Target detection method in ground clutter based on global statistical characteristics of clutter

Abstract

The present invention proposes a target detection method based on the global statistical characteristics of clutter, comprising the following steps: step 1, using histogram statistics amplitude value for the range Doppler amplitude value, and segmenting by a method based on one-dimensional statistical grayscale histogram; step 2, calculating the expected number of target detection results; step 3, clutter global background statistics: including calculating the range clutter width, calculating the azimuth Doppler width and the number of Doppler channels, and calculating the global statistical mean of the clutter background; step 4, determining the detection threshold according to the results of steps 1 to 3, and performing target detection according to the detection threshold. The method of the present invention can be effectively applied to the real-time requirements of high-frequency and high-resolution radar target detection and tracking signal processing of high-speed moving carriers, can effectively improve the effectiveness, accuracy and reliability of radar target detection, and has significant engineering application value.

Description

Technical Field

[0001] The present invention relates to the field of radar signal processing technology, and in particular to a method for realizing target detection in a ground clutter background based on global statistical characteristics of clutter. Background Art

[0002] The radar detection and guidance system has the advantages of working around the clock and in all weather conditions. It can obtain real-time images of the carrier's flight track or the target area and nearby typical landforms, identify, confirm and track suspected targets, and give control signals required for guidance. At present, radar guidance technology has become the development trend of precision guidance technology, and plays a very important role in target strikes in multi-target and complex environments. However, with a large amount of natural interference (such as ground / sea background, rain and snow, buildings, etc.) and human interference in the target scene, the information obtained by the radar contains a large number of irrelevant targets, interference, decoys, clutter, etc. The complex target environment will seriously affect the accuracy of target detection and tracking.

[0003] How to accurately detect the target by radar detector is the key issue to ensure accurate target tracking. The general processing method is to perform pulse compression and coherent or non-coherent accumulation processing on the received original echo, and use the local CFAR detection method to condense and measure the output target point information to obtain the accurate position and velocity information of the target. The basic assumption of the CFAR detection scheme is that the reference unit is pure clutter and the clutter of the reference unit has the same statistical characteristics as the clutter of the detection unit, and the target has a strong contrast with the background. When the target background is complex, the probability of missed detection or false alarm will increase, thereby affecting the accurate tracking of the target. Therefore, the target detection method based on the global statistical characteristics of clutter has important engineering application value for solving the problem of target detection under complex backgrounds. Summary of the invention

[0004] The present invention provides a method for detecting targets in a ground clutter background based on global statistical characteristics of clutter, so as to solve the problem of low accuracy of target detection in a complex background existing in the prior art.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A target detection method based on global statistical characteristics of clutter includes the following steps:

[0007] Step 1, using histogram statistics amplitude value for the range Doppler amplitude value, and using a method based on one-dimensional statistical grayscale histogram for segmentation;

[0008] Step 2, calculate the expected number N of target detection results;

[0009] Step 3, clutter global background statistics: including calculation of range clutter width, calculation of azimuth Doppler width and Doppler channel number, and calculation of clutter background global statistical mean;

[0010] Step 4: determine the detection threshold according to the results of step 1, step 2 and step 3, and perform target detection according to the detection threshold.

[0011] Furthermore, the step 1 specifically includes the following sub-steps:

[0012] Step 11, perform histogram statistics on the range Doppler amplitude values ​​obtained by radar echo processing to obtain the distribution characteristics of multiple amplitude values, and record the maximum amplitude value f max and the minimum amplitude f min , use formula 1 to calculate the histogram quantization interval:

[0013] Δf=(f max -f min ) / 256-1 (1)

[0014] Step 12, traverse all range Doppler amplitude values, and use formula 2 to quantize the range Doppler amplitude value f corresponding to each amplitude value unit to obtain the quantized amplitude value f corresponding to each amplitude value unit N :

[0015] f N =f / Δf (2)

[0016] Among them, f N is the quantized amplitude value, ranging from 0 to 255; f is the range Doppler amplitude value;

[0017] Step 13: quantize the amplitude value f N Perform statistics and convert the quantized amplitude value f N The numbers of 0-255 are stored in the array Num respectively, and the values ​​in the array represent the number of samples with an amplitude of 0-255.

