CN113885003A - Target detection method based on non-coherent accumulation mode of coherent radar - Google Patents

Abstract

The invention discloses a target detection method based on a non-coherent accumulation mode of coherent radar, which comprises the following steps: s1, sampling to obtain a frame of two-dimensional echo data; s2, performing pulse compression on each data of the S1; s3, carrying out distance walking calibration on the data of the S2; s4, carrying out coherent accumulation on the data of S3; s5, repeating S1-S4 to obtain multi-frame coherent accumulation data; s6, performing interframe alignment on the multiframe coherent accumulation data obtained in the S5; s7, performing non-coherent accumulation on the multi-frame coherent accumulation data after the inter-frame alignment; and S8, carrying out target detection based on the non-coherent accumulated data of S7. The advantages are that: the method realizes distance walking calibration by adopting an autocorrelation method between pulses, and realizes the detection of a moving target accumulated by long-time phase coherence; according to the method, after peak value target alignment is carried out on a two-dimensional data plane between adjacent multiframes, non-coherent accumulation is carried out, and then target detection is carried out, so that the purpose of improving the signal-to-noise ratio is achieved.

Description

Target detection method based on non-coherent accumulation mode of coherent radar

Technical Field

The invention relates to the technical field of target detection, in particular to a target detection method based on a non-coherent accumulation mode of a coherent radar.

Background

Space-based warning radar is an important application field for the development of space-based radar. The space-based warning radar has an important significance in realizing a moving target indication function, and a target detection method of the space-based warning radar becomes a hotspot of radar signal processing research. The accumulation of weak target energy is the main research direction for improving the signal to noise ratio of the echo and realizing the detection of the weak target. The method is limited by the design requirements of lightweight and low power consumption of the space-based radar, the transmitting power, the antenna gain and the like of the space-based radar are limited, and when a target is alarmed, capacity accumulation is often realized through long-time coherent accumulation, so that a target for remote detection is realized. However, as an alarm radar, the priori knowledge is often less, and the priori information such as the speed of the target cannot be obtained. On the one hand, under the condition of long-time coherent accumulation, for a moving object with unknown moving speed, the moving object has the problems of distance walking, phase walking and the like within an accumulation period. On the other hand, under the design principle of lightweight and low power consumption of space-based radar, the total resource of a radar system is limited, and when detection is performed, the detection signal-to-noise ratio is not high, so that the detection signal-to-noise ratio needs to be further improved through an information processing method, and the target detection probability of the radar system is improved.

Aiming at the first problem, namely the problem of distance and phase walk of a moving target with unknown moving speed under the condition of long-time coherent accumulation, the current commonly used algorithm comprises a keystone algorithm and the like. However, the algorithms are huge in operation amount and limited in application to real-time processing platforms of the space-based radar. Meanwhile, for space-based warning radars, if the motion speed is unknown and the problem of Doppler frequency folding exists, the difficulty is increased for distance walking compensation and phase supplement. Under the condition that the motion information of the target is unknown, if the speed information of the target is to be obtained by an accumulation method, the general method is to set a large speed search range and a certain speed search step length, then carry out envelope cyclic shift on pulse echoes according to the calculated number of echo envelope walking caused by each search speed, then carry out non-coherent accumulation on the pulse echoes subjected to the envelope cyclic shift, and finally compare the accumulated results of each possible search speed, wherein the speed corresponding to the maximum accumulation result is the speed estimation value of the target. This has the consequence of being at the expense of a high number of search operations.

On the other hand, the signal-to-noise ratio of the detection needs to be further improved by an information processing method. For coherent radar, pulse compression and coherent accumulation are conventional signal processing methods for improving the signal-to-noise ratio. How to maximize the utilization of echo data among multiple frames and further improve the target detection performance is a technical problem which can be further researched.

