Complex environment distributed spaceborne radar moving target adaptive fusion detection method

Abstract

The invention belongs to the technical field of radar moving target detection, and particularly relates to a complex environment distributed spaceborne radar moving target adaptive fusion detection method, which comprises the following steps of S1, constructing amplitude and phase fusion detection quantity of each spaceborne radar; s2, estimating statistical distribution characteristics of amplitude and phase fusion detection quantity of each spaceborne radar; s3, optimally designing the detection false alarm probability of each spaceborne radar; s4, determining a detection threshold value of each spaceborne radar; s5, obtaining a binary detection result of each spaceborne radar; s6, performing optimal weighted fusion on binary detection results of all spaceborne radars in the distributed system to obtain a global detection quantity; and S7, judging a moving target through global detection. According to the method, the airspace freedom degree of the distributed multi-star radar can be better utilized, and the problem that the detection probability of the weak and small moving target under the strong clutter background is low in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of radar moving target detection, and further relates to a low signal-to-noise ratio moving target adaptive detection method for distributed spaceborne radar under a non-uniform background in the field of distributed spaceborne radar moving target detection. The present application can be used for moving target detection in a non-uniform scene based on a distributed spaceborne radar. BACKGROUND

[0002] The distributed spaceborne radar moving target detection technology can perform high-resolution imaging on an observation scene, and can still complete detection and positioning of ground moving targets under a non-uniform background, and is widely used in the fields of city traffic monitoring and military reconnaissance and early warning.

[0003] The distributed spaceborne radar is composed of a plurality of satellite formations, each satellite carries a radar system, a plurality of radar antenna beams are synthesized to form a large beam, and a high-resolution image is obtained through array signal processing on collected echo data. At present, the research on the ground moving target detection technology for the distributed spaceborne radar mainly includes three types of detection methods, namely, amplitude detection, phase detection and amplitude-phase joint detection. Among them, the amplitude detection is the amplitude information moving target detection on the image after adaptive clutter suppression; the phase detection is mainly interference processing on the data between multiple channels of the radar, and the interference phase is extracted, and the ground moving target detection is realized by using the interference phase information of the signal; the amplitude-phase joint detection is to jointly use the amplitude information after image adaptive clutter suppression and the multi-channel interference phase information of the image as the judgment criterion to realize the ground moving target detection, and the related technologies are as follows:

[0004] 1. The first paper file: the paper "Two-Step detector for RADARSAT-2's experimental GMTI mode" (IEEE Transactions on Geoscience & Remote Sensing) proposes a two-step detection method. The method first uses the phase center offset antenna DPCA (Displaced Phase Center Antenna) technology to suppress the clutter on the multi-channel radar image data, uses the signal amplitude information after the clutter suppression to construct a first-step detection statistic to preliminarily detect the radar image, and then uses the two-channel along-track interferometric ATI (Along track interferometric) phase to detect the radar image in the second step, and the final detection result is the logical AND result of the two-step detection. The deficiency of the method is that the second-step ATI phase detection statistic only uses two-channel echo data, the spatial degree of freedom is wasted for a multi-channel radar system, and the phase statistic is sensitive to the measurement error of the echo data, resulting in poor detection performance of low signal-to-noise ratio moving targets;

[0005] 2. It is found by searching that the Chinese invention with the publication number CN103454634A is a SAR moving target detection method based on Doppler analysis, which proposes a SAR moving target detection method based on Doppler analysis. The method takes a rectangular neighborhood of the scattering point in the SAR image, and solves the Doppler frequency center and the Doppler standard deviation by clutter locking for each Doppler spectrum in each neighborhood, then constructs a detection measure based on the Doppler frequency center and the Doppler standard deviation, and determines the detection threshold based on the constant false alarm rate method to realize the moving target detection. The method has the disadvantages that only single-channel SAR image data is used, the spatial freedom degree is low, the clutter suppression capability is limited, the moving target detection performance is sharply deteriorated under strong clutter background, and the operation complexity is high, so that real-time processing is not easy to realize.

[0006] In summary, in order to overcome the problems of low spatial freedom degree and unstable detection performance of the prior art, a distributed spaceborne radar low SNR moving target adaptive detection method under non-uniform background is proposed. SUMMARY

[0007] The purpose of the present application is to overcome the shortcomings of the prior art, and provide a distributed spaceborne radar low SNR moving target adaptive detection method under non-uniform background, which can better utilize the spatial freedom degree of distributed multi-satellite radar, and solve the problems of difficult detection of moving targets under non-uniform background and unstable detection performance of the prior art.

[0008] The technical scheme adopted by the present application is as follows:

[0009] A complex environment distributed spaceborne radar moving target adaptive fusion detection method, comprising the following steps:

[0010] S1: assuming that the distributed radar system is composed of M single-satellite radars, the amplitude and phase fusion detection quantity corresponding to each radar is constructed;

[0011] The implementation step S1 specifically comprises the following steps:

[0012] S11: assuming that the distributed radar system comprises M spaceborne radars, each spaceborne radar comprises N antenna channels, for the echo data of each antenna channel of each radar, synthetic aperture radar imaging technology is adopted, and MxN range Doppler images are obtained, which are respectively corresponding to each antenna channel of each radar;

[0013] S12: for the mth radar unit, the range Doppler images corresponding to the former N-1 and the latter N-1 antenna channels are respectively subjected to space-time adaptive processing, and two clutter suppression residual images are obtained, and are respectively denoted as y m,1 (k) and y m,2(k), where k represents the pixel number, and the value range is k = 1, 2, ..., K.

