A method for detecting aerial moving targets using dynamic grid imaging along target trajectory

Abstract

The present invention provides a method for detecting aerial moving targets using dynamic grid imaging along the target trajectory, belonging to the field of radar signal processing. The present invention utilizes the motion trajectory of an aerial moving target to achieve dynamic grid imaging of the radar's echo over a period of tens of seconds, and uses a phase gradient autofocusing algorithm to further compensate for the accumulated performance degradation caused by motion trajectory errors. Compared to direct coherent integration in the range-Doppler domain, the method proposed by the present invention overcomes the range migration and Doppler spread caused by the relative motion between the platform and the aerial moving target, resulting in a greater signal-to-noise ratio gain and further improving the probability of detecting aerial moving targets.

Description

Technical Field

[0001] The present invention relates to the field of radar signal processing, and in particular to a method for detecting aerial moving targets by dynamic grid imaging along target tracks. Background Art

[0002] High-orbit radars operate in geosynchronous orbit, offering a wide field of view and a shorter revisit period for a specific area of ​​interest. This facilitates long-term observation of a specific region, possessing significant civilian and military value and representing a key direction for future spaceborne radar systems. However, the unique nature of their location also presents numerous challenges in signal processing.

[0003] Because high-orbit radars orbit at an altitude of approximately 36,000 km, the slant range between the satellite platform and the target on the Earth's surface is extremely large, resulting in extremely weak signals received by the platform, making direct target detection difficult. Given that the radar antenna cannot be adjusted, the only way to improve the signal-to-noise ratio of the target signal is to increase the radar echo integration time. However, detecting moving targets in the air often requires tens of seconds of coherent integration time. Over such a long period, using traditional two-dimensional coherent integration methods in the range-Doppler domain will result in significant range migration and Doppler spread due to the motion of the moving target and the platform, significantly reducing the integration efficiency. In some cases, the trajectory of a moving target over a period of time is known or predictable, such as the "air corridors" that airliners must pass through to land at certain airports. Integrating this known target trajectory information into the signal integration process can effectively improve the integration efficiency and increase the probability of detecting moving targets in the air. Summary of the Invention

[0004] In light of this, and to address the existing technical issue of weak echo signals from high-orbit radars detecting aerial moving targets, the present invention provides a method for detecting aerial moving targets using dynamic grid imaging along the target's trajectory. This method utilizes the target's trajectory to achieve dynamic grid imaging of the radar's echo signals over a period of tens of seconds. A phase gradient autofocusing algorithm is then used to further compensate for the performance degradation caused by accumulated trajectory errors. Compared to direct coherent integration in the range-Doppler domain, the proposed method overcomes the range migration and Doppler spread caused by the relative motion between the platform and the aerial moving target, resulting in a greater signal-to-noise ratio gain and further improving the probability of detecting aerial moving targets.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for detecting aerial moving targets using dynamic grid imaging along the target trajectory utilizes the motion trajectory of the aerial moving target to achieve dynamic grid imaging of the radar's echo over a period of tens of seconds, and uses a phase gradient autofocusing algorithm to further compensate for the accumulated performance degradation caused by motion trajectory errors.

[0007] Preferably, the method comprises the following steps:

[0008] Step (1), constructing a matched filter based on each pulse echo and waveform parameters such as pulse duration and modulation frequency, performing matched filtering on each pulse to obtain a pulse compressed signal;

[0009] Step (2): Calculate the range and azimuth resolutions based on the system parameters of the frequency modulation bandwidth, the equivalent synthetic aperture angle, and the ground-grazing angle of the beam center, determine the grid size based on the resolution, set the center moment scene imaging grid, and then inject the predicted target position based on the known target trajectory at each pulse moment into the center moment imaging grid to form a dynamic grid at each pulse moment;

[0010] Step (3), calculating the beam off-axis angle between the radar and all target points in the dynamic grid at each pulse moment based on the antenna pointing direction, platform position, and target position, and determining whether the target point is illuminated by the beam;