[0018] Furthermore, the expected number of target detection results in step 2 is calculated using formula 3:

[0019] N=k*3*(L*cosθ / (C / 2 / F s )) (3)

[0020] Where N is the expected number of target detection results, k is the broadening coefficient, which is 1 without interference and 1.2 with interference; L is the target length; is the ground-rubbing angle, C is the speed of light; F s is the system sampling rate, and the number of Doppler units across the target in azimuth is 3.

[0021] Furthermore, the step 3 specifically includes the following sub-steps:

[0022] Step 31, calculate the range clutter width, the specific operation is as follows:

[0023] Step 311, calculate the distance unit corresponding to the sampling of the beam center:

[0024] N R =(R c -R0) / Fs (4)

[0025] Among them, R c is the center slope distance, R0 is the sampling gate start, N R The distance unit corresponding to the sample representing the beam center;

[0026] Step 312, according to the incident angle of the beam, the beam range width and the platform height, the number of sampling points M in the range direction is calculated using Formula 5:

[0027] M=(H*tan(α+β / 2)-H*tan(α-β / 2))*Sin(α) / Fs (5)

[0028] Where M represents the number of sampling points in the range direction, which is used to display the effective beam range, point N is the location of the radar platform, P is the beam center, point A and point B are the leading and trailing edges of the beam range projection on the ground, H is the platform height, α and β are the beam incidence angle and beam range width, respectively;

[0029] Step 313: The distance unit N corresponding to the sampling of the beam center obtained in step 311 is R The starting and ending coordinates of the distance unit are calculated by the following formula: R –M / 2, N R +M / 2;

[0030] Step 32, calculating the azimuth Doppler width and the number of Doppler channels, the specific operation is as follows:

[0031] The Doppler values ​​of point C and point D are calculated using equations 6 and 7 respectively, and the number of Doppler channels corresponding to the clutter is calculated using equation 8:

[0032] f c =2 / λ*(V x *X c -V y *Y c +V z *Z c ) / R c (6)

[0033] f d=2 / λ*(V x *X d -V y *Y d +V z *Z d ) / R c (7)

[0034] N f =(f d -f c ) / Prf*N p (8)

[0035] Among them, point C and point D are the beam azimuth edges, and their coordinates are (X c ,Y c ,Z c )、(X d ,Y d ,Z d ), θ is the azimuth beam width, V x 、V y 、V z are the three-dimensional speeds of the platform, Prf is the repetition frequency, N p is the azimuth channel number, the starting position of the clutter channel is N P / 2–N f , N P / 2–1+N f / 2; λ is the wavelength;

[0036] Step 33, calculate the global statistical mean of the clutter background, the specific operation is as follows:

[0037] Step 331, for the range Doppler amplitude value N P / 2–N f ~N P / 2–1+N f / 2 channel and N R –M / 2~N R +M / 2 distance unit range amplitude sum, get the amplitude sum, calculate the ratio of the amplitude sum to the number of samples to get the sample amplitude mean f mean ;

[0038] Step 332, traverse the range Doppler amplitude values ​​again and filter out the amplitude values ​​greater than f mean *k max Strong target information and less than f mean *k min The weak noise information of the target information and the amplitude value corresponding to the weak noise information are expressed by f mean Substitute to get the sample data after filtering, k max and k min Take 2.0 and 0.6 respectively;

[0039] Step 333: Take the sample data obtained in step 332 as the processing object, execute step 331, and obtain a new sample amplitude mean f′ mean , which is used as the final global statistical mean of the clutter background.

[0040] Furthermore, the step 4 specifically includes the following sub-steps:

[0041] Step 41, according to the expected number N of target detection results obtained in step 2, the array Num obtained in step 13 is statistically analyzed to determine the number of stop bits n, and the detection threshold Th1 is calculated according to the number of stop bits n;

[0042] Step 42, set the clutter background detection threshold Th2 = f' mean +13dB, f′ mean The global statistical mean of the clutter background is obtained for step 3; the final global threshold for target detection is Th=max(Th1, Th2);

[0043] Step 43, traverse the range Doppler amplitude values ​​described in step 1, save the target information that is greater than the global target detection threshold Th, and the target detection process is completed.

