77GHz millimeter wave radar rapid imaging method

Abstract

As the distance between the radar and the target changes along with time, azimuth signals corresponding to the point target will be located in different distance resolution units. In order to solve the problem that the calculation amount is too large due to the fact that two-dimensional interpolation is needed for range migration correction in the SAR imaging processing process of a 77GHz millimeter-wave radar, the invention provides a 77GHz millimeter-wave radar rapid SAR imaging algorithm based on table look-up. According to the method, the two-dimensional (range direction and azimuth direction) sinc interpolation factor required to be used in range migration correction processing is pre-calculated and locally stored, and the value of the sinc interpolation factor can be directly used only by calling the position relation of a table look-up method in the SAR imaging processing process, so that the step of repeatedly calculating the interpolation factor in each cycle is omitted, and the processing efficiency is improved. The purpose of effectively reducing the calculation amount of the 77GHz millimeter wave radar SAR imaging algorithm is achieved.

Description

Technical Field

[0001] The invention belongs to the technical field of radar imaging, and in particular relates to a 77 GHz millimeter wave radar rapid imaging method. Background Art

[0002] 77GHz millimeter wave radar adopts the frequency modulated continuous wave (FMCW) system, which has the advantages of low cost, light weight, low power consumption, high spatial resolution, large working bandwidth, etc. It is not afraid of weather influences such as rain, fog, dust and haze, and has all-day and all-weather detection capabilities and long detection distance. It has become one of the core sensors of Advanced Driver Assistance System (ADAS) technology and is vigorously promoted by most domestic and foreign automobile manufacturers. In recent years, many scholars have begun to try to use 77GHz millimeter wave radar as a synthetic aperture radar (SAR) payload to carry out high-resolution imaging algorithm verification work on vehicle-mounted platforms. For example, Feger et al. reported a vehicle-mounted SAR system based on a 77 GHz millimeter-wave radar and experimental results, verifying the feasibility of using a 77 GHz millimeter-wave radar to obtain high-resolution SAR images [R.Feger, A.Haderer, A.Stelzer. Experimental verification of a 77-GHz synthetic aperture radar system for automotive applications [C]. IEEE MTT-S International Conference on Microwaves for Intelligent Mobility (ICMIM), March 2017: 111-114.]. However, there are not many reports on the use of 77 GHz millimeter-wave radar as a SAR imaging payload in public research at home and abroad. The main reason is that the 77 GHz millimeter-wave radar has a high operating frequency and a large available signal bandwidth, which makes it easy to achieve high-resolution imaging in the range direction. However, for SAR imaging processing, the higher the range resolution, the more serious the range cell migration (RCM) problem. Specifically, RCM is the relationship between the distance change between the radar carrier and the illuminated ground target, and this change relationship is valid for both spotlight SAR and strip SAR. From the SAR imaging geometry, we know that since the radar moves in azimuth with the carrier platform, the distance between the radar and the target changes over time. This change causes the same target echo signal to have different positions in the fast time (range direction) corresponding to the sampling pulses in different azimuths. When the distance difference in this position exceeds one range resolution unit, the azimuth signal corresponding to the point target will be in a different range resolution unit, which is the fundamental reason for the RCM phenomenon.In order to effectively eliminate or correct RCM to obtain fully focused (range and azimuth) SAR imaging results, a two-dimensional interpolation operation must be performed. However, two-dimensional interpolation requires a lot of calculations. Therefore, it is necessary to invent a fast imaging method to reduce the amount of calculations in the 77GHz millimeter-wave radar SAR imaging algorithm. Summary of the invention

[0003] In order to solve the problem of excessive calculation amount in the SAR imaging process of 77GHz millimeter wave radar due to the need for two-dimensional interpolation of range migration correction, the present invention proposes a 77GHz millimeter wave radar fast SAR imaging algorithm based on table lookup. This method pre-calculates and stores the two-dimensional (range and azimuth) sinc interpolation factors required in the range migration correction process locally. In the SAR imaging process, the value of the sinc interpolation factor can be directly used by calling the position relationship of the table lookup method, eliminating the step of repeatedly calculating the interpolation factor in each cycle, thereby effectively reducing the calculation amount of the 77GHz millimeter wave radar SAR imaging algorithm. It includes the following steps:

[0004] Step 1: Perform range-wise Dechirp processing on the 77GHz millimeter-wave radar echo data to obtain s IF (n,τ).

[0005] Step 2: s IF (n,τ) performs distance FFT processing to obtain S r (n,f r ).

[0006] Step 3: S r (n,f r ) performs azimuth FFT processing to obtain S rd (f a ,f r ).