Disclosure of Invention

The invention aims to provide a target detection method based on a non-coherent accumulation mode of coherent radar, which is mainly used for solving the problem of detection of a moving target and a small target under a space-based alarm radar; meanwhile, the method can utilize the distance walking information to carry out preliminary estimation of the target speed; performing coherent accumulation on the calibrated data to obtain an accumulated two-dimensional data plane; after peak value target alignment is carried out on two-dimensional data planes between adjacent multiframes, non-coherent accumulation is carried out, and then target detection is carried out, so that the purpose of improving the signal-to-noise ratio is achieved, and the target detection probability of the warning radar is improved.

In order to achieve the purpose, the invention is realized by the following technical scheme:

a target detection method based on a non-coherent accumulation mode of coherent radar comprises the following steps:

s1, sampling to obtain a frame of two-dimensional echo data;

s2, performing pulse compression on the sampling data of each pulse of the two-dimensional echo data in the step S1;

s3, carrying out distance walking calibration on the data after pulse compression;

s4, carrying out coherent accumulation on the data after the distance walk calibration;

s5, repeating the steps S1-S4 to obtain multi-frame coherent accumulation data;

s6, performing interframe alignment on the multiframe coherent accumulation data obtained in the step S5;

s7, performing non-coherent accumulation on the multi-frame coherent accumulation data after the inter-frame alignment;

s8, target detection is performed based on the non-coherent accumulated data of the step S7.

Optionally, in step S1, the two-dimensional echo data is in the form of an LFM signal, the number of pulses of each frame of the LFM signal is N, each pulse samples M points, and the pulse repetition period is T_rPulse width of T_PEach sub-pulse has a frequency modulation slope of k and a sub-pulse bandwidth of B, and k is B/T_P(ii) a Setting the initial radial distance of the uniform motion target as R₀Radial velocity of motion v₀The target distance of the ith pulse is delayed by tau_i＝2(R₀+v₀iT_r) C, then the target baseband echo signal s_r(n, t) is:

wherein:

in the above formula, c represents the speed of light, and the length of t is M;

simultaneously, noise Noi (n, t) with a certain signal-to-noise ratio is superposed on a target baseband echo signal s_r(n, t) obtaining two-dimensional echo data SN_r(n,t)：

SN_r(n,t)＝Noi(n,t)+s_r(n,t) (3)。

Optionally, in step S2, the matching function of the pulse compression is:

the pulse compression is carried out in a frequency domain, and the two-dimensional echo data of the ith pulse is set as SN_ri(n, t), then the pulse compression algorithm is:

sc_i(n,t)＝IFFT(FFT(SN_ri(n,t))·conj(FFT(s_ref(t)))) (5)

wherein, sc_i(N, t) is pulse pressure data of the ith pulse, i is 1,2, …, N.

Optionally, in step S3, the distance walking calibration is performed on the data after pulse compression by using an inter-pulse cross-correlation method.

Optionally, in step S3, the pulse pressure data sc of the 1 st pulse is used₁(t) data sc obtained by compressing the pulse of the ith pulse with reference to_i(t) pulse pressure data s of the 1 st pulse_cm1(t) performing a cross-correlation process, a cross-correlation function R_1i(t) is:

find R_1i(t) position P where peak value is located_Xi，P_XiI.e. represents sc_i(t) relative to sc₁(t) the number of distance units, and pulse pressure data sc of the ith pulse_i(t) Cyclic Displacement P_XiPoint, obtain scx_i(t), implementing distance walk calibration; distance walk calibration is carried out on the data after pulse compression of each pulse, distance walk calibration of one frame of two-dimensional echo data is realized, and a new two-dimensional data plane S is formed_xc(n,t)。

sc_iCross-correlation function R of last pulse in (n, t) two-dimensional array_1N(t) position P where peak value is located_XNThe value is expressed in the Nth pulse and the 1 st pulseNumber of distance units P of distance traveled generated between pulses_XNThen, the corresponding walking distance dR is:

dR＝C·P_XN·t_s (7)

the estimated target speed v_eComprises the following steps:

optionally, in step S4, a new two-dimensional data plane S is processed_xc(n, t) performing coherent accumulation, and obtaining the column vector L corresponding to each distance unit_xcm(N) coherent accumulation, where M is 1,2, …, M, N is 1,2, …, N, i.e. the column vector L_xcm(n) performing FFT to obtain F_xcm(n):