[0044] Furthermore, the specific operations of step 41 are as follows:

[0045] Step 411, add the values ​​of the 255th and 254th bits in the array Num to obtain m;

[0046] Step 412, determine whether m=N is established, if yes, record the value of n and execute step 414, otherwise, execute step 413;

[0047] Step 413, continue to read the value of the next bit in the array Num, add it to m and assign it to m, and execute step 413;

[0048] Step 414, calculate the detection threshold Th1 = n*256+f min .

[0049] The target detection method based on the global statistical characteristics of clutter in the embodiment of the present invention adopts a histogram to count the amplitude value distribution of the range Doppler result, and sets the detection threshold in combination with the expected number of detection results calculated based on the target prior information; parameterizes the calculation of the clutter spectrum width to achieve accurate global statistics of the clutter background, and comprehensively sets the detection threshold to complete target detection under complex background.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. The expected number of detection results is calculated by using target prior information, system operating parameters, and carrier inertial navigation information, and the detection threshold is calculated by combining the statistical distribution of the range Doppler amplitude histogram to estimate the number of target detection results more accurately.

[0052] 2. According to the system operating parameters and carrier inertial navigation parameters, the effective area of ​​clutter background is automatically calculated, and the global iterative statistics of the clutter background area are performed to eliminate strong target information and weak noise, and the background of the ground clutter area is calculated more accurately, which is beneficial to the target detection efficiency under complex background.

[0053] 3. The detection threshold is adaptively calculated according to the characteristics of the target and the clutter characteristics, which effectively avoids the unreasonable setting of the detection threshold of the traditional method, resulting in missed target detection or excessive false alarm. It solves the problems of missed target detection, split surface target detection results or missing target information in the traditional local CFAR detection under complex terrain background, and also significantly improves the real-time performance of the system processing.

[0054] In summary, the present invention adopts adaptive parameterized calculation to adapt to the influence of changes in the environment or target area characteristics during the tracking process in radar target detection. The method is easy to implement and can be effectively applied to the real-time requirements of high-frequency, high-resolution radar target detection and tracking signal processing of high-speed moving carriers. It can effectively improve the effectiveness, accuracy and reliability of radar target detection and has significant engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart of the target detection method based on global statistical characteristics of clutter of the present invention;

[0056] Figure 2 This is the global statistical flow chart of clutter background;

[0057] Figure 3 It is the distance profile of the spatial geometric relationship between the target and the platform;

[0058] Figure 4 It is the azimuth profile of the spatial geometric relationship between the target and the platform;

[0059] Figure 5 This is the target detection flow chart;

[0060] Figure 6 This is a comparison diagram of the target detection effect obtained by using the traditional local CFAR and the method of the present invention under the complex background of the embodiment of the present invention. Among them, (a) and (b) are the detection results of the traditional method and the method of the present invention respectively. DETAILED DESCRIPTION

[0061] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose, the following is a further detailed description in conjunction with the accompanying drawings and preferred embodiments.

[0062] like Figure 1 As shown, the target detection method based on the global statistical characteristics of clutter of the present invention takes the range Doppler amplitude value obtained after conventional pulse compression, motion compensation and MTD processing of radar echo as the processing object. It includes the following steps:

[0063] Step 1, using histogram statistics amplitude value for the range Doppler amplitude value, and using a method based on one-dimensional statistical grayscale histogram for segmentation, specifically including the following sub-steps:

[0064] Step 11, perform histogram statistics on the range Doppler amplitude values ​​obtained by radar echo processing to obtain the distribution characteristics of multiple amplitude values, and record the maximum amplitude value f max and the minimum amplitude f min , use formula 1 to calculate the histogram quantization interval:

[0065] Δf=(f max -f min ) / 256-1 (1)

[0066] Step 12, traverse all range Doppler amplitude values, and use formula 2 to quantize the range Doppler amplitude value f corresponding to each amplitude value unit to obtain the quantized amplitude value f corresponding to each amplitude value unit N :

[0067] f N =f / Δf (2)

[0068] Among them, f N is the quantized amplitude value, ranging from 0 to 255; f is the range Doppler amplitude value;

[0069] Step 13: quantize the amplitude value f N Perform statistics and convert the quantized amplitude value f N The numbers of 0-255 are stored in the array Num (size is 256), and the values ​​in the array represent the number of samples with an amplitude of 0-255.