[0007] Step 4: S rd (f a ,f r ) performs distance migration correction processing based on table lookup, and obtains

[0008] Step 5: Perform azimuth pulse pressure processing to obtain

[0009] Step 6: Perform azimuth IFFT processing to obtain s rd (x,y).

[0010] Specifically:

[0011] (1) Establishing signal model

[0012] Assume that the analog pulse linear frequency modulation signal emitted by the 77GHz millimeter wave radar is in the form of:

[0013]

[0014] Where, T p is the pulse width, k r is the modulation frequency, f c is the carrier frequency, then the transmitted signal bandwidth is B = k r T p After being reflected by the point target P(x0, y0) in the imaging area, Figure 1 As shown, the echo signal received by the radar is:

[0015]

[0016] In the formula, A p is the scattering intensity of point target P, T a is the synthetic aperture time, t k is the fast time (distance direction), t m =n×T p is the slow time (azimuth), n is the nth pulse, t=t m +t k is the total time, r(t) is the distance between the radar and the point target P, and v is the speed of the radar platform.

[0017] It should be pointed out here that, unlike the traditional radar, which first mixes to remove the carrier frequency and then performs pulse compression, the 77GHz millimeter wave radar uses dechirp processing to achieve distance measurement of the target. Specifically, it combines the received signal (RX) and the transmitted signal (TX) with a certain delay (the delay corresponds to the reference distance) through a mixer to generate the corresponding intermediate frequency (IF) signal. Figure 2 shown.

[0018] After the range-direction Dechirp processing, the echo signal of the nth pulse (azimuth sampling) of the millimeter-wave radar is the echo signal model for SAR imaging processing, which can be expressed as:

[0019]

[0020] In the formula, r c is the reference distance, △r=r(t m )-r c , τ = t k -2r c / c.

[0021] (2) Distance migration

[0022] Performing distance Fourier transform on equation (4) with respect to τ yields:

[0023]

[0024] In the formula, sinc(x)=sin(πx) / πx is the sinc function. The integral term in formula (5) can be expressed as:

[0025]

[0026] Therefore, the phase term Φ in equation (7) b It can be expressed as:

[0027]

[0028] Since the RCM phenomenon is applicable to both spotlight SAR and strip SAR, in order to simplify the description, the present invention takes the positive side-view SAR mode as an example to illustrate the distance change process between the radar and the point target in the SAR imaging process. Figure 3 (a) shows the geometric relationship between the radar and the point target P in the positive side-view SAR mode. In the figure, the platform height H = 0, and the point target P is located on the y-axis (distance direction), that is, x0 = 0. At positions S and E, the distance between the carrier and the point target P is the largest, and at the aperture center M, the distance between the carrier and the point target P is the smallest, denoted as r0, then r0 = y0. Therefore, Figure 3 The distance relationship between the radar and the point target P in (a) can be expressed as:

[0029]

[0030] Furthermore, we expand r(t) by a second-order Taylor series around t=0, and we get:

[0031]

[0032] At this time, substitute the result of equation (9) back into equation (5), and temporarily ignore the influence of the phase term, then we have:

[0033]

[0034] In the formula, △r0=r0-r c From formula (9), it is not difficult to find that for the same point target, at different slow times t m , the position of the sinc term peak is different, showing a parabola (about the slow time t m The quadratic function of the distance migration path is Figure 3 (as shown in (b)).

[0035] (3) Range Migration Correction

[0036] The purpose of range migration correction is to turn this parabola into a straight line. The process is shown in the attached figure. Fig. 9 shown.

[0037] For 77GHz millimeter wave radar, f r <<f c , 4πk r △r 2 <<c 2 , then formula (7) can be approximately expressed as:

[0038]

[0039] Where λ = c / f c is the radar wavelength. The first phase term in equation (11) is a constant, and the second phase term is the azimuth linear frequency modulation phase generated by the Doppler frequency shift. Therefore, the azimuth frequency modulation obtained by millimeter wave radar SAR imaging processing is:

[0040]

[0041] Substitute the results of equations (11) and (12) into equation (10), and then perform a slow time dimension (azimuth) Fourier transform on equation (10). According to the residence phase principle (PSP), the signal form in the range-azimuth two-dimensional frequency domain can be obtained, namely:

[0042]

[0043] In the formula, B a =|k a T a | is the bandwidth of the azimuth composite signal, and |·| is the absolute value operator.