[0014] S13: Calculate the complex n-look interferometry processing result of the clutter suppression residual map of the m-th radar element according to the following formula.

[0015]

[0016] Among them, z m (k) represents the data corresponding to the k-th pixel in the m-th radar interferogram, where n represents the number of views, and y m,1 (k) and y m,2 (k) represents the value of the kth pixel in the clutter suppression residual map corresponding to the N-1 channels before and N-1 channels after the mth radar, respectively. * indicates the conjugate operation, and E[·] indicates the expectation operation.

[0017] S14: Calculate the amplitude feature of the k-th pixel of the m-th radar element according to the following formula.

[0018] t m (k)=|z m (k)|

[0019] Among them, t m (k) represents the amplitude feature of the k-th pixel of the m-th radar, where m represents the radar unit number, k represents the pixel number, and z m (k) represents the data corresponding to the k-th pixel in the m-th radar interferogram, and |·| represents the operation of taking the complex modulus.

[0020] S15: Calculate the phase feature of the k-th pixel of the m-th radar element according to the following formula.

[0021]

[0022] Among them, ò m (k) represents the phase feature of the m-th radar pixel k, b m (k) represents the feature vector of the k-th pixel of the m-th radar, expressed as: j represents the imaginary unit. The interference phase of the m-th radar pixel k is represented as: arg[·] represents taking the phase angle, z m (k) represents the data corresponding to the k-th pixel in the m-th radar interferogram, T represents the transpose operation, exp(·) represents the exponent with the natural number e as the base, and b0 represents the reference feature vector, expressed as b0=[1,1]. T .

[0023] S16: Calculate the amplitude and phase fusion detection quantity of the kth pixel point of the mth radar unit according to the following formula

[0024] γ m (k) = t m (k)(1-ò m (k))

[0025] wherein γ m (k) represents the amplitude and phase fusion detection quantity of the kth pixel point of the mth radar, t m (k) represents the amplitude feature quantity of the kth pixel point of the mth radar, ò m (k) represents the phase feature quantity of the kth pixel point of the mth radar.

[0026] S2: According to the single-star radar echo data, estimate the statistical distribution characteristics of the corresponding amplitude and phase fusion detection quantity, and denote the probability density function of the amplitude and phase fusion detection quantity of the mth radar as f m (γ; H1), and denote the probability density function of the amplitude and phase fusion detection quantity of the background detected by the mth radar as f m (γ; H0), γ represents the amplitude and phase fusion detection quantity, H1 and H0 represent the target hypothesis and the no-target hypothesis respectively;

[0027] S3: Design the false alarm probability of the amplitude and phase fusion detection of each single-star radar;

[0028] The step S3 specifically comprises the following steps:

[0029] S31: Initialize the detection false alarm probability of each radar unit, denoted as P f (1), …, P f (m), …, P f (M), P f (m) represents the false alarm probability of the amplitude and phase fusion detection quantity of the mth radar, m represents the radar unit serial number of the distributed radar, and the value range is m ∈ {1, 2, 3, …, M}.

[0030] S32: Calculate the amplitude and phase fusion detection threshold η m

[0031]

[0032] wherein P f (m) represents the false alarm probability of the amplitude and phase fusion detection quantity of the mth radar, η m represents the detection threshold of the mth radar, f m (γ; H0) represents the probability density function of the amplitude and phase fusion detection quantity of the background detected by the mth radar.

[0033] S33: Calculate the mth radar threshold η m Corresponding amplitude and phase fusion detection probability

[0034]

[0035] Wherein, P d (m) represents the detection probability of the mth radar amplitude and phase fusion detection quantity, η m represents the detection threshold of the mth radar, f m (γ;H1) represents the probability density function of the amplitude and phase fusion detection quantity corresponding to the moving target of the mth radar.

[0036] S34: Determine the moving target detection result of each radar with the detection threshold, wherein the pixel point whose amplitude and phase fusion detection quantity exceeds the threshold is marked as +1, and the pixel point whose amplitude and phase fusion detection quantity does not exceed the threshold is marked as-1. From the perspective of probability theory, there are 2 M M kinds of detection result combinations, wherein the radar serial number whose detection result is +1 is marked with set D1(l), and the radar serial number whose detection result is-1 is marked with set D0(l), l represents the detection result serial number of M radar units, and the value range is l∈{1,2,…,2 M}.

[0037] S35: Calculate the total false alarm probability of the distributed spaceborne radar according to the following formula

[0038]

[0039] Wherein, P f0 represents the total false alarm probability of the distributed spaceborne radar, M represents the number of radar units, D1(l) represents the radar serial number whose detection result is marked as +1 in the lth detection result combination, D0(l) represents the radar serial number whose detection result is-1 in the lth detection result combination, l represents the detection result combination serial number of M radar units, p and q represent the radar unit serial number of the distributed radar, P f (p) represents the false alarm probability of the amplitude and phase fusion detection quantity of the pth radar.