[0011] Step (4), for the target point illuminated by the beam, the equivalent slant range and echo delay between the radar and the target point are calculated according to the platform position and the target position;

[0012] Step (5), performing range interpolation processing on the pulse compression signal, and extracting the echo signal corresponding to the target point at the pulse time;

[0013] Step (6), compensating the azimuth phase factor of the echo signal extracted in step (5), and performing coherent superposition processing on each pulse to obtain a preliminary dynamic grid imaging result along the target trajectory;

[0014] Step (7) performs phase gradient self-focusing on the preliminary dynamic grid imaging result along the target trajectory obtained in step (6) to further improve the accumulation effect and obtain a phase gradient self-focusing result of dynamic grid imaging along the target trajectory.

[0015] Preferably, in step (1), each pulse echo is a matrix of size Na*Nr, where Na is the number of pulses and Nr is the number of sampling points of each pulse;

[0016] The matched filter is 1*N r The matched filtering result is also a matrix of Na*Nr.

[0017] Preferably, in step (2), for the nth pulse, the position of the moving target at the pulse moment predicted based on the prior trajectory information is used to compensate the entire center moment imaging grid for its relative displacement relative to the center moment position of the known trajectory to obtain a dynamic grid at each pulse moment.

[0018] Preferably, in step (2), the position vector of any grid point at the central moment is expressed as Then at the nth pulse moment, the position vector of any grid point is expressed as

[0019] Preferably, in step (7), the specific steps of phase gradient self-focusing include:

[0020] Step (71), find the position of the strong point target in the image, select the range gate where the strong point target is located, and perform circular shift in the azimuth direction so that the strong point is shifted to the center position in the azimuth direction;

[0021] Step (72), setting a window length of appropriate size for the azimuth of the strong point target so that the main energy of the azimuth of the point target is retained in the window, and obtaining a windowed result;

[0022] Step (73), after azimuth circular shifting and windowing, performing azimuth inverse Fourier transform on the obtained image to obtain range time domain and azimuth frequency domain signals;

[0023] Step (74), estimating the phase error using a linear unbiased minimum estimate;

[0024] Step (75): Integrate all azimuth phase errors to obtain the final phase error that needs to be compensated.

[0025] Step (76): multiply each point of the image by the corresponding to perform phase error compensation.

[0026] Preferably, in step (74), in actual application, the phase errors of the adjacent azimuth directions are estimated by directly using the conjugate multiplication between the adjacent azimuth directions and taking the sum.

[0027] The present invention provides a method for detecting aerial moving targets using dynamic grid imaging along the target trajectory. This method utilizes the trajectory of an aerial moving target to achieve dynamic grid imaging of the radar's echo over a period of tens of seconds. It also uses a phase gradient autofocusing algorithm to further compensate for the cumulative performance degradation caused by trajectory errors. This method is suitable for situations where the trajectory of an aerial moving target is known over a period of time, and allows the known trajectory of the aerial moving target to have a very large azimuth position error and a certain range position error relative to the true trajectory. Compared to existing technologies, this method has the following beneficial effects:

[0028] (1) Compared with direct coherent integration in the range-Doppler domain, the method proposed in this invention overcomes the range migration and Doppler spread caused by the relative motion between the platform and the aerial moving target, brings a greater signal-to-noise ratio gain, and further improves the detection probability of the aerial moving target.

[0029] (2) By incorporating prior information about some aerial moving targets into the imaging method, the present invention achieves effective coherent accumulation of signals on the order of tens of seconds when detecting aerial moving targets. This effectively overcomes the problems of weak signals detected by high-orbit radars, severe range migration caused by the movement of aerial moving targets on the order of tens of seconds, and the difficulty in coherent accumulation of signals caused by Doppler spread. This improves the detection probability of aerial moving targets by high-orbit radars. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 The present invention is a flowchart of the steps of a method for detecting aerial moving targets by dynamic grid imaging along the target trajectory.

[0031] Figure 2 This is the preliminary dynamic grid imaging result along the target trajectory.

[0032] Figure 3 This is the phase gradient self-focusing result after dynamic grid imaging along the target trajectory.