F_xcm(n)＝FFT(L_xcm(n)) (9)

For each column vector L_xcm(n) all carry out coherent accumulation to form a new two-dimensional data plane F_xc1(n,t)。

Compared with the prior art, the invention has the following advantages:

the invention relates to a target detection method based on a non-coherent accumulation mode of a coherent radar, which is characterized in that under the condition of unknown target speed, distance walking calibration is realized by adopting an autocorrelation method based on pulses, and the detection of a moving target accumulated by long-time coherent is realized; meanwhile, the method can utilize the distance walking information to carry out preliminary estimation of the target speed; performing coherent accumulation on the calibrated data to obtain an accumulated two-dimensional data plane; after peak value target alignment is carried out on two-dimensional data planes between adjacent multiframes, non-coherent accumulation is carried out, and then target detection is carried out, so that the purpose of improving the signal-to-noise ratio is achieved, and the target detection probability of the warning radar is improved.

Furthermore, the method can realize the distance walk calibration of the moving target under the condition of long accumulation time without prior information such as the speed of the target and resolving the speed of the target, and can further meet the requirement of rapid detection of the alarm radar.

Furthermore, compared with other distance calibration algorithms, the distance walk calibration method provided by the method is small in calculation amount and more suitable for the signal processing capacity of the space-based platform.

Furthermore, the method can realize the preliminary estimation of the target speed by utilizing a single-frame signal, has short required time and can meet the requirement of the alarm radar for quick detection.

Furthermore, after coherent accumulation, non-coherent accumulation is carried out, so that the signal-to-noise ratio can be further improved, and the radar detection probability is improved.

Drawings

FIG. 1 is a schematic diagram of a target detection method based on a non-coherent accumulation mode of coherent radar according to the present invention;

FIG. 2(a) is a time domain simulation of an LFM echo signal;

FIG. 2(b) is a time domain diagram of an LFM echo signal superimposed with noise;

FIG. 3(a) is a two-dimensional data plane after pulse pressing a single pulse, i.e., the range dimension, in the echo signal of FIG. 2 (a);

FIG. 3(b) is a two-dimensional data plan view after a distance walk calibration of FIG. 3 (a);

FIG. 4(a) is a two-dimensional data plane obtained by direct coherent integration of FIG. 3 (a);

FIG. 4(b) is a two-dimensional data plane obtained by direct coherent integration of FIG. 3 (b);

FIG. 5(a) is data in a velocity dimension corresponding to a peak in the two-dimensional data plane of FIG. 4 (a);

FIG. 5(b) is data in a velocity dimension corresponding to a peak in the two-dimensional data plane of FIG. 4 (b);

FIG. 6(a) is a schematic diagram of a two-dimensional data plane after coherent accumulation of three consecutive frames of signals;

FIG. 6(b) is the distance dimension data corresponding to the peak of different frames;

FIG. 7(a) is a schematic diagram of a two-dimensional data plane after non-coherent accumulation of three consecutive frames of signals;

fig. 7(b) is a schematic diagram of distance dimension data corresponding to the peak of fig. 7 (a).

Detailed Description

The present invention will now be further described by way of the following detailed description of a preferred embodiment thereof, taken in conjunction with the accompanying drawings.

As shown in fig. 1, a method for detecting a target based on a coherent radar in a non-coherent accumulation mode according to the present invention includes:

and S1, sampling to obtain a frame of two-dimensional echo data.