[0040] S36: Calculate the total detection probability of the distributed spaceborne radar according to the following formula

[0041]

[0042] Wherein, P d0Ptotal (M) represents the total detection probability of the distributed spaceborne radar, M represents the number of radar units, D1 (1) represents the radar serial number whose detection result is marked as +1 in the 1th detection result combination, D0 (1) represents the radar serial number whose detection result is -1 in the 1th detection result combination, 1 represents the serial number of the detection result combination of the M radar units, p and q represent the serial number of the radar unit of the distributed radar, P d (p) represents the detection probability of the 1th radar amplitude and phase fusion detection quantity.

[0043] S37: It is judged whether the total false alarm probability and the total detection probability of the distributed spaceborne radar meet the system index requirements, if yes, the detection false alarm probability of each unit radar at present is taken as the design result, otherwise the detection false alarm probability initialization value of each radar unit is updated, and steps S32 to S37 are repeated until the system index requirements are met.

[0044] S38: The false alarm probability of each single spaceborne radar in the design is recorded as m represents the serial number of the radar unit of the distributed radar, and the value range is m ∈ {1, 2, 3, …, M}.

[0045] S4: According to the statistical distribution characteristics in step S2 and the false alarm probability design result of the amplitude and phase fusion detection in step S3, the detection threshold of each single spaceborne radar is determined;

[0046] S5: For each single spaceborne radar, the pixel point whose amplitude and phase fusion detection quantity is greater than the detection threshold is marked as +1, and the pixel point which does not exceed the detection threshold is marked as -1, and the binary detection result of each single spaceborne radar is obtained;

[0047] S6: The binary detection result of each single spaceborne radar is optimally weighted and fused to obtain a global detection quantity;

[0048] The implementation of step S6 specifically includes the following steps:

[0049] S61: The binary detection result of the kth pixel point of the mth radar is recorded as u m (k), and the value range is u m (k) ∈ {-1, +1}, m represents the serial number of the radar unit of the distributed radar, and the value range is m ∈ {1, 2, 3, …, M}, k represents the serial number of the pixel point, and the value range is k ∈ {1, 2, 3, …, K}.

[0050] S62: The threshold corresponding to the amplitude and phase fusion detection probability

[0051]

[0052] wherein, denotes the detection probability of the mth radar amplitude and phase fusion detection quantity, denotes the detection threshold of the mth radar, f m denotes the amplitude and phase fusion detection quantity probability density function corresponding to the moving target of the mth radar, m represents the radar unit serial number of the distributed radar, and the value range is m∈{1, 2, 3, …, M}.

[0053] S63: The detection confidence of the mth radar unit is used to obtain the corresponding weighted fusion coefficient according to the following formula

[0054]

[0055] wherein a m denotes the weighted fusion coefficient obtained by the detection confidence of the mth radar unit, denotes the detection probability corresponding to the mth radar, denotes the false alarm probability of the mth radar amplitude and phase fusion detection, u m denotes the binary detection result of the mth radar, m represents the radar unit serial number of the distributed radar, and the value range is m∈{1, 2, 3, …, M}.

[0056] S64: The global detection quantity of the target is obtained by weighted fusion of the binary detection result corresponding to each radar unit, and is denoted as

[0057]

[0058] wherein β(k) denotes the global detection quantity corresponding to the kth pixel point, k represents the pixel point serial number, M represents the number of radar units of the distributed radar, m represents the radar unit serial number of the distributed radar, and the value range is m∈{1, 2, 3, …, M}, a m denotes the fusion coefficient of the mth radar unit, u m (k) denotes the binary detection result corresponding to the kth pixel point of the mth radar, and the value range is u m (k)∈{-1,+1}, η G denotes the global detection threshold.

[0059] S7: The pixel point whose global detection quantity exceeds the given global detection threshold is determined as a moving target, and is outputted.

[0060] The technical effects obtained by the present application are:

[0061] The complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application can improve the detection performance of the satellite-borne radar on the weak and small moving target under the strong clutter background, overcome the problem that the detection performance of the existing detection technology is easily affected by the clutter and the false alarm probability is high, and make the complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application not easily affected by the strong clutter interference in engineering practice, thereby improving the robustness of the satellite-borne radar on the moving target detection under the strong clutter background.

[0062] The complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application can improve the detection performance of the satellite-borne radar on the weak and small moving target under the strong clutter background, overcome the problem that the detection performance of the existing detection technology is easily affected by the clutter and the false alarm probability is high, and make the complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application not easily affected by the strong clutter interference in engineering practice, thereby improving the robustness of the satellite-borne radar on the moving target detection under the strong clutter background. Figure 2

[0063] The complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application can improve the detection performance of the satellite-borne radar on the weak and small moving target under the strong clutter background, overcome the problem that the detection performance of the existing detection technology is easily affected by the clutter and the false alarm probability is high, and make the complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application not easily affected by the strong clutter interference in engineering practice, thereby improving the robustness of the satellite-borne radar on the moving target detection under the strong clutter background. Figure 3 BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 The complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application can improve the detection performance of the satellite-borne radar on the weak and small moving target under the strong clutter background, overcome the problem that the detection performance of the existing detection technology is easily affected by the clutter and the false alarm probability is high, and make the complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application not easily affected by the strong clutter interference in engineering practice, thereby improving the robustness of the satellite-borne radar on the moving target detection under the strong clutter background.