[0033] Figure 4 This is the phase gradient self-focusing magnification result after dynamic grid imaging along the target trajectory.

[0034] Figure 5 This is a slice diagram of the phase gradient from the focusing distance after dynamic grid imaging along the target trajectory.

[0035] Figure 6 This is an azimuthal slice diagram of the phase gradient autofocusing after dynamic grid imaging along the target trajectory.

[0036] Figure 7 This is a graph of mixed accumulation results for comparative experiments.

[0037] Figure 8 This is a diagram of the mixed accumulation and magnification results of the comparative experiment.

[0038] Figure 9 Mixed accumulation distance dimension slice map for comparison experiment.

[0039] Figure 10 Slice diagram of the mixed accumulation velocity dimension for comparative experiments. DETAILED DESCRIPTION

[0040] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0041] In the task of detecting aerial moving targets by high-orbit radar, the echo energy is extremely weak and requires long-term accumulation. In addition, the aerial moving targets move greatly over a long time span, making it difficult for traditional methods to achieve effective accumulation. The present invention provides an aerial moving target detection method using dynamic grid imaging along the target trajectory. The overall flow chart is as follows: Figure 1 As shown, the present invention utilizes the motion trajectory of an aerial moving target to realize dynamic grid imaging of the radar's echo of the aerial moving target for tens of seconds, and uses a phase gradient autofocusing algorithm to further compensate for the accumulated performance degradation caused by the motion trajectory error.

[0042] In the present invention, the radar system parameters adopt the parameters shown in Table 1, which are as follows:

[0043] Table 1 Radar system parameters

[0044] parameter value Average antenna transmit power 10kw Antenna effective transmit and receive gain 60dB Receiver noise temperature 293k Receiver noise figure 2dB System loss 10dB carrier frequency 3GHz bandwidth 6MHz Pulse repetition frequency 200Hz Accumulation time 50s

[0045] The present invention specifically comprises the following steps:

[0046] Step (1): According to each pulse echo Echo and waveform parameter pulse duration T r , frequency modulation K r , construct a matched filter, perform matched filtering on each pulse, and obtain the pulse compression signal S rc .

[0047] Wherein, each pulse echo Echo is preferably a matrix of size Na*Nr, Na is the number of pulses, and Nr is the number of sampling points of each pulse. The matched filter is a 1*N r The matched filtering result is also a matrix of Na*Nr.

[0048] Among them, the construction steps of the matched filter can be operated by the following formula:

[0049] hrc=(abs(tr)<Tr / 2)*exp(1j*pi*Kr*tr 2 )

[0050] Hrc=conj(fftshift(fft(fftshift(hrc))))

[0051] Where hrc is the time domain matched filter, abs() is the absolute value operation, t r is the pulse fast time one-dimensional vector, T r is the pulse duration, which is 0.5ms in this example. exp() is the natural exponential function, pi is the circumference of a circle, and K r is the linear frequency modulation slope. Hrc is the frequency domain matched filter, conj() is the complex conjugate operation, fft() is the fast Fourier transform operation, and fftshift() is the operation of shifting the zero-frequency component to the center of the spectrum.

[0052] The matched filtering operation for each pulse can be expressed as:

[0053] srfft=fftshift(fft(fftshift(Echo,2),Nr,2),2)

[0054] src=ifftshift(ifft(ifftshift(srfft*Hrc,2),Nr,2),2)

[0055] Where srfft is the frequency domain form of the pulse compressed signal, src is the time domain form of the pulse compressed signal, ifft() is the inverse fast Fourier transform operation, and ifftshift() is the inverse zero-frequency shift operation.

[0056] Step (2), frequency modulation bandwidth B according to system parameters w , the equivalent synthetic aperture angle Δi, the ground-grazing angle of the beam center, and the range and azimuth resolution ρ are calculated r , ρ az The grid sizes dx and dy are determined based on the resolution, and the central moment scene imaging grid is set. The predicted target positions based on the known target trajectory at all pulse moments are then injected into the central moment imaging grid to form a dynamic grid at each pulse moment.