[0070] Step 2: Calculate the expected number of target detection results N. The calculation formula is as follows:

[0071] N=k*3*(L*cosθ / (C / 2 / F s )) (3)

[0072] Where N is the expected number of target detection results, k is the broadening coefficient, which is 1 without interference and 1.2 with interference; L is the target length; is the ground-rubbing angle, C is the speed of light; F s is the system sampling rate, and the number of Doppler units across the target in azimuth is 3.

[0073] This step automatically calculates the expected number of detection results through parameters to ensure that the number of target detection results remains within a reasonable range, avoiding the problem of too many or too few results caused by unreasonable detection threshold settings.

[0074] Step 3, clutter global background statistics: Figure 2 As shown, it includes calculating the range clutter width, calculating the azimuth Doppler width and the number of Doppler channels, and calculating the global statistical mean of the clutter background.

[0075] In order to accurately realize the global background statistics of clutter, the clutter area is calculated according to the system inertial navigation parameters, including the calculation of range clutter width, azimuth Doppler width and global statistical mean of clutter background.

[0076] Step 31, calculate the range clutter width, as follows:

[0077] Step 311, calculate the distance unit corresponding to the sampling of the beam center:

[0078] N R =(R c -R0) / Fs (4)

[0079] Where R c is the center slope distance, R0 is the start of the sampling gate, N R The distance unit corresponding to the sample representing the beam center;

[0080] Step 312, according to the incident angle of the beam, the beam range width and the platform height, the number of sampling points M in the range direction is calculated using Formula 5:

[0081] M=(H*tan(α+β / 2)-H*tan(α-β / 2))*Sin(α) / Fs (5)

[0082] Where M represents the number of sampling points in the range direction, which is used to display the effective beam range. Its geometric relationship is as follows: Figure 3 As shown in the figure, point N is the position of the radar platform, P is the beam center, point A and point B are the front and rear edges of the beam range projection on the ground, H is the platform height, α and β are the beam incidence angle and beam range width, respectively;

[0083] Step 313: The distance unit N corresponding to the sampling of the beam center obtained in step 311 is R The starting and ending coordinates of the distance unit are calculated by the following formula: R –M / 2, N R+M / 2.

[0084] Step 32, calculating the azimuth Doppler width and the number of Doppler channels, specifically:

[0085] See also Figure 4 , assuming that point C and point D are the beam azimuth edges, use formulas 6 and 7 to calculate the Doppler values ​​of point C and point D respectively, and use formula 8 to calculate the number of Doppler channels corresponding to the clutter:

[0086] f c =2 / λ*(V x *X c -V y *Y c +V z *Z c ) / R c (6)

[0087] f d =2 / λ*(V x *X d -V y *Y d +V z *Z d ) / R c (7)

[0088] N f =(f d -f c ) / Prf*N p (8)

[0089] Among them, point C and point D are the beam azimuth edges, and their coordinates are (X c ,Y c ,Z c )、(X d ,Y d ,Z d ), θ is the azimuth beam width, V x 、V y 、V z are the three-dimensional velocities of the platform, the angles between the resultant velocity V and NC and ND are α1 and α2 respectively; Prf is the repetition frequency, N p is the azimuth channel number, the starting position of the clutter channel is N P / 2–N f , N P / 2–1+N f / 2; λ is the wavelength;

[0090] Step 33, calculate the global statistical mean of the clutter background, the specific operation is as follows:

[0091] Step 331, for the range Doppler amplitude value NP / 2–N f ~N P / 2–1+N f / 2 channel and N R –M / 2~N R +M / 2 distance unit range amplitude sum, get the amplitude sum, calculate the ratio of the amplitude sum to the number of samples to get the sample amplitude mean f mean .

[0092] Step 332, traverse the range Doppler amplitude values ​​again and filter out the amplitude values ​​greater than f mean *k max Strong target information and less than f mean *k min The weak noise information of the target information and the amplitude value corresponding to the weak noise information are expressed by f mean Substitute to get the sample data after filtering, k max and k min Take 2.0 and 0.6 respectively.