[0065] Figure 2 The complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application can improve the detection performance of the satellite-borne radar on the weak and small moving target under the strong clutter background, overcome the problem that the detection performance of the existing detection technology is easily affected by the clutter and the false alarm probability is high, and make the complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application not easily affected by the strong clutter interference in engineering practice, thereby improving the robustness of the satellite-borne radar on the moving target detection under the strong clutter background.

[0066] Figure 3 The complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application can improve the detection performance of the satellite-borne radar on the weak and small moving target under the strong clutter background, overcome the problem that the detection performance of the existing detection technology is easily affected by the clutter and the false alarm probability is high, and make the complex environment distributed satellite-borne radar moving target adaptive fusion detection method of the application not easily affected by the strong clutter interference in engineering practice, thereby improving the robustness of the satellite-borne radar on the moving target detection under the strong clutter background. DETAILED DESCRIPTION

[0067] In order to make the purpose and advantages of the application more clear and obvious, the application is specifically described below in combination with the embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the application, and does not strictly limit the specific protection scope of the application.

[0068] Embodiment one:

[0069] As shown in Figure 1 , a complex environment distributed satellite-borne radar moving target adaptive fusion detection method comprises the following steps:

[0070] S1: assuming that the distributed radar system is composed of M single satellite radars, the amplitude and phase fusion detection quantity corresponding to each radar is constructed;

[0071] ​​S11: Assuming that the distributed radar system contains M spaceborne radars, each spaceborne radar includes N antenna channels, for each antenna channel of each radar, the echo data is processed by using synthetic aperture radar imaging technology, and M×N range Doppler images are obtained, which are respectively corresponding to each antenna channel of each radar;

[0072] S12: For the mth radar unit, the range Doppler images corresponding to the former N-1 and the latter N-1 antenna channels are respectively processed by using space-time adaptive processing, and two clutter suppression residual images are obtained, and are respectively denoted as y m,1 (k) and y m,2 (k), wherein k represents the pixel point sequence number, and the value range is k=1, 2, …, K.

[0073] The specific operation process of step S12 is as follows:

[0074] For the mth radar unit, the range Doppler images corresponding to the former N-1 and the latter N-1 antenna channels are respectively processed by using space-time adaptive processing, and two clutter suppression residual images are obtained, and are respectively denoted as y

[0075] x 1,m (k) = [x m (1, k), x m (2, k), …, x m (N-1, k)] T

[0076] Wherein, x 1,m (k) represents the data vector corresponding to the former N-1 antenna channels of the kth pixel point of the mth radar, m represents the radar unit sequence number of the distributed radar, and the value range is m∈{1, 2, 3, …, M}, k represents the pixel point sequence number, and x m (n0, k) represents the data of the kth pixel point in the range Doppler image corresponding to the n0th antenna channel of the mth radar, and T represents the transposition operation.

[0077] Correspondingly, the data vector is constructed by using the range Doppler image corresponding to the latter N-1 antenna channels of the mth radar unit as

[0078] x 2,m (k) = [x m (2, k), x m (3, k), …, x m (N, k)] T

[0079] Wherein, x 2,m (k) represents the data vector corresponding to the latter N-1 antenna channels of the kth pixel point of the mth radar.

[0080] For the data vectors x 1,m (k) and x 2,m(k), using the same space-time adaptive processing weight vector, expressed as

[0081]

[0082] wherein w m (k) represents the clutter suppression weight vector of the kth pixel point of the mth radar, R m (k) represents the clutter plus noise covariance matrix of the kth pixel point of the mth radar, a m (k) represents the target steering vector of the kth pixel point of the mth radar, and H represents the conjugate transpose.

[0083] The data vector x 1,m (k) and x 2,m (k) is subjected to clutter suppression processing, and the output result is

[0084]

[0085] wherein y m,1 (k) and y m,2 (k) respectively represent the clutter suppression residual image data of the kth pixel point of the mth radar using the first N-1 antennas and using the last N-1 antennas, m represents the radar unit sequence number of the distributed radar, and takes the value range of m∈{1, 2, 3, …, M}, k represents the pixel point sequence number, w m (k) represents the clutter suppression weight vector of the kth pixel point of the mth radar, and H represents the conjugate transpose, x 1,m (k) represents the data vector corresponding to the first N-1 antenna channels of the kth pixel point of the mth radar, x 2,m (k) represents the data vector corresponding to the last N-1 antenna channels of the kth pixel point of the mth radar.

[0086] S13: using the clutter suppression residual image data, calculating the complex n-view interference processing result of the kth pixel point of the mth radar unit according to the following formula

[0087]

[0088] wherein z m (k) represents the data corresponding to the kth pixel point in the interference image of the mth radar, n represents the view number, y m,1 (k) and y m,2 (k) respectively represent the kth pixel point value in the clutter suppression residual image corresponding to the first N-1 and last N-1 antenna channels of the mth radar, * represents the conjugate operation, and E[·] represents the expectation operation.

[0089] S14: calculating the amplitude feature quantity of the kth pixel point of the mth radar unit according to the following formula

[0090] t m (k) = |z m (k)|

[0091] wherein t m (k) represents the amplitude feature quantity of the kth pixel point of the mth radar, m represents the radar unit serial number, and k represents the pixel point serial number. z m (k) represents the data corresponding to the kth pixel point in the mth radar interferogram, and |·| represents the complex modulus value operation.