[0057] In step (2), for the nth pulse, the moving target position predicted based on the prior trajectory information at the pulse moment is The center position relative to the known trajectory Relative displacement The position compensation is performed on the imaging grid of the entire central moment to obtain the dynamic grid at each pulse moment.

[0058] If the position vector of any grid point at the central moment can be expressed as Then at the nth pulse moment, the position vector of any grid point can be expressed as

[0059] In this invention, the scene center is set at an altitude of 10,000 meters at the center of the beam illumination. The north-south direction is the range direction. The positive x-axis of the scene coordinate system is set to the north direction, the east-west direction is the azimuth direction, and the y-axis of the scene coordinate system is set to the east direction. According to the right-hand coordinate system, the positive z-axis is close to the center of the earth. According to theoretical calculations, the range resolution of the slant range plane of the imaging processing can be expressed as:

[0060]

[0061] Where c is the speed of light and B is the radar signal bandwidth.

[0062] Considering that the ground plane and the slant range plane do not coincide, the resolutions of the two planes cannot be equal. The range resolution of the ground plane can be expressed as:

[0063]

[0064] The angle γ is the angle between the beam line of sight and the ground plane.

[0065] There is no precise expression for the azimuth resolution of a high-orbit radar. Due to its high orbital altitude, the influence of the Earth's rotation cannot be ignored, and the calculation methods of the Doppler bandwidth and equivalent velocity of the airborne model and the low-orbit model cannot be directly applied. Using the generalized ambiguity function derivation, the azimuth resolution can be expressed as:

[0066]

[0067] Where λ is the wavelength of the radar signal, ω Σ is the relative angular velocity between the platform and the target point around the center of the earth, T c is the effective synthetic aperture time, then θ is the total rotation angle of the platform relative to the target point around the center of the earth. Due to the rotation of the earth and the movement of the satellite platform, the direction of the azimuth resolution is always along ω Σ Direction of angular velocity In fact, for a fixed point on the ground Since the synthetic aperture angle is very small, if the position vector of the satellite platform at the beginning and end of the synthetic aperture time is known, the total rotation angle θ can be approximated as:

[0068]

[0069] in, is the starting time of synthetic aperture, the time between satellite platform and target point The unit vector between is the end time of synthetic aperture, the satellite platform and the target point The unit vector between . Here All vectors are calculated in the Earth-centered Earth-fixed coordinate system (ECEF).

[0070] It is also important to note that since the target to be imaged is a moving target, its own motion contributes to the relative angular velocity between the platform and the target around the center of the Earth. Therefore, the target's own motion also affects the azimuth resolution of the imaging result. Furthermore, the phase gradient autofocus algorithm can also improve azimuth resolution to a certain extent.

[0071] In the present invention, the range ground resolution is calculated to be 95.9m, and the azimuth ground resolution, which takes into account the motion of moving targets, is calculated to be 75.6m. Due to phase gradient autofocusing, the actual values ​​may be smaller. When setting the scene grid size, the embodiment selects a range grid size of 50m and an azimuth grid size of 50m.

[0072] Then determine the center coordinate ImageCenter in the scene coordinate system. The longitude, latitude, and altitude of this coordinate are Lmbda0, Theta0, and Elevation0. Then, perform grid division. The x-axis and y-axis coordinates of each grid scene coordinate system can be expressed as:

[0073] BPImagePx=(-0.5*Nx:1:0.5*Nx-1)*dx+BPImageCenter(1)

[0074] BPImagePy=(-0.5*Ny:1:0.5*Ny-1)*dy+BPImageCenter(2)

[0075] Where Nx is the number of grid points in the distance direction, dx is the grid size in the distance direction, and BPImageCenter(1) is the x-axis coordinate of the center coordinate; Ny is the number of grid points in the azimuth direction, dy is the grid size in the azimuth direction, and BPImageCenter(2) is the y-axis coordinate of the center coordinate. The conversion from the scene coordinate system to the latitude and longitude coordinate system can be expressed as:

[0076] GridElevation=Elevation0

[0077] GridLambdat=Lambda0+BPImagePy / (Re_center+GridElevation)

[0078] GridThetat=Theta0+BPImagePx / (Re_center+GridElevation)

[0079] Where Re_center is the equivalent earth radius at the center coordinate.