[0044] The second term in the parentheses of the sinc term in formula (13) is the range migration factor △R that causes the RCM phenomenon in the range-azimuth two-dimensional frequency domain, that is:

[0045]

[0046] At this time, the distance migration correction of △R can be achieved by using the two-dimensional sinc interpolation method, that is:

[0047]

[0048] In the formula, △R=△R int +du, is the positive part of △R, This is a rounding operation, where du is the decimal part of △R.

[0049] Obviously, equation (15) requires that for each azimuth frequency f a and each distance frequency f r Perform N-point sinc interpolation operation, that is, use the sinc function with a window length of N points and S rd Since the interpolation function sinc contains trigonometric operations, repeated calculations are very time-consuming, so you can consider pre-calculating all the interpolation function values ​​sinc(du-n) and the integer part of the distance migration factor △R int and the decimal part du, establish the corresponding lookup table {sinc_du}, {△R int}, {du}, and then use the table lookup method to replace the real-time calculation process of the above function.

[0050] The calculation process of the fast range migration correction method based on table lookup is as follows:

[0051]

[0052] If the range migration factor is corrected by the fast range migration correction method based on table lookup, the corrected range-azimuth two-dimensional frequency domain function Should be:

[0053]

[0054] (4) Azimuthal pulse pressure

[0055] The azimuth pulse pressure reference function is:

[0056] h a (t m )=exp(jπk a t m 2 ) (17)

[0057] Where, T a is the synthetic aperture time, and x0 is the range coordinate of the point target. According to the dwell phase principle (PSP), we can get:

[0058]

[0059] Taking the conjugate of equation (18) and multiplying it with equation (16) can realize the azimuthal pulse pressure, that is,

[0060]

[0061] In the formula, (·) * is the conjugate operation.

[0062] (5) Azimuth IFFT

[0063] By performing inverse Fourier transform on equation (19), the target imaging result in the final scene can be obtained as follows:

[0064]

[0065] Where x = vt m is the azimuth coordinate scale (unit, m), y = cf r / 2k r is the range coordinate scale (unit, m), and x0 is the azimuth coordinate of the point target P. After 77GHz millimeter wave radar fast SAR imaging processing, the position where the sinc peak of the point target P appears on the slow time axis should be The position where the sinc peak appears on the range Doppler frequency axis should be f r P =2k r (r c -y0) / c.

[0066] The beneficial effects of the present invention are

[0067] The present invention adopts a table-based sinc interpolation method in the range migration correction step, which saves the operation process of calculating the sinc interpolation factor in each cycle, and effectively improves the computational efficiency of the 77GHz millimeter wave radar imaging algorithm. The simulation results show that under the same computer hardware conditions and the same simulation parameter scenarios, the computational time required for the fast SAR imaging algorithm proposed by the present invention is about 11% of that of the SAR imaging algorithm using direct sinc interpolation, which effectively improves the computational efficiency of the SAR imaging algorithm and is conducive to the promotion and application of 77GHz millimeter wave radar SAR imaging technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 Schematic diagram of the geometric relationship of UAV SAR imaging.

[0069] Figure 2 This is a diagram of the Dechirp processing process.

[0070] Figure 3 Schematic diagram of the distance migration process.

[0071] Figure 4 Schematic diagram of the distance migration correction process.

[0072] Figure 5 This is the 77GHz millimeter wave radar SAR imaging processing flow chart.

[0073] Figure 6 is the two-dimensional time domain amplitude spectrum |s IF (n,τ)|Schematic diagram.

[0074] Figure 7 is the range frequency domain-azimuth time domain amplitude spectrum |S r (n,f r )|Schematic diagram.

[0075] Figure 8 is the two-dimensional frequency domain amplitude spectrum |S rd (f a ,f r )|Schematic diagram.

[0076] Fig. 9 is the two-dimensional frequency domain amplitude spectrum Schematic diagram.

[0077] Fig.10 The two-dimensional frequency domain amplitude spectrum of the azimuth pulse pressure processing Schematic diagram.

[0078] Fig.11 This is the imaging result after range migration correction.

[0079] Fig.12 This is the imaging result without correction for range migration. DETAILED DESCRIPTION

[0080] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] The specific implementation of the 77 GHz millimeter wave radar fast imaging algorithm of the present invention is as follows:

[0082] Step 1: Set the echo data simulation parameters of the 77GHz millimeter wave radar and the carrier platform

[0083] Speed ​​of light: c = 3 × 10 8 m / s; radar frequency: f c =77GHz; signal bandwidth: B = 2GHz; sampling rate: f s =10MHz; pulse repetition period: T p =1.024×10 -4 s; platform speed: v = 10m / s; platform height: H = 0m; reference distance: r c=0m; synthetic aperture length: Ls=10m; number of range points: rN=1024; number of azimuth points: aN=16384; sinc interpolation factor: N=8; number of targets: 5; target range coordinates: [20, 20, 60, 60, 40]m; target azimuth coordinates: [-2, 2, -4, 4, 0]m; computer processor: i7-4790 CPU@3.6GHz ; Computer memory (RAM): 16GB; Computer operating system: Windows 7 Enterprise Edition 64-bit Service Pack 1.