[0092] S15: the phase feature quantity of the kth pixel point of the mth radar unit is calculated according to the following formula

[0093]

[0094] wherein t m (k) represents the phase feature quantity of the kth pixel point of the mth radar, b m (k) represents the feature vector of the kth pixel point of the mth radar, and the expression is j represents the imaginary unit, represents the interference phase of the kth pixel point of the mth radar, and is expressed as arg[·] represents the phase angle, z m (k) represents the data corresponding to the kth pixel point in the mth radar interferogram, T represents the transpose operation, exp(·) represents the exponential with the natural number e as the base, b0 represents the reference feature vector, and is expressed as b0 = [1, 1] T .

[0095] S16: the amplitude and phase fusion detection quantity of the kth pixel point of the mth radar unit is calculated according to the following formula

[0096] γ m (k) = t m (k)(1-ò m (k))

[0097] wherein γ m (k) represents the amplitude and phase fusion detection quantity of the kth pixel point of the mth radar, t m (k) represents the amplitude feature quantity of the kth pixel point of the mth radar, and ò m (k) represents the phase feature quantity of the kth pixel point of the mth radar.

[0098] S2: according to the single-star radar echo data, the statistical distribution characteristics of the corresponding amplitude and phase fusion detection quantity are estimated, and the probability density function of the moving target amplitude and phase fusion detection quantity of the mth radar is denoted as f m (γ; H1), and the probability density function of the amplitude and phase fusion detection quantity of the mth radar detection background is denoted as fm (γ; H0), γ represents an amplitude and phase fusion detection quantity, H1 and H0 respectively represent a target hypothesis and a no-target hypothesis;

[0099] S3: design a false alarm probability of amplitude and phase fusion detection of each single-star radar;

[0100] The step S3 specifically comprises the following steps:

[0101] S31: initialize a detection false alarm probability of each radar unit, denoted as P f (1), …, P f (m), …, P f (M), P f (m) represents a false alarm probability of an amplitude and phase fusion detection quantity of the mth radar, m represents a radar unit serial number of the distributed radar, and the value range is m ∈ {1, 2, 3, …, M}.

[0102] S32: calculate an amplitude and phase fusion detection threshold η m

[0103]

[0104] wherein P f (m) represents a false alarm probability of an amplitude and phase fusion detection quantity of the mth radar, η m represents a detection threshold of the mth radar, f m (γ; H0) represents a probability density function of an amplitude and phase fusion detection quantity of the mth radar for detecting background, and m represents a radar unit serial number of the distributed radar, and the value range is m ∈ {1, 2, 3, …, M}.

[0105] S33: calculate an amplitude and phase fusion detection probability of the mth radar corresponding to the threshold η m

[0106]

[0107] wherein P d (m) represents a detection probability of an amplitude and phase fusion detection quantity of the mth radar, η m represents a detection threshold of the mth radar, f m (γ; H1) represents a probability density function of an amplitude and phase fusion detection quantity of the mth radar corresponding to a moving target, and m represents a radar unit serial number of the distributed radar, and the value range is m ∈ {1, 2, 3, …, M}.

[0108] ​S34: Determine the moving target detection result of each radar with a detection threshold, wherein the pixel point with the amplitude and phase fusion detection quantity exceeding the threshold is marked as +1, and the pixel point with the amplitude and phase fusion detection quantity not exceeding the threshold is marked as -1. From the perspective of probability theory, there are 2M detection result combinations, wherein the radar serial number with the detection result of +1 is marked with set D1(l), and the radar serial number with the detection result of -1 is marked with set D0(l), and l represents the detection result serial number of the M radar units, and the value range is l∈{1, 2, …, 2M}. M M

[0109] S35: Calculate the total false alarm probability of the distributed spaceborne radar according to the following formula

[0110]

[0111] wherein P f0 represents the total false alarm probability of the distributed spaceborne radar, M represents the number of radar units, D1(l) represents the radar serial number with the detection result marked as +1 in the lth detection result combination, D0(l) represents the radar serial number with the detection result of -1 in the lth detection result combination, l represents the detection result combination serial number of the M radar units, p and q represent the radar unit serial number of the distributed radar, P f (p) represents the false alarm probability of the pth radar amplitude and phase fusion detection quantity.

[0112] S36: Calculate the total detection probability of the distributed spaceborne radar according to the following formula

[0113]

[0114] wherein P d0 represents the total detection probability of the distributed spaceborne radar, M represents the number of radar units, D1(l) represents the radar serial number with the detection result marked as +1 in the lth detection result combination, D0(l) represents the radar serial number with the detection result of -1 in the lth detection result combination, l represents the detection result combination serial number of the M radar units, p and q represent the radar unit serial number of the distributed radar, P d (p) represents the detection probability of the pth radar amplitude and phase fusion detection quantity.

[0115] S37: Determine whether the total false alarm probability and the total detection probability of the distributed spaceborne radar meet the system index requirements, if yes, take the detection false alarm probability of each unit radar at present as the design result, otherwise update the detection false alarm probability initialization value of each radar unit, repeat steps S32 to S37 until the system index requirements are met.

