Indoor moving target detection method based on Doppler frequency modulation profile and range gating filtering

Abstract

The invention discloses an unmanned aerial vehicle-mounted through-the-wall radar indoor moving target detection method. The method comprises the step of providing a moving target detection method based on a Doppler frequency profile and range gating filtering. And the imaging result of the moving target is defocused and is easily submerged in strong static clutters, so that the detection rate of the moving target is reduced. In order to solve the problem, the Doppler frequency modulation rate characteristics of the static clutter and the moving target are analyzed. Then, the range migration of the target echo is corrected in the range Doppler domain through sinc interpolation; thirdly, estimating the Doppler frequency modulation rate of each range gate of the echo by applying fractional Fourier transform, comparing the estimated value of the frequency modulation rate with the theoretical value of the frequency modulation rate of the stationary target, and designing a range gating filter to suppress clutters generated by the wall body and the stationary object based on the difference of the Doppler frequency modulation rates; and finally, constructing an azimuth compression filter to realize focusing imaging of the moving target. Simulation and actual measurement experiments verify feasibility and effectiveness of the method.

Description

Technical Field

[0001] The present invention relates to the field of radar signal processing, and in particular to a method for detecting indoor moving targets by a through-wall radar carried by an unmanned aerial vehicle. Background Art

[0002] Through-the-wall radar uses electromagnetic waves in the L / S band to detect dangerous moving targets indoors. Through-the-wall radar mounted on drones uses the advantages of flexible deployment, rapid response and wide-area coverage of drones to achieve remote, accurate, non-contact detection and positioning of moving targets, which is especially suitable for complex environments and inaccessible areas. However, the speed of the moving target causes the imaging result to be defocused, resulting in the moving target being easily submerged in strong stationary clutter, reducing the detection rate of the moving target. The indoor moving target detection method of the through-the-wall radar mounted on drones proposed in the present invention aims to propose a technology that can suppress stationary clutter while focusing the moving target for imaging.

[0003] At present, through-the-wall radars for detecting indoor moving targets are mainly based on stationary platforms, and there are relatively few studies on indoor moving target detection by through-the-wall radars carried by drones. However, the ground moving target indication (GMTI) technology of airborne synthetic aperture radar (SAR) has always been a research focus and hot topic in the field of SAR, and can provide a technical reference for moving target detection by through-the-wall radars carried by drones. Airborne SAR-GMTI technology can be divided into multi-channel SAR moving target detection methods and single-channel SAR moving target detection methods according to the number of transceiver antennas used.

[0004] Multi-channel SAR moving target detection methods increase the number of channels along the track and use spatial and temporal information to achieve better clutter suppression effects. Typical methods include Displaced Phase Center Antenna (DPCA) technology, Along Track Interferometry (ATI) technology, Velocity Synthetic Aperture Radar (VSAR) technology, Subspace Projection (SP) technology, and Robust Principal Component Analysis (RPCA) technology. However, these multi-channel SAR moving target detection methods need to meet strict channel registration conditions to suppress stationary clutter, and due to the small phase difference between channels of azimuth moving targets, these methods will cause serious energy loss of azimuth moving targets and make it difficult to detect azimuth moving targets.

[0005] Single-channel SAR has the advantages of low hardware cost and simple engineering implementation. Many related moving target detection methods have been proposed. G.Lv et al. proposed a moving target detection method based on symmetric defocusing. This method uses two symmetric matched filters with opposite phase signs to defocus the SAR image respectively, and obtains two defocused SAR images. The moving target is detected by comparing the sharpness of the two defocused SAR images. However, different moving targets require different defocus filter parameters, and when the signal-to-noise ratio is low, defocusing will further weaken the target energy, which is not conducive to target detection. J.Wang halved the Doppler spectrum, and each half generated a range Doppler (Range Doppler, RD) imaging result, and calculated the difference of the imaging results pixel by pixel, and then set the threshold to detect the moving target. Z.Wang split the Doppler spectrum to generate a multi-view image sequence, and then described the motion characteristics of the moving target response in the image sequence, and introduced the optical flow field to realize the moving target detection. W.Pu did not constrain the low rank of stationary clutter and proposed an unsupervised robust principal component analysis autoencoder network, but the running time of this method was long. In addition, these methods do not consider the impact on target detection when the target cross-distance unit migration is obvious.