[0080] The conversion relationship from the latitude and longitude coordinate system to the ECEF coordinate system can be expressed as:

[0081] Gridxgt=(Re*Rp / sqrt(Rp 2 *cos 2 (GridThetat)+Re 2 *sin 2 (GridThetat))+GridElevation)*cos(GridThetat)*cos(GridLambdat)

[0082] Gridygt=(Re*Rp / sqrt(Rp 2 *cos(GridThetat) 2 +Re 2 *sin(GridThetat) 2 )+GridElevation)*cos(GridThetat)*sin(GridLambdat)

[0083] Where Re is the Earth's equatorial radius, Rp is the Earth's polar radius, sqrt() is the square root operation, GridThetat is the latitude of the target point, GridLambdat is the longitude of the target point, and GridElevation is the altitude of the target point. The obtained coordinate posg_Grid in the ECEF coordinate system is (Gridxgt, Gridygt, Gridzgt).

[0084] In the present invention, the moving target point T is set, and its RCS is 10m 2 Its central position at the moment is [0m, 0m, -10000m] at the center of the scene, and its true speed is [0m / s, -340m / s, 0m / s]. That is, at an altitude of 10000m, it moves at a constant speed of 340m / s in the west direction.

[0085] The known moving target trajectory of the present invention is assumed to have a velocity of [0 m / s, -300 m / s, 0 m / s]. This means that the target trajectory injected into the dynamic grid image along the target trajectory has an azimuth velocity error of 40 m / s compared to the actual target trajectory. In the present invention, the known moving target trajectory is injected directly in the ECEF coordinate system. For each pulse moment, the operation can be expressed as:

[0086] posg_Grid_movinng=posg_Grid+posg_target_moving

[0087] Among them, posg_Grid_movinng is the dynamic grid position vector along the target trajectory that needs to be used for imaging accumulation after each pulse moment, and posg_target_moving is the relative displacement vector of the position obtained at the pulse moment according to the known moving target trajectory in the ECEF coordinate system relative to the position obtained at the center moment according to the known moving target trajectory.

[0088] In the present invention, since the trajectory of the known moving target in the air is set as a relatively simple uniform linear motion model, posg_target_moving can be simply obtained by multiplying the estimated speed by time, that is:

[0089] posg_target_moving=v_track*(t-t_center)

[0090] Where v_track is the velocity vector of the known target trajectory, t is the time variable at the current pulse moment, and t_center is the time variable at the center moment. If the known target trajectory has a more complex motion model, it can also be injected into the center moment grid to construct a dynamic grid at each pulse moment.

[0091] Step (3), according to the antenna pointing Platform location and target location Calculate the beam off-axis angle θ between the radar and all target points at the current azimuth moment a , to determine whether the target point is illuminated by the beam. In the present invention, the antenna diameter is set to 50m, the carrier frequency is 3GHz, and the half-beam width of the antenna in the range and azimuth directions is about 0.05°. The calculation of the beam off-axis angle is usually carried out in the antenna coordinate system. The coordinate origin of the antenna coordinate system is the phase center of the antenna, the positive direction of the x-axis points to the true flight direction of the satellite, the y-axis is along the antenna sighting line, and the direction pointing to the earth is positive. The z-axis is pointed out by the right-hand rule, so that the coordinate system constitutes a right-hand rectangular coordinate system. In other words, the direction vector pointed by the beam can be expressed as (0,1,0). If the position of the target point is converted to the coordinates in the antenna coordinate system as (posa_target_x,posa_target_y,posa_target_z), then the size of the off-axis angle can be expressed as:

[0092]

[0093] Step (4): the target point of the beam irradiation is determined according to the platform position. and target location Calculate the equivalent slant range R and two-way delay τ between the radar and the target point.