In this embodiment, the signal form of the two-dimensional echo data is an LFM signal (linear frequency hopping signal), where N is the number of pulses of each frame of the LFM signal, M points are sampled per pulse, and the sampling interval is t_sWith a pulse repetition period of T_rPulse width of T_PEach sub-pulse has a frequency modulation slope of k and a sub-pulse bandwidth of B, and k is B/T_P(ii) a Setting the initial radial distance of the uniform motion target as R₀Radial velocity of motion v₀The target distance of the ith pulse is delayed by tau_i＝2(R₀+v₀iT_r) C, then the target baseband echo signal s_r(n, t) is:

wherein:

in the above formula, c represents the speed of light, and t has a length of M.

Simultaneously, noise Noi (n, t) with a certain signal-to-noise ratio is superposed on a target baseband echo signal s_r(n, t) obtaining two-dimensional echo data SN_r(n,t)：

SN_r(n,t)＝Noi(n,t)+s_r(n,t) (3)。

In this embodiment, the target detection method based on the non-coherent accumulation mode of the coherent radar in fig. 1 is used to simulate the echo accumulated by long-time coherent accumulation, and the echo data is a two-dimensional array, where the X dimension represents the pulse accumulation number (slow time dimension) and the Y dimension represents the distance dimension (fast time dimension). Eye settingThe distance of the echo signal is R₀60km, 800m/s speed, 256 pulse accumulation, 1kHz repetition frequency, 200us sub-pulse width, 40MHz clock, t sampling period_s25 ns. The number of sampling points in the range gate is 2000, the power ratio of the target echo signal to the noise is-40 dB, the echo time domain waveform without the noise superimposed is shown in fig. 2(a), the echo time domain waveform with the noise superimposed is shown in fig. 2(b), and as can be seen from fig. 2(a) and 2(b), the signal is completely submerged by the noise.

S2, the sampled data of each pulse of the two-dimensional echo data of step S1 is pulse compressed.

In step S2, the matching function of pulse compression is:

the pulse compression is carried out in a frequency domain, and the two-dimensional echo data of the ith pulse is set as SN_ri(n, t), then the pulse compression algorithm is:

sc_i(n,t)＝IFFT(FFT(SN_ri(n,t))·conj(FFT(s_ref(t)))) (5)

wherein, sc_i(N, t) is pulse pressure data of the ith pulse, i is 1,2, …, N.

In this embodiment, the two-dimensional data plane after pulse compression is shown in fig. 3(a), which is a two-dimensional data plane after pulse compression of a single pulse, i.e., a distance dimension, in the echo data of fig. 2 (a). It can be seen that the pulse pressure peak of different pulses has a distance walk condition due to the existing speed of the target.

And S3, performing distance walking calibration on the data after pulse compression.

In this embodiment, the inter-pulse cross correlation method is used to perform the range walk calibration on the data after the pulse compression.

In the step S3, the pulse pressure data sc of the 1 st pulse is used in the step S3₁(t) data sc obtained by compressing the pulse of the ith pulse with reference to_i(t) pulse pressure number of the 1 st pulseAccording to s_cm1(t) performing a cross-correlation process, a cross-correlation function R_1i(t) is:

find R_1i(t) position P where peak value is located_Xi，P_XiI.e. represents sc_i(t) relative to sc₁(t) the number of distance units, and pulse pressure data sc of the ith pulse_i(t) Cyclic Displacement P_XiPoint, obtain scx_i(t), implementing distance walk calibration; distance walk calibration is carried out on the data after pulse compression of each pulse, distance walk calibration of one frame of two-dimensional echo data is realized, and a new two-dimensional data plane S is formed_xc(n,t)。

The distance walk calibration was performed on fig. 3(a) using inter-pulse cross-correlation, and the two-dimensional data plane after calibration is shown in fig. 3 (b). As shown in fig. 3(b), a new two-dimensional data plane S is obtained after distance walk calibration by inter-pulse cross-correlation_xc(n, t). It can be seen that the range walk has been significantly corrected, with the range peaks of the different pulses all being at the same location.

sc_iCross-correlation function R of last pulse in (n, t) two-dimensional array_1N(t) position P where peak value is located_XNThe value represents the number of distance units P traveled between the Nth pulse and the 1 st pulse_XNThen, the corresponding walking distance dR is:

dR＝C·P_XN·t_s (7)

the estimated target speed v_eComprises the following steps:

from the above, the motion velocity of the target can be preliminarily estimated through the formula (8), and the target velocity in one frame signal can be quickly estimated.