[0006] Based on the above research, we believe that the existing airborne moving target detection methods are difficult to achieve strong static clutter suppression and moving target focused imaging in the wall penetration scenario. In order to solve the above problems, we proposed a new indoor moving target detection method. Our method first derived and analyzed the Doppler modulation rate variation range of the target and the wall, and then designed a moving target detection method based on the difference in Doppler modulation rate, including range migration correction, Doppler linear modulation rate profile extraction, static clutter suppression and focused imaging. Simulation and experimental results show that the proposed method can improve the signal-to-clutter ratio and focusing effect of the imaging results.

[0007] Compared with the prior art, the present invention has significant innovations in the following aspects:

[0008] 1. When detecting through walls, the modulation frequencies of moving targets, stationary objects and walls are theoretically derived and analyzed, providing theoretical support for clutter suppression;

[0009] 2. A static clutter suppression method is proposed based on the frequency modulation difference, which provides technical support for moving target detection;

[0010] 3. An azimuth compression filter is constructed based on the Doppler modulation frequency estimation value to achieve focused imaging of indoor moving targets. Summary of the invention

[0011] In order to solve the difficult technical problem that when detecting indoor moving targets by UAV-mounted through-wall radar, the moving targets are submerged in strong stationary clutter, resulting in a reduced moving target detection rate, the present invention proposes an indoor moving target detection method based on Doppler modulation rate profile and range gating filtering, comprising:

[0012] Step S1: obtaining the original data of the echo of the moving target of the wall-penetrating radar carried by the UAV, performing distance compression on the original data, and transforming the echo data after the distance compression into the range Doppler domain;

[0013] Step S2: Calculate the precise range migration of the moving target according to the range Doppler domain expression of the target echo, and perform range migration correction by sinc interpolation;

[0014] Step S3: Obtain the echo data of each range gate by range indexing, and estimate the Doppler modulation rate by fractional Fourier transform.

[0015] Step S4: construct a range-gated filter based on the Doppler modulation rate estimation value to suppress clutter from walls and stationary objects.

[0016] Step S5: construct an azimuth compression filter according to the Doppler modulation rate estimation value to achieve focused imaging of indoor moving targets.

[0017] Furthermore, the specific method of step S1 is:

[0018] UAV-mounted radar transmits linear frequency modulation signal

[0019]

[0020] where rect(·), t, T p 、f c , K r They are rectangular gate function, fast time variable, pulse width, signal center frequency and modulation frequency of linear frequency modulation signal. After receiving the radar echo data, it is compressed by distance to obtain

[0021] s r (t a ,t)=s w (t a ,t)+s sta (t a ,t)+s mov (t a ,t) (2)

[0022] where t a is the slow time variable, s w (t a ,t),s sta (t a ,t) and smov (t a ,t) are wall echo, indoor stationary target echo and moving target echo respectively. The moving target echo can be expressed as

[0023]

[0024] where w a (t a ) is the azimuth rectangular gate function, B is the signal bandwidth, λ is the signal wavelength corresponding to the signal center frequency, R(t a ) is the slant range between the target and the radar, which can be written as

[0025]

[0026] where x0, y0 are t a = The azimuth and range positions of the target relative to the radar at time 0, v x 、v y are the target’s azimuth velocity and range velocity, R w is the additional propagation distance of the signal generated during the wall penetration process, which is relative to R(t a ) is small and can be ignored.

[0027] Taking the derivative of equation (4), we can get the slow time when the slant distance is the smallest:

[0028]

[0029] The corresponding shortest slope distance is

[0030]

[0031] The static target echo is essentially the moving target echo in v x =0, v y = 0, it can be written as follows

[0032]

[0033] Ideally, the wall echo can be regarded as the result of the superposition of multiple closely adjacent static point target echoes. When the complex reflectivity of these point targets is the same, the wall echo after range compression can be expressed as

[0034]

[0035] Where R wall R represents the distance between the UAV-mounted wall-penetrating radar and the wall. wall Can be considered as a constant.