[0093] Step 333: Take the sample data obtained in step 332 as the processing object, execute step 331, and obtain a new sample amplitude mean f′ mean , which is used as the final global statistical mean of the clutter background.

[0094] Step 4: Determine the detection threshold based on the results of steps 1, 2, and 3, and perform target detection based on the detection threshold. Figure 5 As shown, the specific operations are as follows:

[0095] Step 41 , according to the expected number N of target detection results obtained in step 2 , the array Num obtained in step 1 is statistically analyzed to determine the number of stop bits n, and the detection threshold Th1 is calculated according to the number of stop bits n.

[0096] The specific operations are as follows:

[0097] Step 411, add the values ​​of the 255th and 254th bits in the array Num to obtain m;

[0098] Step 412, determine whether m=N is established, if yes, record the value of n and execute step 414, otherwise, execute step 413;

[0099] Step 413, continue to read the value of the next bit in the array Num, add it to m and assign it to m, and execute step 413;

[0100] Step 414, calculate the detection threshold Th1 = n*256+f min ;

[0101] This detection threshold ensures that the expected number of detection results is N.

[0102] Step 42: According to the empirical value, the target is required to be at least 13 dB higher than the clutter background threshold, so the clutter background detection threshold Th2 = f′ is set. mean +13dB, f′ mean The global statistical mean of the clutter background is obtained for step 3. Taking into account the expected number of target detection results and the clutter background threshold, the final target detection global threshold is selected as Th=max(Th1, Th2).

[0103] Step 43, traverse the range Doppler amplitude values ​​described in step 1, save the target information that is greater than the global target detection threshold Th, and the target detection process is completed.

[0104] Figure 6 This is a comparison chart of target detection effects obtained by using traditional local CFAR and the method of the present invention under a complex ground background. The target is a ship model with a complex ground background. An airborne platform is used for flight verification. Using the detection method of the present invention, the processing time of 512*4096 data is reduced from the original 60ms to 15ms, and the characteristics of the target are more complete and rich.

[0105] Compared with the traditional CFAR detection sliding window statistical background and calculation detection threshold, the method of the present invention has a fast calculation speed and greatly improves the detection efficiency. At the same time, the background of the ground clutter area is more accurately calculated to solve the problems of traditional CFAR detection in complex ground background, easy target leakage, surface target detection result splitting or target information missing, etc., and has a good effect in achieving effective target detection in complex target background.

Claims

1. A target detection method based on global statistical characteristics of clutter, characterized in that: The steps include: Step 1, using histogram statistics amplitude value for the range Doppler amplitude value, and using a method based on one-dimensional statistical grayscale histogram for segmentation; Step 2, calculate the expected number N of target detection results; Step 3, clutter global background statistics: including calculation of range clutter width, calculation of azimuth Doppler width and Doppler channel number, and calculation of clutter background global statistical mean; Step 4, determining the detection threshold according to the results of step 1, step 2 and step 3, and performing target detection according to the detection threshold, specifically includes the following sub-steps: Step 41, according to the expected number N of target detection results obtained in step 2, the array Num obtained in step 13 is statistically determined to determine the number of stop bits n, and the detection threshold Th1 is calculated according to the number of stop bits n. The specific operations are as follows: Step 411, add the values ​​of the 255th and 254th bits in the array Num to obtain m; Step 412, determine whether m=N is true, if so, record the value of n and execute step 414, otherwise, execute step 413; Step 413, continue to read the value of the next bit in the array Num, add it to m and assign it to m, and execute step 413; Step 414, calculate the detection threshold Th1=n*256+f min ; Step 42, set the clutter background detection threshold Th2 = f m ' ean +13dB, f m ' ean The global statistical mean of the clutter background is obtained for step 3; the final global threshold for target detection is Th=max(Th1, Th2); Step 43, traverse the range Doppler amplitude values ​​described in step 1, save the target information that is greater than the global target detection threshold Th, and the target detection process is completed.