And S4, performing coherent accumulation on the data after the distance walk calibration.

In the step S4, a new two-dimensional data plane S is processed_xc(n, t) performing coherent accumulation, and obtaining the column vector L corresponding to each distance unit_xcm(n) coherent accumulation, where M is 1,2, …, M (M corresponds to the column vector L)_xcmSubscript of (N), N is 1,2, …, N, i.e., a pair column vector L_xcm(n) performing FFT to obtain F_xcm(n):

F_xcm(n)＝FFT(L_xcm(n)) (9)

For each column vector L_xcm(n) all carry out coherent accumulation to form a new two-dimensional data plane F_xc1(n, t) (see FIG. 4 (b)).

Fig. 4(a) is a two-dimensional data plane obtained by directly performing coherent integration on fig. 3(a), and fig. 4(b) is a two-dimensional data plane obtained by performing coherent integration on data subjected to distance walk calibration on fig. 3 (b). The comparison shows that the two-dimensional plane data subjected to coherent accumulation after calibration has higher signal peak value and more concentrated energy.

In this embodiment, fig. 5(a) is data of a velocity dimension corresponding to a peak value of the two-dimensional data plane of fig. 4(a), and fig. 5(b) is data of a velocity dimension corresponding to a peak value of the two-dimensional data plane of fig. 4(b), that is, data of a velocity dimension corresponding to a peak value after coherent accumulation, as can be seen by comparison, peak energy after distance walk calibration is more concentrated, and the signal-to-noise ratio of fig. 5(b) is improved by 10 dB.

And S5, repeating the steps S1-S4 to obtain multi-frame coherent accumulation data.

In this embodiment, taking three frames as an example, that is, the number of frames K is 3, the data planes after three coherent accumulation are respectively F_xc1(n,t)、F_xc2(n,t)、F_xc3(n,t)。

F_xc1(n,t)、F_xc2(n,t)、F_xc3The peak positions of (n, t) are moved from frame to frame in the distance dimension, as shown in fig. 6(a) and 6 (b).

Fig. 6(a) and 6(b) show the position intention of the target after coherent accumulation of signals of three continuous frames, and characterize that when the moving speed of the target exists, even after intra-frame distance walk calibration, distance walk still exists between frames. As shown in fig. 6(a), three points in the graph represent positions of objects in different frames, respectively, and fig. 6(b) is distance dimensional data corresponding to peaks in different frames.

And S6, performing interframe alignment on the multiframe coherent accumulation data obtained in the step S5.

Taking the number of frames K as 3 as an example, find F respectively_xc1(n,t)、F_xc2(n,t)、F_xc3Position P where peak of (n, t) is located₁(x₁,y₁)、P₂(x₂,y₂)、P₃(x₃,y₃) In a first frame F_xc1(n, t) is taken as a reference, the number dP of inter-frame distance walking units generated by the target in the 2 nd frame and the 3 rd frame is calculated, and the calculation method comprises the following steps:

dP₂＝y₂-y₁ (10)

dP₃＝y₃-y₁ (11)

respectively converting the 2 nd frame data F_xc2(n, t), frame No. 3F_xc3(n, t) cyclically shifting dP in the range dimension₂、dP₃Each distance unit obtains a new two-dimensional data plane and records the new two-dimensional data plane as F_nxc2(n,t)、F_nxc2And (n, t), realizing the alignment among the data of multiple frames.