[0036] Transforming the distance of equation (3) to Fourier transform yields

[0037]

[0038] where f r represents the distance frequency, and the azimuth Fourier transform of equation (9) can be obtained

[0039]

[0040] where f a represents the Doppler frequency, f ac and D(f a ) represent the Doppler center frequency and the range migration factor respectively, and their expressions are as follows

[0041]

[0042]

[0043] The square root of the phase in equation (10) is r / f c = 0 Taylor expansion

[0044]

[0045] The quadratic and cubic distance frequencies in formula (13) will cause the distance of the two-dimensional time domain echo to be defocused upward. In order to solve this problem, a quadratic distance compression filter is constructed as follows:

[0046]

[0047] Multiplying equation (13) and equation (14) yields

[0048]

[0049] The echo expression of the echo data in the range Doppler domain obtained by inverse Fourier transform of equation (15) is:

[0050]

[0051] Furthermore, the specific method of step S2 is:

[0052] According to formula (16), the distance migration of the moving target is calculated as follows:

[0053]

[0054] Similarly, for a stationary target, the range migration is

[0055]

[0056] From equations (17) and (18), we can see that the range migration of a stationary target is related to the radar speed, and the range migration of a moving target is related to both the radar speed and the speed of the moving target.

[0057] Range Cell Migration Correction (RCMC) is performed by sinc interpolation. Unlike the traditional sinc interpolation method, it can suppress the part of the echo with a large degree of range migration by setting the number of sampling points of the interpolation kernel function. This method can be expressed as

[0058]

[0059] Where S rcmc (n a ,n r ) is the result after range migration correction, S mov (n a ,n r +i) is the echo data before range migration correction, sinc(Δn r -i) is the interpolation kernel function, which can be regarded as the weight of the original sampling signal, Δn r is ΔR sta (f a ) is the distance migration corresponding to the interpolation function, and P is the number of sampling points of the interpolation function. The result of Sinc interpolation is

[0060]

[0061] Where W r (f a ) is the azimuth gate function after range migration correction.

[0062] Applying azimuth IFFT to equation (20) yields the RCMC result in the two-dimensional time domain:

[0063]

[0064] Furthermore, the specific method of step S3 is:

[0065] Taylor expansion of equation (12) yields

[0066]

[0067] Substituting equation (22) into equation (20), we get

[0068]

[0069] The one-dimensional azimuth signal obtained by indexing the range direction of equation (23) is:

[0070]

[0071] Transforming equation (8) to the range Doppler domain yields

[0072]

[0073] Where δ(f a ) represents the impact signal. The one-dimensional azimuth signal of the range gate in equation (25) is

[0074]

[0075] From equations (24) and (26), it can be seen that after range migration correction, the target echo can be directly extracted in the range Doppler domain and the Doppler modulation rate can be estimated.

[0076] The fractional Fourier transform of equation (24) yields

[0077]

[0078] When the following relationship is satisfied, the peak value appears in equation (27):

[0079]

[0080] That is, according to the angle α corresponding to the peak * Azimuth modulation frequency can be estimated

[0081]

[0082] Similarly, the frequency modulation rate of the wall is estimated to be

[0083]

[0084] Furthermore, the specific method of step S4 is:

[0085] The Doppler modulation frequency estimate Theoretical value of frequency modulation with stationary object Subtract

[0086]

[0087] Among them, ΔK(t) measures the difference between the estimated frequency modulation and the theoretical value of the frequency modulation of a stationary object. The range-oriented gating filter is constructed as

[0088]

[0089] Where α1 is a threshold. For any range gate's range Doppler domain signal, if the difference between the estimated Doppler frequency modulation value and the theoretical value of the stationary object frequency modulation value is less than α1, W1(t) is set to 0. After the echo data in the range Doppler domain is weighted by W1(t), the stationary object clutter can be suppressed.

[0090] Next, in order to suppress wall clutter, based on The distance-gated filter is constructed as

[0091]

[0092] α2 is a threshold. Since the theoretical value of the Doppler frequency modulation of the wall is 0, α2 is also a threshold to measure the difference between the estimated value of the frequency modulation and the theoretical value of the wall frequency modulation. For the range Doppler domain signal of any range gate, if the estimated value of the Doppler frequency modulation is less than α2, the filter is set to 0.