[0094] In addition, it should be noted that the coordinate solution of both the satellite platform and the ground target point should finally be converted to the inertial coordinate system, usually the Earth-centered inertial coordinate system (ECI), to avoid coordinate and slant range solution errors caused by changes in other coordinate systems at different time points.

[0095] Step (5): pulse compression signal S rc Perform distance interpolation to obtain S rcup , and extract the echo signal sig corresponding to the target point at the pulse time temp .

[0096] The specific operations can be:

[0097] After Fourier transforming the pulse pressure signal, the result is padded with zeros at both ends, and then the inverse Fourier transform is performed on the zero-padded result. The upsampling factor RangeInterpolation in the present invention is 4 times, which means that after the zero-padding operation, the total number of signal points is quadrupled. A pulse signal with an original number of sampling points Nr will be converted to Nrup = RangeInterpolation * Nr points. The specific operation can be expressed as:

[0098] srcfftup=zeros(Na,Nrup)

[0099] srcfftup(:,floor((Nrup-Nr) / 2)+(1:Nr))=srfft*Hrc

[0100] srcup=ifftshift(ifft(ifftshift(srfftup,2),Nrup,2),2)

[0101] Where Nrup is the number of signal points after upsampling, srcfftup is the frequency domain signal after zero padding, srcup is the pulse compression signal after upsampling, step (6), the echo signal sig extracted in step (5) temp , to compensate for its azimuth phase factor And perform coherent superposition processing on each pulse to obtain the preliminary dynamic grid imaging result Image along the target trajectory. The specific operation process is that for each target point at each moment, there is a calculated instantaneous slant distance R, and then the distance unit corresponding to the slant distance is selected from the signal at the corresponding moment. However, the instantaneous slant distance is often not exactly equal to the value of the distance unit, and is often between two distance units. Assume R min is the distance represented by the minimum distance unit that the signal can be detected, and dr is the size of each distance unit. Then the following variables can be calculated:

[0102] N=(RR min ) / dr

[0103] N int =floor(N)

[0104] coe=NN int

[0105] In the above formula, N is the non-integer number of distance units corresponding to the slant distance R, N int The floor() operation is the rounding operation for the distance unit where the slope distance R is located. int and Nth int +1 distance unit. coe can be understood as R to Nth int The distance unit represents the degree of distance. When coherently accumulating energy at the target point at this moment, the following operations can be performed:

[0106] sig temp =(1-coe)*srcup(N)+coe*srcup(N+1)

[0107] Where srcup is the signal after four times upsampling pulse compression. temp The temporary value of the energy accumulated at the target point at that moment.

[0108] Then directly to sig temp Phase compensation can be performed directly through sig temp *exp(j4πR / λ) is calculated for compensation. Each target will calculate the phase compensation sig at each moment. temp The final value of the target point is obtained by adding the values. The initial imaging result along the target trajectory is as follows: Figure 2 As shown in the figure, under the effect of imaging along the moving target trajectory, the energy of the moving target in the air is focused to a certain extent, but it is still severely defocused, forming a straight line. At this time, the signal-to-noise ratio is very low, and effective detection cannot be performed.

[0109] Step (7): performing phase gradient autofocusing (PGA) on the preliminary dynamic grid imaging result Image along the target trajectory obtained in step (6) to further improve the accumulation effect and obtain the dynamic grid imaging phase gradient autofocusing result Image_PGA along the target trajectory.

[0110] The main principle of the PGA algorithm is to estimate the phase information of the azimuth direction according to the defocus of the special point, and use the iterative idea to continuously estimate and compensate the phase error until the phase error value meets the requirements.

[0111] In step (7), the specific steps of phase gradient autofocusing (PGA) include:

[0112] Step (71) finds the strong point target position in the image Image, selects the range gate m where the strong point target is located, and performs a circular shift in the azimuth direction so that the strong point is shifted to the azimuth center position M / 2+1, where M is the total number of points in the azimuth direction.