And S7, performing non-coherent accumulation on the multi-frame coherent accumulation data after the inter-frame alignment.

Specifically, non-coherent accumulation is performed on the aligned data to obtain a new two-dimensional data plane F_nacc(n,t)：

F_nacc(n,t)＝F_xc1(n,t)+F_nxc2(n,t)+...+F_nxK(n,t) (12)

FIG. 7(a) is a non-coherent accumulated two-dimensional data plane F obtained by performing modulo value calculation on three continuous frames of signals, performing data alignment with reference to the first frame, and accumulating the three frames of data_nacc(n, t). Fig. 7(b) shows distance dimension data corresponding to the peak of fig. 7(a), and comparing fig. 7(b) and fig. 5(b), it can be seen that the signal-to-noise ratio is significantly improved.

S8, not based on step S7And carrying out target detection on the data obtained by coherent accumulation. Namely: data F based on non-coherent accumulation_naccAnd (n, t), calculating the signal-to-noise ratio and carrying out target detection.

In summary, the target detection method based on the non-coherent accumulation mode of the coherent radar is particularly suitable for the field of space-based target detection and alarm, typical long-time coherent accumulation scenes, and scenes such as ultrahigh-speed target motion, high-resolution radar and the like. The method is mainly used for solving the problem of detecting a moving target and a small target under a space-based warning radar, and under the condition that the target speed is unknown, the method realizes distance walking calibration by adopting an autocorrelation method based on pulses, and realizes the detection of the moving target accumulated by long-time phase coherence; meanwhile, the method utilizes the distance walking information to carry out preliminary estimation of the target speed; performing coherent accumulation on the calibrated data to obtain an accumulated two-dimensional data plane; after the two-dimensional data planes between the adjacent multiframes are subjected to target alignment, non-coherent accumulation is carried out, and then target detection is carried out, so that the aim of improving the signal-to-noise ratio is fulfilled, and the target detection probability of the warning radar is improved.

Furthermore, the method can solve the problems of distance walking, phase walking and the like caused by long-time coherent accumulation, can further improve the signal-to-noise ratio of target detection through multi-frame non-coherent accumulation, realizes stable detection and tracking and improves the detection probability of a radar system; the operation of the method has the characteristics of relatively small operation amount and high detection speed, and further meets the requirement of the space-based alarm radar for quick alarm.

Furthermore, the method realizes distance walking calibration by adopting a cross-correlation method based on pulses, realizes detection of a moving target accumulated by long-time phase coherence, and has the characteristic of small operand.

Furthermore, the method realizes the data alignment of a two-dimensional data plane through the peak value alignment among the multi-frame signals, and then carries out the non-coherent accumulation among the multi-frame signals, thereby being suitable for the static target detection and the target detection with speed, improving the signal-to-noise ratio of the detection, and having simple calculation process and small operand.

Furthermore, the method for realizing distance walking calibration and peak alignment among multi-frame signals based on the inter-pulse autocorrelation method has the advantages that target speed priori knowledge is not needed, the method can be well suitable for application scenes of space-based warning radars, the rapid estimation of the motion speed of the target can be realized through the algorithm provided by the patent, and the method can be more suitable for application scenes without target priori knowledge.

Furthermore, the non-coherent accumulation method provided by the method can further improve the detection signal-to-noise ratio and improve the target detection probability of the radar on the basis of the traditional coherent accumulation.

While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.

Claims (7)

1. A target detection method based on a non-coherent accumulation mode of coherent radar is characterized by comprising the following steps:

s1, sampling to obtain a frame of two-dimensional echo data;

s2, performing pulse compression on the sampling data of each pulse of the two-dimensional echo data in the step S1;

s3, carrying out distance walking calibration on the data after pulse compression;

s4, carrying out coherent accumulation on the data after the distance walk calibration;

s5, repeating the steps S1-S4 to obtain multi-frame coherent accumulation data;

s6, performing interframe alignment on the multiframe coherent accumulation data obtained in the step S5;

s7, performing non-coherent accumulation on the multi-frame coherent accumulation data after the inter-frame alignment;

s8, target detection is performed based on the non-coherent accumulated data of the step S7.