[0093] At this point, the stationary target and wall clutter are suppressed in the range Doppler domain, and the moving target is retained. The echo expression in the range Doppler domain after the stationary clutter is suppressed is:

[0094] S a (f a ,t)=S rcmc (f a ,t)×W1(t)×W2(t) (34)

[0095] Furthermore, the specific method of step S5 is:

[0096] according to The azimuth matched filter is constructed as

[0097]

[0098] in is the estimated frequency modulation rate of each distance unit. Multiplying equation (34) and equation (35) yields

[0099] S a (f a ,t)=S a (f a ,t)H ac (f a ,t) (36)

[0100] Transform equation (36) inversely to obtain

[0101]

[0102] It can be seen from formula (37) that the moving target is focused in azimuth and distance.

[0103] Beneficial effects:

[0104] (1) The present invention proposes an indoor moving target detection method based on Doppler frequency modulation profile and range gating filtering, which is different from the traditional moving target detection technology. This technology does not require complex calculation processes and large amounts of data processing, has higher processing efficiency, and can quickly achieve the suppression of stationary clutter and the focused imaging of moving targets;

[0105] (2) The present invention proposes a method for detecting indoor moving targets based on Doppler frequency modulation profile and range gating filtering, which is different from the traditional moving target detection technology. The drone used in this technology is equipped with radar and can perform large-scale search and detection. This makes the technology not limited by the ground environment, floor height, etc., and has a very large effective search range;

[0106] (3) The present invention proposes an indoor moving target detection method based on Doppler frequency modulation profile and range gating filtering, which is different from the traditional moving target detection technology. This technology does not require channel alignment and echo cancellation, and retains the energy of the moving target to a great extent, which is more conducive to the detection of azimuth moving targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0107] Figure 1 A schematic flow chart of an indoor moving target detection method based on Doppler modulation rate profile and range gating filtering provided by the present invention;

[0108] Figure 2 A schematic diagram of modeling of an indoor moving target detection method based on Doppler modulation rate profile and range gating filtering provided by the present invention;

[0109] Figure 3 A schematic diagram of the structure of a radar antenna array for indoor moving target detection method based on Doppler modulation frequency profile and range gating filtering provided by the present invention;

[0110] Figure 4 A schematic diagram of a simulation scene of an indoor moving target detection method based on Doppler modulation rate profile and range gating filtering provided by the present invention;

[0111] Figure 5 The modulation rate estimation and stationary clutter suppression results of an indoor moving target detection method based on Doppler modulation rate profile and range gating filtering provided by the present invention;

[0112] Figure 6 The invention provides an imaging result of an indoor moving target detection method based on Doppler frequency modulation profile and range gating filtering. DETAILED DESCRIPTION

[0113] The present invention is described in detail below in conjunction with the accompanying drawings and embodiments.

[0114] like Figure 1The method for detecting indoor moving targets based on Doppler modulation rate profile and range gating filtering includes:

[0115] Step S1: Obtain the original data of the moving target echo of the UAV-mounted through-wall radar, perform distance compression on the original data, and transform the echo data after distance compression into the range Doppler domain. The specific method is as follows:

[0116] Consider a closed space surrounded by four walls, such as Figure 2 As shown in Figure 1. The drone flies parallel to the front wall at a speed V. The radar antenna array is as follows Figure 3 As shown in the figure, the antenna array consists of a transmitting antenna and a receiving antenna, and the antenna spacing d is very small and can be ignored. The drone-mounted radar transmits a linear frequency modulation signal

[0117]

[0118] where rect(·), t, T p 、f c , K r They are rectangular gate function, fast time variable, pulse width, signal center frequency and modulation frequency of linear frequency modulation signal. After receiving the radar echo data, it is compressed by distance to obtain

[0119] s r (t a ,t)=s w (t a ,t)+s sta (t a ,t)+s mov (t a ,t) (39)

[0120] where t a is the slow time variable, s w (t a ,t),s sta (t a ,t) and s mov (t a ,t) are wall echo, indoor stationary target echo and moving target echo respectively. The moving target echo can be expressed as

[0121]

[0122] where w a (t a ) is the azimuth rectangular gate function, B is the signal bandwidth, λ is the signal wavelength corresponding to the signal center frequency, R(t a ) is the slant range between the target and the radar, which can be written as

[0123]

[0124] where x0, y0 are t a = The azimuth and range positions of the target relative to the radar at time 0, v x 、v y are the target’s azimuth velocity and range velocity, R w is the additional propagation distance of the signal generated during the wall penetration process, which is relative to R(t a ) is small and can be ignored.