[0116] S38: Record the false alarm probability of each single spaceborne radar in the design as ​​m represents the radar unit sequence number of the distributed radar, and m∈{1, 2, 3, …, M}.

[0117] S4: determining each single-star radar detection threshold according to the statistical distribution characteristics in step S2 and the false alarm probability design result of the amplitude and phase fusion detection in step S3;

[0118] Specifically, the amplitude and phase fusion detection threshold η of the mth radar is calculated according to the following formula m

[0119]

[0120] wherein, indicates the amplitude and phase fusion detection false alarm probability of the mth radar designed in step S3, indicates the detection threshold of the mth radar, f m (γ; H0) indicates the amplitude and phase fusion detection quantity probability density function of the mth radar detecting the background, γ indicates the amplitude and phase fusion detection quantity, and m represents the radar unit sequence number of the distributed radar, and m∈{1, 2, 3, …, M}.

[0121] S5: for each single-star radar, marking the pixel point with an amplitude and phase fusion detection quantity greater than the detection threshold as +1, and marking the pixel point not exceeding the detection threshold as -1, to obtain the binary detection result of each single-star radar;

[0122] S6: performing optimal weighted fusion on the binary detection result of each single-star radar to obtain a global detection quantity;

[0123] The step S6 specifically includes the following steps:

[0124] S61: recording the binary detection result of the kth pixel point of the mth radar as u m (k), and u m (k)∈{-1, +1}, m represents the radar unit sequence number of the distributed radar, and m∈{1, 2, 3, …, M}, and k represents the pixel point sequence number, and k∈{1, 2, 3, …, K}.

[0125] S62: calculating the threshold corresponding to the amplitude and phase fusion detection probability of the mth radar

[0126]

[0127] wherein, indicates the detection probability of the amplitude and phase fusion detection quantity of the mth radar, indicates the detection threshold of the mth radar, f m(γ;H1) represents the amplitude and phase fusion detection quantity probability density function corresponding to the mth radar of the moving target, m represents the radar unit sequence number of the distributed radar, and the value range is m∈{1, 2, 3, …, M}.

[0128] S63: the detection credibility of the mth radar unit is used to obtain the corresponding weighted fusion coefficient

[0129]

[0130] Wherein, a m The weighted fusion coefficient obtained by the detection credibility of the mth radar unit, The detection probability corresponding to the mth radar, The false alarm probability of the amplitude and phase fusion detection of the mth radar, u m The binary detection result of the mth radar, m represents the radar unit sequence number of the distributed radar, and the value range is m∈{1, 2, 3, …, M}.

[0131] S64: the global detection quantity of the target is obtained by weighting and fusing the binary detection result corresponding to each radar unit, which is represented as

[0132]

[0133] Wherein, β(k) represents the global detection quantity corresponding to the kth pixel point, k represents the pixel point sequence number, M represents the number of radar units of the distributed radar, m represents the radar unit sequence number of the distributed radar, and the value range is m∈{1, 2, 3, …, M}, a m The fusion coefficient of the mth radar unit, u m (k) represents the binary detection result corresponding to the kth pixel point of the mth radar, and the value range is u m (k)∈{-1,+1}, η G The global detection threshold.

[0134] S7: the pixel point whose global detection quantity β(k) exceeds the given global detection threshold η G is determined as a moving target, and is output, wherein k represents the pixel point sequence number.

[0135] In summary, since the present application uses the distributed spaceborne radar data to design the single-satellite amplitude and phase fusion detection and the multi-satellite collaborative optimal fusion detection criterion, the detection performance of the spaceborne radar on the weak and small moving target under the strong clutter background is improved, the problem that the detection performance of the existing detection technology is easily affected by the clutter and the false alarm probability is high is overcome, so that the present application is not easily affected by the strong clutter interference in engineering practice, and the robustness of the spaceborne radar on the moving target detection under the strong clutter background is improved.

[0136] Embodiment one:

[0137] As Figure 1 shown, a complex environment distributed satellite-borne radar moving target adaptive fusion detection method given for embodiment one needs to be put in simulation environment for testing, as follows:

[0138] 1. Simulation conditions:

[0139] The simulation experiment environment of the application is: MATLAB R2010a, Intel(R) Core(TM) 2 Duo CPU 3.4GHz, WindowXP Professional Edition.

[0140] 2. Simulation content and result analysis:

[0141] The effectiveness of the algorithm is verified by simulation data. In the simulation process, the distributed radar is composed of 4 single satellite radars, each radar has 8 antennas, the radar wavelength is 0.25 meters, and the channel spacing is 0.125 meters. Each satellite observes the same moving target from different angles, and due to the difference in angle, the radial velocity and clutter-to-noise ratio of the moving target relative to each satellite are different. The non-uniformity of the clutter background adopts the measurement method in the literature (C.H. Gierull, I. Sikaneta, and D. Cerutti-Maori, “Two-step detector for RADARSAT-2’s experimental GMTI mode,” IEEE Transactions on Geoscience and Remote Sensing, vol. 51, no. 1, pp. 436-454, 2013.), and the clutter non-uniformity parameters of the 4 satellite radars are 3, 5, 11, and 13, respectively. Note that the smaller the non-uniformity parameter, the more uneven the clutter background, that is, the greater the fluctuation of the scattering coefficient of the scene, and the range of change of the clutter-to-noise ratio is 15dB to 60dB. The radial velocities of the moving target relative to the 4 satellite radars are 57m / s, 57m / s, 16m / s, and 63m / s, respectively. After clutter suppression processing, the output signal-to-clutter noise ratios of the moving target are 14.77dB, 10dB, 13dB, and 11.76dB, respectively.

