firstly, calculating the amplitude and initial value SumAmp of the M reference sub-windows on the front side. F (m) and the amplitudes of the M reference sub-windows on the rear side, and SumAmp B (m), then calculate the amplitude of the corresponding sub-windows on the front and back sides and take the larger value as MaxSumAmp(m)=max[SumAmp F (m),SumAmp B (m)], and finally the mean background amplitude is obtained. Where m is a positive integer, and 1≤m≤M; Step 4: Calculate the squared amplitude deviation value DevSqu; calculate the squared deviation of each echo amplitude from the mean background amplitude value. Step 5: Estimate the background amplitude standard deviation σ; Starting from the first distance sampling unit of the pulse, use a sliding window to traverse each distance sampling unit sequentially along the distance direction, and calculate the background amplitude standard deviation σ(n) for each distance sampling unit; This includes: firstly, calculating the sum of squared amplitude deviations of the preceding M reference sub-windows, SumDevSqu F (m) and the sum of squared amplitude deviations of the M reference sub-windows behind it, SumDevSqu B (m), then calculate the sum of squared amplitude deviations of the corresponding sub-windows on the front and back sides, taking the larger value as MaxDevSqu(m)=max[SumDevSqu F (m),SumDevSqu B (m)], and finally the standard deviation of the background amplitude is obtained as Step 6: Background Adaptive Threshold Detection; Threshold detection is performed on each distance sampling unit sequentially to obtain the detection result CFAR_Out; including: firstly calculating the detection threshold value of the distance sampling unit. Wherein, α is the detection threshold factor, which can be derived from the clutter amplitude probability density distribution model. Then, the echo amplitude of this distance sampling unit is compared with the detection threshold to obtain the detection result. Step 7: Receive the next radar pulse echo data, then proceed to step 2.

Images