[0125] Taking the derivative of equation (41), we can get the slow time when the slant distance is the smallest:

[0126]

[0127] The corresponding shortest slope distance is

[0128]

[0129] The static target echo is essentially the moving target echo in v x =0, v y = 0, it can be written as follows

[0130]

[0131] Ideally, the wall echo can be regarded as the result of the superposition of multiple closely adjacent static point target echoes. When the complex reflectivity of these point targets is the same, the wall echo after range compression can be expressed as

[0132]

[0133] Where R wall R represents the distance between the UAV-mounted wall-penetrating radar and the wall. wall Can be considered as a constant.

[0134] Transforming the distance of equation (40) to Fourier transform yields

[0135]

[0136] where f r represents the distance frequency, and the azimuth Fourier transform of equation (46) can be obtained

[0137]

[0138] where f a represents the Doppler frequency, f ac and D(f a ) represent the Doppler center frequency and the range migration factor respectively, and their expressions are as follows

[0139]

[0140]

[0141] Replace the square root of the phase in equation (47) with f r / f c = 0 Taylor expansion

[0142]

[0143] The quadratic and cubic distance frequencies in equation (50) will cause the distance of the two-dimensional time domain echo to be defocused upward. To solve this problem, a quadratic distance compression filter is constructed as

[0144]

[0145] Multiplying equation (50) and equation (51) yields

[0146]

[0147] The echo expression of the echo data in the range Doppler domain obtained by inverse Fourier transform of equation (52) is:

[0148]

[0149] Step S2: Calculate the precise range migration of the moving target according to the range Doppler domain expression of the target echo, and perform range migration correction by sinc interpolation. The specific method is as follows:

[0150] According to formula (53), the distance migration of the moving target is calculated as

[0151]

[0152] Similarly, for a stationary target, the range migration is

[0153]

[0154] It can be seen from equations (54) and (55) that the range migration of a stationary target is related to the radar speed, and the range migration of a moving target is related to both the radar speed and the speed of the moving target.

[0155] Range Cell Migration Correction (RCMC) is performed by sinc interpolation. Unlike the traditional sinc interpolation method, it can suppress the part of the echo with a large degree of range migration by setting the number of sampling points of the interpolation kernel function. This method can be expressed as

[0156]

[0157] Where Srcmc (n a ,n r ) is the result after range migration correction, S mov (n a ,n r +i) is the echo data before range migration correction, sinc(Δn r -i) is the interpolation kernel function, which can be regarded as the weight of the original sampling signal, Δn r is ΔR sta (f a ) is the distance migration corresponding to the interpolation function, and P is the number of sampling points of the interpolation function. The result of Sinc interpolation is

[0158]

[0159] Where W r (f a ) is the azimuth gate function after range migration correction.

[0160] Applying azimuth IFFT to equation (57) yields the RCMC result in the two-dimensional time domain:

[0161]

[0162] Step S3: range indexing to obtain echo data of each range gate, and estimate the Doppler modulation rate by fractional Fourier transform. The specific method is as follows:

[0163] Taylor expansion of equation (49) yields

[0164]

[0165] Substituting equation (59) into equation (57), we get

[0166]

[0167] The one-dimensional azimuth signal can be obtained by indexing the range direction of equation (60):

[0168]

[0169] Transforming equation (45) to the range Doppler domain yields

[0170]

[0171] Where δ(f a ) represents the impact signal. The one-dimensional azimuth signal of the range gate in equation (62) is

[0172]

[0173] From equations (61) and (63), it can be seen that after range migration correction, the target echo can be directly extracted in the range Doppler domain and the Doppler modulation rate can be estimated.