2. The method for target detection based on non-coherent accumulation mode of coherent radar according to claim 1,

in step S1, the two-dimensional echo data is in the form of LFM signal, the number of pulses of each frame of LFM signal is N, each pulse samples M points, and the pulse repetition period is T_rPulse width of T_PEach sub-pulse has a frequency modulation slope of k and a sub-pulse bandwidth of B, and k is B/T_P(ii) a Setting the initial radial distance of the uniform motion target as R₀Radial velocity of motion v₀The target distance of the ith pulse is delayed by tau_i＝2(R₀+v₀iT_r) C, then the target baseband echo signal s_r(n, t) is:

wherein:

in the above formula, c represents the speed of light, and the length of t is M;

simultaneously, noise n (n, t) with a certain signal-to-noise ratio is superposed on the target baseband echo signal Noi (n, t) to obtain two-dimensional echo data SN_r(n,t)：

SN_r(n,t)＝Noi(n,t)+s_r(n,t) (3)。

3. The method for target detection based on the non-coherent accumulation mode of coherent radar according to claim 2,

in step S2, the matching function of pulse compression is:

the pulse compression is carried out in a frequency domain, and the two-dimensional echo data of the ith pulse is set as SN_ri(n, t), the pulse pressure thereofThe reduction algorithm is as follows:

sc_i(n,t)＝IFFT(FFT(SN_ri(n,t))·conj(FFT(s_ref(t)))) (5)

wherein, sc_i(N, t) is pulse pressure data of the ith pulse, i is 1,2, …, N.

4. The method for target detection based on non-coherent accumulation mode of coherent radar according to claim 1,

in step S3, the distance walk calibration is performed on the data after the pulse compression by using the inter-pulse cross correlation method.

5. The method for target detection based on the non-coherent accumulation mode of coherent radar according to claim 3,

in step S3, the pulse pressure data sc of the 1 st pulse is used₁(t) data sc obtained by compressing the pulse of the ith pulse with reference to_i(t) pulse pressure data s of the 1 st pulse_cm1(t) performing a cross-correlation process, a cross-correlation function R_1i(t) is:

find R_1i(t) position P where peak value is located_Xi，P_XiI.e. represents sc_i(t) relative to sc₁(t) the number of distance units, and pulse pressure data sc of the ith pulse_i(t) Cyclic Displacement P_XiPoint, obtain scx_i(t), implementing distance walk calibration; distance walk calibration is carried out on the data after pulse compression of each pulse, distance walk calibration of one frame of two-dimensional echo data is realized, and a new two-dimensional data plane S is formed_xc(n,t)。

6. The method for target detection based on non-coherent accumulation mode of coherent radar according to claim 5,

sc_i(n, t) two-dimensional arrayIn the cross-correlation function R of the last pulse_1N(t) position P where peak value is located_XNWhich represents the number of distance units P traveled by the distance generated between the Nth pulse and the 1 st pulse_XNThen, the corresponding walking distance dR is:

dR＝C·P_XN·t_s (7)

the estimated target speed v_eComprises the following steps:

7. the method for target detection based on non-coherent accumulation mode of coherent radar according to claim 5,

in the step S4, a new two-dimensional data plane S is processed_xc(n, t) performing coherent accumulation, and obtaining the column vector L corresponding to each distance unit_xcm(N) coherent accumulation, where M is 1,2, …, M, N is 1,2, …, N, i.e. the column vector L_xcm(n) performing FFT to obtain F_xcm(n):

F_xcm(n)＝FFT(L_xcm(n)) (9)

For each column vector L_xcm(n) all carry out coherent accumulation to form a new two-dimensional data plane F_xc1(n,t)。

Images