2. The target detection method based on global statistical characteristics of clutter according to claim 1, characterized in that: The step 1 specifically includes the following sub-steps: Step 11, perform histogram statistics on the range Doppler amplitude values ​​obtained by radar echo processing to obtain the distribution characteristics of multiple amplitude values, and record the maximum amplitude value f max and the minimum amplitude f min , use formula 1 to calculate the histogram quantization interval: Δf=(f max -f min ) / 256-1 (1) Step 12, traverse all range Doppler amplitude values, and use formula 2 to quantize the range Doppler amplitude value f corresponding to each amplitude value unit to obtain the quantized amplitude value f corresponding to each amplitude value unit N : f N =f / Δf (2) Among them, f N is the quantized amplitude value, ranging from 0 to 255; f is the range Doppler amplitude value; Step 13: quantize the amplitude value f N Perform statistics and convert the quantized amplitude value f N The numbers of 0-255 are stored in the array Num respectively, and the values ​​in the array represent the number of samples with an amplitude of 0-255.

3. The target detection method based on global statistical characteristics of clutter according to claim 1, characterized in that: The expected number of target detection results in step 2 is calculated using formula 3: Where N is the expected number of target detection results, k is the broadening coefficient, which is 1 without interference and 1.2 with interference; L is the target length; is the ground-rubbing angle, C is the speed of light; F s is the system sampling rate, and the number of Doppler units across the target in azimuth is 3.

4. The target detection method based on global statistical characteristics of clutter according to claim 1, characterized in that: The step 3 specifically includes the following sub-steps: Step 31, calculate the range clutter width, the specific operation is as follows: Step 311, calculate the distance unit corresponding to the sampling of the beam center: N R =(R c -R0) / Fs(4) Among them, R c is the center slope distance, R0 is the start of the sampling gate, N R The distance unit corresponding to the sample representing the beam center; Step 312, according to the incident angle of the beam, the beam range width and the platform height, the number of sampling points M in the range direction is calculated using Formula 5: M=(H*tan(α+β / 2)-H*tan(α-β / 2))*Sin(α) / Fs (5) Where M represents the number of sampling points in the range direction, which is used to display the effective beam range, H is the platform height, α and β are the beam incident angle and beam range width respectively; Step 313: The distance unit N corresponding to the sampling of the beam center obtained in step 311 is R The starting and ending coordinates of the distance unit are calculated by the following formula: R -M / 2, N R +M / 2; Step 32, calculating the azimuth Doppler width and the number of Doppler channels, the specific operation is as follows: The Doppler values ​​of point C and point D are calculated using equations 6 and 7 respectively, and the number of Doppler channels corresponding to the clutter is calculated using equation 8: f c =2 / λ*(V x *X c -V y *Y c +V z *Z c ) / R c (6) f d =2 / λ*(V x *X d -V y *Y d +V z *Z d ) / R c (7) N f =(f d -f c ) / Prf*N p (8) Among them, point C and point D are the beam azimuth edges, and their coordinates are (X c ,Y c ,Z c )、(X d ,Y d ,Z d ), θ is the azimuth beam width, V x 、V y 、V z are the three-dimensional speeds of the platform respectively; Prf is the repetition frequency, N p is the azimuth channel number, the starting position of the clutter channel is N P / 2-N f , N P / 2-1+N f / 2; λ is the wavelength; Step 33, calculate the global statistical mean of the clutter background, the specific operation is as follows: Step 331, for the range Doppler amplitude value N P / 2-N f ~N P / 2-1+N f / 2 channel and N R -M / 2~N R +M / 2 distance unit range amplitude sum, get the amplitude sum, calculate the ratio of the amplitude sum to the number of samples to get the sample amplitude mean f mean ; Step 332, traverse the range Doppler amplitude values ​​again and filter out the amplitude values ​​greater than f mean *k max Strong target information and less than f mean *k min The weak noise information of the target information and the amplitude value corresponding to the weak noise information are expressed by f mean Substitute to get the sample data after filtering, k max and k min Take 2.0 and 0.6 respectively; Step 333: Take the sample data obtained in step 332 as the processing object, execute step 331, and obtain a new sample amplitude mean value f m ' ean , which is used as the final global statistical mean of the clutter background.

Images