[0142] Figure 2 The local detection results of satellite 1 are given. It can be seen that compared with only using the amplitude detection method, the amplitude and phase fusion detection method proposed in the application can significantly reduce the number of false alarms and improve the single satellite detection accuracy.

[0143] Figure 3 The detection performance curves of the 4 satellites after optimal weighted fusion are given. Figure 3The performance curves of the single star radar using the amplitude characteristic quantity proposed in the application for detection are respectively represented by "t1", "t2", "t3" and "t4", and "β" represents the distributed satellite adaptive detection global detection quantity proposed in the application. It can be seen that in the low false alarm probability region, the method can significantly improve the detection probability of the moving target, and further improve the detectability of the moving target.

[0144] The above only describes the preferred embodiments of the present application. It should be noted that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which should also be considered as the protection scope of the present application. The structures, devices and operation methods not specifically described and explained in the present application are implemented according to the conventional means in the art, unless otherwise specified and limited.

Claims

1. An adaptive fusion detection method for moving targets using distributed spaceborne radar in complex environments, characterized in that: Includes the following steps: S1: Assume that the distributed radar system consists of M single-satellite radars, and construct the amplitude and phase fusion detection quantities for each radar. S2: Based on the single-satellite radar echo data, estimate the statistical distribution characteristics of the corresponding amplitude and phase fusion detection quantities. Let f be the probability density function of the moving target amplitude and phase fusion detection quantity of the m-th radar. m (γ; H1), let f be the probability density function of the amplitude and phase fusion detection quantity of the m-th radar background detection. m (γ; H0), where γ represents the amplitude and phase fusion detection quantity, and H1 and H0 represent the target hypothesis and the non-target hypothesis, respectively; S3: Design the false alarm probability for amplitude and phase fusion detection of each single-satellite radar; S4: Based on the statistical distribution characteristics in step S2 and the false alarm probability design results of amplitude and phase fusion detection in step S3, determine the detection threshold for each single-satellite radar. S5: For each single-satellite radar, pixels with amplitude and phase fusion detection values ​​greater than the detection threshold are marked as +1, and pixels with values ​​less than the detection threshold are marked as -1, thus obtaining the binary detection result for each single-satellite radar. S6: Perform optimal weighted fusion on the binary detection results of each single-satellite radar to obtain the global detection quantity; S7: Identify pixels whose global detection count exceeds the given global detection threshold as moving targets and output the result.

2. The adaptive fusion detection method for moving targets in a distributed spaceborne radar system in complex environments according to claim 1, characterized in that: Step S1 includes the following steps: S11: Assume that the distributed radar system contains M spaceborne radars, each of which includes N antenna channels. For the echo data of each antenna channel of each radar, synthetic aperture radar imaging technology is used to obtain a total of M×N range Doppler images, which correspond one-to-one with each antenna channel of each radar. S12: For the m-th radar element, perform space-time adaptive processing using the range Doppler images corresponding to the first N-1 and last N-1 antenna channels respectively to obtain two clutter suppression residual images, which are denoted as y. m,1 (k) and y m,2 (k), where k represents the pixel number, and the value range is k = 1, 2, ..., K. S13: Based on the clutter suppression residual map data, calculate the complex n-look interferometry result of the k-th pixel of the m-th radar unit according to the following formula. Among them, z m (k) represents the data corresponding to the k-th pixel in the m-th radar interferogram, where n represents the number of views, and y m,1 (k) and y m,2 (k) represents the value of the kth pixel in the clutter suppression residual map corresponding to the N-1 channels before and N-1 channels after the mth radar, respectively. * indicates the conjugate operation, and E[·] indicates the expectation operation. S14: Calculate the amplitude feature of the k-th pixel of the m-th radar element according to the following formula. t m (k)=|z m (k)| Among them, t m (k) represents the amplitude feature of the k-th pixel of the m-th radar, where m represents the radar unit number, k represents the pixel number, and z m (k) represents the data corresponding to the k-th pixel in the m-th radar interferogram, and |·| represents the operation of taking the complex modulus. S15: Calculate the phase feature of the k-th pixel of the m-th radar element according to the following formula. Among them, ò m (k) represents the phase feature of the m-th radar pixel k, b m (k) represents the feature vector of the k-th pixel of the m-th radar, expressed as: j represents the imaginary unit. The interference phase of the m-th radar pixel k is represented as: arg[·] represents taking the phase angle, z m (k) represents the data corresponding to the k-th pixel in the m-th radar interferogram, T represents the transpose operation, exp(·) represents the exponent with the natural number e as the base, and b0 represents the reference feature vector, expressed as b0=[1,1]. T . S16: Calculate the amplitude and phase fusion detection quantity of the m-th radar element according to the following formula. c m (k)=t m (k)(1-ò m (k)) Where, γ m (k) represents the amplitude and phase fusion detection value of the m-th radar pixel k, t m (k) represents the amplitude feature of the m-th radar pixel k, ò m (k) represents the phase feature of the m-th radar and the k-th pixel.