[0113] The specific operation can be expressed as:

[0114] Scene_temp(:,i)=circshift(image_temp(:,i),[-m(i)+M / 2+1 0])

[0115] Wherein, image_temp is the imaging result of step 6. For the i-th range gate, in the first dimension of the image, that is, the azimuth direction, the cyclic shift size is -m(i)+M / 2+1, m(i) is the brightest point position of the range gate, and M is the number of points in the image azimuth direction, which is 200 in this invention.

[0116] Step (72) sets a window length [winL, winR] of appropriate size for the azimuth of the strong point target, so that the main energy of the point target azimuth is retained in the window, and the result after windowing is g n This operation can reduce the impact of other point targets and clutter on the accuracy of phase error estimation. In this embodiment, the window length setting threshold is 0.05 of the maximum energy point size of each range gate.

[0117] Step (73), after azimuth circular shift and windowing, the resulting image is represented by g n Indicates that g n Perform inverse Fourier transform and get G n , the specific operation can be expressed as the following form:

[0118] G n (m)=|G n (m)|exp{j[φ α (m)+θ n (m)]}

[0119] Where m is the image azimuth and n is the image distance. n (m)| is the amplitude. φ α (m) is the phase error, θ n (m) is the error-free phase value.

[0120] Step (74) uses linear unbiased minimum estimation to estimate the phase error. In practical applications, the phase error is estimated by directly using the conjugate multiplication between adjacent azimuth directions and taking the sum. The phase error of adjacent azimuth directions can be obtained. The specific operation can be expressed as:

[0121]

[0122] Step (75), then calculate all azimuth phase errors By integrating, we can get the phase error that needs to be compensated. This step can be expressed as:

[0123]

[0124] Step (76), for image G n Multiply each point by the corresponding Phase error compensation is performed. The phase error is determined during the iteration process. If the phase error is less than π / 4, the image is considered well focused and no compensation is required. Generally, the phase error requirement of less than π / 4 can be achieved within 10 iterations.

[0125] The imaging result along the moving target trajectory after self-focusing is as follows Figure 3 In order to better observe the results, the enlarged diagram is shown as Figure 4 As shown, the range and azimuth slices are as follows Figure 5 、 Figure 6 shown.

[0126] In order to verify the effectiveness of the present invention, CFAR detection is performed on the final image domain signal to achieve aerial moving target detection. In this embodiment, CAFR detection uses the CA-CFAR method, the size of the protection unit in both directions is set to 3, the size of the training unit in both directions is set to 2, and the false alarm rate is set to 10 -5 CFAR is implemented using the two-dimensional CFAR detector of the phased toolbox in Matlab. The construction of the detector can be expressed as:

[0127] detector=phased.CFARDetector2D('TrainingBandSize',[3,3],'ThresholdFactor','Auto','GuardBandSize',[2,2],'ProbabilityFalseAlarm',1e-5,'Method','CA')

[0128] After 1000 Monte Carlo experiments, the final detection probability was 91.2%.

[0129] In order to further verify the effectiveness of the method proposed in the present invention, a comparative experiment was carried out.

[0130] The commonly used range-Doppler domain integration method, hybrid integration, is used to accumulate the same echoes. Similarly, the known trajectory of the moving target in the air is used to phase compensate the echoes and then accumulate them to achieve the effect of controlling the variables.

[0131] The accumulated results are as follows Figure 7 As shown. In addition, the accumulation result interpolation amplification result is as follows Figure 8 As shown, the distance dimension and velocity dimension slice results are as follows Figure 9 、 Figure 10 As shown in Figure 2, the signal-to-noise ratio of the cumulative results of the comparative experiment is significantly different from that of the cumulative results of the present invention, far less than the 13dB required for reliable detection. Using the same CFAR detector configuration, 1000 Monte Carlo experiments were performed, and the final detection probability was 10.7%.

[0132] The comparison shows that the present invention has a significant improvement in the accumulation effect compared to the existing methods, which verifies the effectiveness of the method. It is a significant improvement compared to the commonly used range-Doppler domain accumulation method.