[0084] Step 2: Generate range-direction Dechirp echo data s according to the simulation parameters set in step 1 IF (n,τ), and obtain the two-dimensional time domain amplitude spectrum |s IF (n,τ)|, as attached Figure 6 shown.

[0085] Step 3: s IF (n,τ) performs FFT processing in the range direction to obtain the frequency domain in the range direction and the time domain amplitude spectrum in the azimuth direction |S r (n,f r )|, as attached Figure 7 As shown in the figure, after the range FFT processing, the five simulated targets form a wide range of parabolic sinc peak curves in the azimuth time domain spectrum, that is, there is a range migration phenomenon in the azimuth direction.

[0086] Step 4: S r (n,f r ) performs azimuth FFT processing to obtain a two-dimensional frequency domain amplitude spectrum |S rd (f a ,f r )|, as attached Figure 8 As shown in the figure, after the range FFT processing and the azimuth FFT processing, the five simulated targets still have parabolic sinc peak curves in the azimuth frequency domain spectrum, but the azimuth width has become narrower, that is, the range migration phenomenon still exists after the azimuth FFT processing.

[0087] Step 5: Use the direct calculation sinc interpolation method and the table-based sinc interpolation method to calculate S rd (f a ,f r ) to perform range migration correction processing to obtain a two-dimensional frequency domain amplitude spectrum that eliminates the range migration phenomenon As attached Fig. 9The results in the figure show that after the sinc interpolation processing, the range migration problem of the five simulated targets in the azimuth frequency domain spectrum has been corrected, and the five simulated targets form three straight sinc peak curves in the azimuth frequency domain spectrum.

[0088] According to statistics, the time required for the direct calculation sinc interpolation method is 178.6518 seconds, and the time required for the table lookup sinc interpolation method proposed in the present invention is 20.2852 seconds, which is about 11% of the former. This shows that the fast imaging method proposed in the present invention has higher sinc interpolation operation efficiency.

[0089] Step 6: Perform azimuth pulse pressure processing to obtain a two-dimensional frequency domain amplitude spectrum of azimuth pulse pressure processing As attached Fig.10 shown.

[0090] Step 7: Perform azimuth IFFT processing to obtain the range frequency domain-azimuth time domain amplitude spectrum after SAR imaging|s rd (x,y)|, as attached Fig.11 As a comparison, Fig.12 The figure shows the SAR imaging result obtained by directly processing the azimuth pulse compression without correcting the range migration phenomenon. By comparison, we can easily find that Fig.11 The five simulated targets after range motion correction have achieved good focusing effects in both distance and azimuth dimensions. Fig.12 The imaging results of the five simulated targets show certain defocusing problems, especially for targets at close range, where the defocusing problem is more serious.

[0091] The present invention is not limited to the above specific implementations, and the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made to the above implementations based on the technical essence of the present invention shall be included in the protection scope of the present invention.

Claims

1. A 77 GHz millimeter wave radar rapid imaging method, characterized in that: It includes the following six steps: Step 1: Perform range-wise Dechirp processing on the 77GHz millimeter-wave radar echo data to generate a range-wise Dechirp echo signal s IF (n,τ), and obtain the two-dimensional time domain amplitude spectrum |s IF (n,τ)|; Step 2: s IF (n,τ) performs FFT processing in the range direction to obtain the frequency domain in the range direction and the time domain amplitude spectrum S in the azimuth direction r (n,f r )|; Step 3: S r (n,f r ) performs azimuth FFT processing to obtain a two-dimensional frequency domain amplitude spectrum |S rd (f a ,f r )|; Step 4: Use direct calculation interpolation method and table-based interpolation method to calculate S rd (f a ,f r ) to perform range migration correction processing to obtain a two-dimensional frequency domain amplitude spectrum that eliminates the range migration phenomenon Step 5: Perform azimuth pulse pressure processing to obtain the two-dimensional frequency domain amplitude spectrum of azimuth pulse pressure processing Step 6: Perform azimuth IFFT processing to obtain the range frequency domain-azimuth time domain amplitude spectrum after SAR imaging|s rd (x,y)|.