[0174] The fractional Fourier transform of equation (61) yields

[0175]

[0176] When the following relationship is satisfied, the peak value appears in equation (64):

[0177]

[0178] That is, according to the angle α corresponding to the peak * Azimuth modulation frequency can be estimated

[0179]

[0180] Similarly, the frequency modulation rate of the wall is estimated to be

[0181]

[0182] Step S4: construct a range-gated filter based on the Doppler modulation frequency estimation value to suppress clutter from walls and stationary objects. The specific method is as follows:

[0183] The Doppler modulation frequency estimate Theoretical value of frequency modulation with stationary object Subtract

[0184]

[0185] Among them, ΔK(t) measures the difference between the estimated frequency modulation and the theoretical value of the frequency modulation of a stationary object. The range-oriented gating filter is constructed as

[0186]

[0187] Where α1 is a threshold. For any range gate's range Doppler domain signal, if the difference between the estimated Doppler frequency modulation value and the theoretical value of the stationary object frequency modulation value is less than α1, W1(t) is set to 0. After the echo data in the range Doppler domain is weighted by W1(t), the stationary object clutter can be suppressed.

[0188] Next, in order to suppress wall clutter, based on The distance-gated filter is constructed as

[0189]

[0190] α2 is a threshold. Since the theoretical value of the Doppler frequency modulation of the wall is 0, α2 is also a threshold to measure the difference between the estimated value of the frequency modulation and the theoretical value of the wall frequency modulation. For the range Doppler domain signal of any range gate, if the estimated value of the Doppler frequency modulation is less than α2, the filter is set to 0.

[0191] At this point, the stationary target and wall clutter are suppressed in the range Doppler domain, and the moving target is retained. The echo expression in the range Doppler domain after the stationary clutter is suppressed is:

[0192] S a (f a ,t)=S rcmc (f a ,t)×W1(t)×W2(t) (71)

[0193] Step S5: construct an azimuth compression filter according to the Doppler modulation rate estimation value to achieve focused imaging of indoor moving targets. The specific method is as follows:

[0194] according to The azimuth matched filter is constructed as

[0195]

[0196] in is the estimated frequency modulation rate of each distance unit. Multiplying equation (71) and equation (72) yields

[0197] S a (f a ,t)=S a (f a ,t)H ac (f a ,t) (73)

[0198] Transform equation (73) inversely to obtain

[0199]

[0200] It can be seen from formula (74) that the moving target is focused in azimuth and distance.

[0201] In order to verify the effectiveness of the proposed method, a practical experiment was designed for analysis. The experimental photos and radar detection schematic diagram are shown in Figure 2. Figure 4 (a) and Figure 4As shown in (b), the speed of the human-borne radar is 4m / s, flying from (-8,0)m to (8,0)m, the distance coordinate of the front wall is 3m, the azimuth length of the room is 16m, the distance length is 12m, there is a stationary person and a person moving in the azimuth in the room, the stationary person's coordinate is (-4,9)m, during the flight of the drone, the moving person's coordinate moves from (-2,7)m to (4,7)m. The system parameters of the drone-borne radar are shown in Table 1.

[0202] Table 1 Radar system parameter settings

[0203]

[0204] Execute step S1 to transform the echo data after range compression into the range Doppler domain.

[0205] Execute step S2 to calculate the range migration of the target echo and perform range migration correction by sinc interpolation.

[0206] Execute step S3 to extract the echo data of each range gate and estimate its Doppler modulation rate using fractional Fourier transform.

[0207] Step S4 is executed to construct a range-gated filter according to the estimated value of the modulation rate to suppress stationary clutter.

[0208] Execute step S5 to construct an azimuth compression filter according to the estimated value of the modulation rate to achieve focusing imaging of the moving target.