3. The adaptive fusion detection method for moving targets in a distributed spaceborne radar system in complex environments according to claim 2, characterized in that: Step S3 includes the following steps: S31: Initialize the false alarm probability of each radar unit, denoted as P. f (1), ..., P f (m), ..., P f (M), P f (m) represents the false alarm probability of the m-th radar amplitude and phase fusion detection quantity, where m represents the radar element number of the distributed radar, and the value range is m∈{1,2,3,…,M}. S32: Calculate the radar amplitude and phase fusion detection threshold η for the m-th radar according to the following formula. m Among them, P f (m) represents the false alarm probability of the m-th radar amplitude and phase fusion detection quantity, η m f represents the detection threshold of the m-th radar. m (γ;H0) represents the probability density function of the amplitude and phase fusion detection quantity of the m-th radar detection background. S33: Calculate the threshold η of the m-th radar according to the following formula. m Corresponding amplitude and phase fusion detection probability Among them, P d (m) represents the detection probability of the m-th radar amplitude and phase fusion detection quantity, η m f represents the detection threshold of the m-th radar. m (γ;H1) represents the probability density function of the amplitude and phase fusion detection quantity corresponding to the moving target of the m-th radar. S34: Determine the moving target detection result of each radar based on the detection threshold. Pixels with amplitude and phase fusion detection values ​​exceeding the threshold are marked as +1, while pixels with amplitude and phase fusion detection values ​​not exceeding the threshold are marked as -1. From a probabilistic perspective, there are a total of 2... M There are several combinations of detection results, where the radar serial numbers with a detection result of +1 are labeled by set D1(l), and the radar serial numbers with a detection result of -1 are labeled by set D0(l), where l represents the detection result ordinal number of the M radar units, and the value range is l∈{1,2,…,2}. M } S35: Calculate the total false alarm probability of the distributed spaceborne radar according to the following formula. Among them, P f0 Let P represent the total false alarm probability of the distributed spaceborne radar, M represent the number of radar elements, D1(l) represent the radar number marked as +1 in the l-th detection result combination, D0(l) represent the radar number marked as -1 in the l-th detection result combination, l represent the detection result combination ordinal number of M radar elements, p and q represent the radar element ordinal numbers of the distributed radar, and P f (p) represents the false alarm probability of the p-th radar amplitude and phase fusion detection quantity. S36: Calculate the total detection probability of the distributed spaceborne radar according to the following formula. Among them, P d0 Let P represent the total detection probability of the distributed spaceborne radar, M represent the number of radar elements, D1(l) represent the radar number marked with +1 in the l-th detection result combination, D0(l) represent the radar number marked with -1 in the l-th detection result combination, l represent the detection result combination ordinal number of M radar elements, p and q represent the radar element ordinal numbers of the distributed radar, and P d (p) represents the detection probability of the p-th radar amplitude and phase fusion detection quantity. S37: Determine whether the total false alarm probability and total detection probability of the distributed spaceborne radar meet the system index requirements. If they meet, use the current detection false alarm probability of each unit radar as the design result. Otherwise, update the initial value of the detection false alarm probability of each radar unit and repeat steps S32 to S37 until the system index requirements are met. S38: Record the false alarm probability of each individual radar in the design. m represents the radar element ordinal number of the distributed radar, and its value range is m∈{1,2,3,…,M}.

4. The adaptive fusion detection method for moving targets in a distributed spaceborne radar system in complex environments according to claim 2, characterized in that: Step S6 includes the following steps: S61: Record the binary detection result of the m-th radar pixel k as u m (k), with a value range of u m (k)∈{-1,+1}, m represents the radar element ordinal number of the distributed radar, with a value range of m∈{1,2,3,…,M}, and k represents the pixel ordinal number, with a value range of k∈{1,2,3,…,K}. S62: Calculate the threshold value of the m-th radar according to the following formula. Corresponding amplitude and phase fusion detection probability in, This represents the detection probability of the m-th radar amplitude and phase fusion detection quantity. f represents the detection threshold of the m-th radar. m (γ;H1) represents the probability density function of the amplitude and phase fusion detection of the moving target corresponding to the m-th radar, where m represents the radar cell number of the distributed radar, and the value range is m∈{1,2,3,…,M}. S63: The weighted fusion coefficient corresponding to the detection confidence of the m-th radar element is obtained according to the following formula. Among them, a m This represents the weighted fusion coefficient obtained from the detection confidence of the m-th radar unit. This represents the detection probability corresponding to the m-th radar. U represents the false alarm probability of the m-th radar amplitude and phase fusion detection. m This represents the binary detection result of the m-th radar, where m represents the radar element number of the distributed radar, and the value range is m∈{1,2,3,…,M}. S64: The global detection quantity of the target is obtained by weighted fusion of the binary detection results corresponding to each radar unit, denoted as: Where β(k) represents the global detection quantity corresponding to the k-th pixel, k represents the pixel ordinal number, M represents the number of radar elements in the distributed radar, and m represents the radar element ordinal number of the distributed radar, with a value range of m∈{1,2,3,…,M}, a m U represents the fusion coefficient of the m-th radar element. m (k) represents the binary detection result corresponding to the k-th pixel of the m-th radar, with a value range of u. m (k)∈{-1,+1},η G This represents the global detection threshold.

Images