2. A 77 GHz millimeter wave radar rapid imaging method according to claim 1, characterized in that: Before step 1, the signal model is also established: (1) Establishing signal model Assume that the analog pulse linear frequency modulation signal emitted by the 77GHz millimeter wave radar is in the form of: Where, T p is the pulse width, k r is the modulation frequency, f c is the carrier frequency, then the transmitted signal bandwidth is B = k r T p ; After being reflected by the point target P(x0,y0) in the imaging area, the echo signal received by the radar is: In the formula, A t is the scattering intensity of point target P, T a is the synthetic aperture time, t k is the distance to speed time, t m =n×T p is the azimuth slow time, n is the nth pulse, t=t m +t k is the total time, r(t) is the distance between the radar and the point target P, and v is the speed of the radar platform.

3. A 77 GHz millimeter wave radar rapid imaging method according to claim 2, characterized in that: In the step 1, the echo signal of the nth pulse after the range dechirp process can be expressed as: In the formula, r c is the reference distance, △r=r(t m )-r c , τ = t k -2r c / c.

4. A 77 GHz millimeter wave radar rapid imaging method according to claim 3, characterized in that: In the step 2, the Fourier transform of equation (4) with respect to τ is performed to obtain: Where sinc(x)=sin(πx) / πx is the sinc function; the integral term in equation (5) can be expressed as: Therefore, the phase term Φ in equation (5) b It can be expressed as: Assume that the platform height H = 0, and the point target P is located on the y-axis, that is, x0 = 0; at the aperture center M, the distance between the carrier and the point target P is the smallest, recorded as r0, then r0 = y0, and the distance relationship between the radar and the point target P can be expressed as: Expand r(t) by a second-order Taylor series around t=0, and we get: At this time, substitute the result of formula (9) back into formula (5), and we have: In the formula, △r0=r0-r c , it can be seen from formula (9) that the same point target, at different slow times t m , the position where the sinc term peak appears is different, showing as a parabola.

5. A 77 GHz millimeter wave radar rapid imaging method according to claim 4, characterized in that: In step 3, for the 77 GHz millimeter wave radar, f r <<f c , 4πk r △r 2 <<c 2 , then formula (7) can be approximately expressed as: Where λ = c / f c is the radar wavelength; the first phase term in equation (11) is a constant, and the second phase term is the azimuth linear frequency modulation phase generated by the Doppler frequency shift, so the azimuth frequency modulation is: Substitute the results of equations (11) and (12) into equation (10), and then perform Fourier transform on equation (10) in the azimuth slow time dimension. According to the resident phase principle, the signal form in the range-azimuth two-dimensional frequency domain can be obtained, namely: In the formula, B a =|k a T a | is the bandwidth of the azimuth composite signal, and |·| is the absolute value operator.

6. A 77 GHz millimeter wave radar rapid imaging method according to claim 5, characterized in that: In step 4, the second term in the parentheses of the sinc term in equation (13) is the range migration factor △R that causes the RCM phenomenon in the range-azimuth two-dimensional frequency domain, that is: At this time, the distance migration correction of △R can be achieved by using the two-dimensional sinc interpolation method, that is: In the formula, △R=△R int +du, is the positive part of △R, For the rounding operation, du is the decimal part of △R; From formula (15), we can see that for each azimuth frequency f a and each distance frequency f r Perform N-point sinc interpolation operation, that is, use the sinc function with a window length of N points and S rd Multiply and add; Pre-calculate all the interpolation function values ​​​​sinc(du-n) and the integer part of the distance migration factor △R int and the decimal part du, establish the corresponding lookup table {sinc_du}, {△R int }, {du}, the table lookup method is used to replace the real-time calculation process of the function.

7. A 77 GHz millimeter wave radar rapid imaging method according to claim 6, characterized in that: In step 5, the azimuth pulse pressure reference function is: h a (t m )=exp(jπk a t m 2 ) (17) Where, T a is the synthetic aperture time, x0 is the range coordinate of the point target, according to the dwell phase principle, we can get: By taking the conjugate of equation (18) and multiplying it with equation (16), the azimuthal pulse pressure can be obtained, namely: In the formula, (·) * is the conjugate operation.

8. A 77 GHz millimeter wave radar rapid imaging method according to claim 7, characterized in that: In step 6, the inverse Fourier transform is performed on equation (19) to obtain the final target imaging result in the scene: Where x = vt m is the azimuth coordinate scale, y=cf r / 2k r is the distance coordinate scale, x0 is the azimuth coordinate of the point target P; After the 77GHz millimeter wave radar rapid imaging processing, the position of the sinc peak of the point target P on the slow time axis should be The position where the sinc peak appears on the range Doppler frequency axis should be

Images