[0133] The proposed method for detecting aerial moving targets using dynamic grid imaging along their trajectory injects the trajectory of an aerial moving target over a period of time into a static imaging grid, generating a dynamic grid along the target's trajectory. This method then uses an imaging algorithm to accumulate energy and further focuses the image using a phase gradient autofocusing method. The above results demonstrate the effective accumulation of energy from aerial moving targets, validating the effectiveness of the method.

[0134] While various embodiments of the present invention have been described above, the above descriptions are intended to be illustrative, non-exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to existing technologies, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for detecting aerial moving targets using dynamic grid imaging along target trajectories, characterized in that: The radar uses the trajectory of a moving target to achieve dynamic grid imaging of the target's echo over a period of tens of seconds. The phase gradient autofocus algorithm is then used to further compensate for the accumulated performance degradation caused by the trajectory error. The specific steps are as follows: Step (1), constructing a matched filter based on each pulse echo and waveform parameters such as pulse duration and modulation frequency, performing matched filtering on each pulse to obtain a pulse compressed signal; Step (2): Calculate the range and azimuth resolutions based on the system parameters of the frequency modulation bandwidth, the equivalent synthetic aperture angle, and the ground-grazing angle of the beam center, determine the grid size based on the resolution, set the center moment scene imaging grid, and then inject the predicted target position based on the known target trajectory at each pulse moment into the center moment imaging grid to form a dynamic grid at each pulse moment. Step (3), based on the antenna pointing, platform position and target position, calculate the beam off-axis angle between the radar and all target points in the dynamic grid at each pulse moment, and determine whether the target point is illuminated by the beam; Step (4), for the target point illuminated by the beam, the equivalent slant range and echo delay between the radar and the target point are calculated based on the platform position and the target position; Step (5), performing range interpolation processing on the pulse compression signal, and extracting the echo signal corresponding to the target point at the pulse time; Step (6), compensating the azimuth phase factor of the echo signal extracted in step (5), and performing coherent superposition processing on each pulse to obtain a preliminary dynamic grid imaging result along the target trajectory; Step (7), performing phase gradient self-focusing on the preliminary dynamic grid imaging result along the target trajectory obtained in step (6), further improving the accumulation effect, and obtaining a phase gradient self-focusing result of dynamic grid imaging along the target trajectory; In step (7), the specific steps of phase gradient self-focusing include: Step (71), find the position of the strong point target in the image, select the range gate where the strong point target is located, and perform a circular shift in the azimuth direction so that the strong point is shifted to the center position in the azimuth direction; Step (72), setting a window length of appropriate size for the azimuth of the strong point target so that the main energy of the azimuth of the point target is retained within the window, and obtaining a windowed result; Step (73), after azimuth circular shifting and windowing, performing azimuth inverse Fourier transform on the obtained image to obtain range time domain and azimuth frequency domain signals; Step (74), estimating the phase error using a linear unbiased minimum estimate; Step (75): Integrate all azimuth phase errors to obtain the final phase error that needs to be compensated. Step (76): Multiply each point of the image by the corresponding to perform phase error compensation.

2. The method for detecting aerial moving targets using dynamic grid imaging along target trajectories according to claim 1, wherein: In step (1), each pulse echo is a The matrix, is the number of pulses, is the number of sampling points for each pulse; the matched filter is a The size of the vector, the matched filtering result is also a The matrix of .

3. The method for detecting aerial moving targets using dynamic grid imaging along target trajectories according to claim 1, wherein: In step (2), for The moving target position predicted according to the prior trajectory information at the pulse moment is compensated for the position of the entire center moment imaging grid by the relative displacement of the center moment position of the known trajectory to obtain the dynamic grid at each pulse moment.

4. The method for detecting aerial moving targets using dynamic grid imaging along target trajectories according to claim 3, wherein: In step (2), the position vector of any grid point at the central moment is expressed as , then At the pulse moment, the position vector of any grid point is expressed as .

5. The method for detecting aerial moving targets by dynamic grid imaging along target trajectory according to claim 1, characterized in that: In step (74), in actual application, the phase errors of adjacent azimuth directions can be estimated by directly using the conjugate multiplication between adjacent azimuth directions and taking the sum.

Images