[0209] The experimental results of frequency modulation estimation and clutter suppression are shown in Figure 5 As shown, Figure 5 (a) is the variation curve of the estimated value of the frequency modulation and the theoretical value of the frequency modulation of the stationary target along the distance direction. Figure 5 (b) is the result of incoherent accumulation of echo data along the azimuth frequency before suppressing stationary clutter and after suppressing stationary clutter according to the frequency modulation. It can be seen from the figure that the estimated frequency modulation of the wall echo is approximately 0, the estimated frequency modulation of the stationary target is approximately equal to the theoretical value, and the estimated frequency modulation of the moving target is lower than the theoretical value of the frequency modulation of the stationary target. This is because the moving target moves in the same direction as the radar azimuth in the experiment. Figure 5 (b) shows that the wall and stationary targets are suppressed, while the energy of the moving targets is preserved.

[0210] In order to highlight the superiority of the algorithm, the imaging results of this method are compared with the original imaging results, such as Figure 6 The original imaging result is shown in Figure 6 (a), we can see that there are walls, stationary targets, and moving targets at the same time, and the target is almost submerged in the wall clutter; Figure 6 (b) shows that the clutter generated by walls and stationary objects is suppressed, and the imaging effect of moving targets is good.

[0211] The processing result of the present invention basically suppresses the static clutter of the wall, and has a good focusing effect on the moving target, which provides technical support for the moving target detection of the wall-penetrating radar on UAVs.

[0212] The above specific embodiments only describe the design principle of the present invention. The shapes and names of the components in the description may be different and are not limited. Therefore, those skilled in the art in the field of the present invention may modify or replace the technical solutions recorded in the above embodiments; and these modifications and replacements do not deviate from the creative purpose and technical solutions of the present invention and should all fall within the protection scope of the present invention.

Claims

1. A method for detecting indoor moving targets based on Doppler modulation rate profile and range gating filtering, characterized in that: include: Step S1: obtaining the original data of the echo of the moving target of the wall-penetrating radar carried by the UAV, performing distance compression on the original data, and transforming the echo data after the distance compression into the range Doppler domain; Step S2: Calculate the precise range migration of the moving target according to the range Doppler domain expression of the target echo, and perform range migration correction by sinc interpolation; Step S3: obtaining echo data of each range gate by range indexing, and estimating Doppler modulation rate by fractional Fourier transform; Step S4: constructing a range-gated filter according to the Doppler modulation frequency estimation value to suppress clutter of walls and stationary objects; Step S5: construct an azimuth compression filter according to the Doppler modulation rate estimation value to achieve focused imaging of indoor moving targets.

2. The method according to claim 1, characterized in that In step S1, the linear frequency modulation signal transmitted by the drone-borne radar is expressed as: where rect(·), t, T p 、f c , K r They are rectangular gate function, fast time variable, pulse width, signal center frequency and modulation frequency of linear frequency modulation signal respectively; after receiving the radar echo data, it is compressed by distance to obtain s r (t a ,t)=s w (t a ,t)+s sta (t a ,t)+s mov (t a ,t) (2) where t a is the slow time variable, s w (t a ,t),s sta (t a ,t) and s mov (t a ,t) are wall echo, indoor stationary target echo and moving target echo respectively.

3. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S1, the moving target echo is expressed as where w a (t a ) is the azimuth rectangular gate function, B is the signal bandwidth, λ is the signal wavelength corresponding to the signal center frequency, c is the speed of light, R(t a ) is the slant range between the target and the radar, which can be written as Where V is the speed of the drone-mounted wall-penetrating radar, x0 and y0 are t a = The azimuth and range positions of the target relative to the radar at time 0, v x 、v y are the target’s azimuth velocity and range velocity, R w is the additional propagation distance of the signal generated during the wall penetration process, which is relative to R(t a ) is small and can be ignored.

4. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S1, the slow time when the slant distance is the smallest can be obtained by deriving equation (4): The corresponding shortest slope distance is The static target echo is essentially the moving target echo in v x =0, v y = 0, it can be written as follows Ideally, the wall echo can be regarded as the result of the superposition of multiple closely adjacent static point target echoes. When the complex reflectivity of these point targets is the same, the wall echo after range compression can be expressed as Where R wall R represents the distance between the UAV-mounted wall-penetrating radar and the wall. wall Can be considered as a constant.

5. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S1, the distance of equation (3) is transformed into Fourier transform to obtain where f r represents the distance frequency, and the azimuth Fourier transform of equation (9) can be obtained where f a represents the Doppler frequency, f ac and D(f a ) represent the Doppler center frequency and range migration factor respectively, and their expressions are as follows The square root of the phase in equation (10) is r / f c = 0 Taylor expansion The quadratic and cubic distance frequencies in formula (13) will cause the distance of the two-dimensional time domain echo to be defocused upward. In order to solve this problem, a quadratic distance compression filter is constructed as follows: Multiplying equation (13) and equation (14) yields The echo expression of the echo data in the range Doppler domain obtained by inverse Fourier transform of equation (15) is:

6. The method for detecting indoor moving targets based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S2, the distance migration amount of the moving target is calculated according to formula (16): Similarly, for a stationary target, the range migration is From equations (17) and (18), we can see that the range migration of a stationary target is related to the radar speed, and the range migration of a moving target is related to both the radar speed and the speed of the moving target.

7. The method for detecting indoor moving targets based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S2, range cell migration correction (RCMC) is performed by sinc interpolation; this method is expressed as Where S rcmc (n a ,n r ) is the result after range migration correction, S mov (n a ,n r +i) is the echo data before range migration correction, sinc(Δn r -i) is the interpolation kernel function, which can be regarded as the weight of the original sampling signal, n a is the azimuth index of the echo data, n r is the range index of the echo data, Δn r is ΔR sta (f a ) corresponds to the distance migration, P is the number of sampling points of the interpolation function; the result of Sinc interpolation is Where W r (f a ) is the azimuth gate function after range migration correction; applying azimuth IFFT to (20) yields the RCMC result in the two-dimensional time domain:

8. The method for detecting indoor moving targets based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S3, the Taylor expansion of equation (12) is obtained: Substituting equation (22) into equation (20), we get Where K mov is the Doppler modulation frequency of the moving target; the one-dimensional azimuth signal can be obtained by indexing the range direction of equation (23): Transforming equation (8) to the range Doppler domain yields Where δ(f a ) represents the impact signal; the one-dimensional azimuth signal of the range gate where equation (25) is located is From equations (24) and (26), it can be seen that after range migration correction, the target echo can be directly extracted in the range Doppler domain and the Doppler modulation rate can be estimated.

9. The method for detecting indoor moving targets based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S3, the fractional Fourier transform of equation (24) yields Where α is the rotation angle and u represents the fractional domain frequency. When the following relationship is satisfied, a peak value appears in equation (27): That is, according to the angle α corresponding to the peak * Azimuth modulation frequency can be estimated Similarly, the frequency modulation of the wall is estimated to be 10. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S4, the Doppler modulation frequency estimation value Theoretical value of frequency modulation with stationary object Subtract ΔK(t) measures the difference between the estimated frequency modulation and the theoretical value of the frequency modulation of a stationary object.

11. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S4, the distance gating filter is constructed as follows: Where α1 is a threshold. For the range Doppler domain signal of any range gate, if the difference between the estimated value of the Doppler frequency modulation and the theoretical value of the frequency modulation of the stationary object is less than α1, W1(t) is set to 0. After the echo data in the range Doppler domain is weighted by W1(t), the suppression of the stationary object clutter can be achieved.

12. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S4, in order to suppress wall clutter, based on the Doppler modulation frequency estimation value The distance-gated filter is constructed as Among them, α2 is a threshold; since the theoretical value of the wall Doppler modulation frequency is 0, α2 is also a threshold for measuring the difference between the modulation frequency estimate and the wall modulation frequency theoretical value; for the range Doppler domain signal of any range gate, if the Doppler modulation frequency estimate is less than α2, the filter is set to 0.

13. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: In step S4, the echo expression in the range Doppler domain after stationary clutter suppression is: S a (f a ,t)=S rcmc (f a ,t)×W1(t)×W2(t) (34)。 14. The indoor moving target detection method based on Doppler modulation rate profile and range gating filtering as claimed in claim 1, characterized in that: The specific method of step S5 is: Doppler modulation rate estimate The azimuth matched filter is constructed as Multiplying (34) and (35) yields S(f a ,t)=S a (f a ,t)H ac (f a ,t) (36) Transform equation (36) inversely to obtain It can be seen from formula (37) that the moving target is focused in azimuth and distance.

Images