A variable resolution radar imaging method

Abstract

The application discloses a variable resolution radar imaging method, which has the following characteristics: step 1, a variable resolution synthetic aperture radar imaging method is proposed, an original scene is divided into multiple sub-scenes, and the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained; step 2, the minimum pulse repetition period required for sampling is obtained according to the azimuth span required to be irradiated at each aperture, and echo data storage is reduced; step 3, scene information is defined according to the scene scattering amplitude distribution; step 4, mutual information between the scene and a synthetic aperture radar image is defined based on information theory; step 5, mutual information between the original scene and the obtained synthetic aperture radar is calculated based on the maximum mutual information principle, and a mathematical model of an optimization problem is constructed; and step 6, an equivalent model of the above problem is constructed, and under certain limitation conditions, the synthetic aperture length of each sub-scene is optimized to maximize the mutual information, so that the scene information is maximally represented.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of radar imaging, in particular to a variable resolution radar imaging method. BACKGROUND

[0002] As one of the indispensable observation means of space reconnaissance system, synthetic aperture radar (SAR) can well make up for the deficiency that optical cameras cannot conduct reconnaissance all day and all weather, and has certain penetration ability and is more likely to identify camouflage and thus find underground military facilities. Since the United States launched the first SAR satellite SEASAT in 1978, many countries have established their own synthetic aperture radar development programs, and various new SAR systems have emerged, which have played an irreplaceable role in military reconnaissance, topographic mapping, environmental monitoring, disaster prevention and other aspects.

[0003] Traditional strip SAR and spotlight SAR have a trade-off problem between high resolution and wide swath, in order to obtain high resolution in the azimuth direction, a large Doppler bandwidth is needed, and in order to avoid azimuth aliasing, a high PRF is needed for sampling; however, a high PRF will limit the range swath width, and if a large range swath width is selected, range ambiguity will be caused. In order to balance this contradiction, high resolution wide swath (HRWS) SAR mode, digital beam forming (DBF) SAR mode, enhanced resolution strip mode (ERSM) SAR mode and the like have been proposed, among which the ERSM SAR mode combines the programmable metamaterial antenna which is popular in recent years to provide a dynamically adjustable directional diagram. However, the current SAR modes do not combine the dynamically adjustable directional diagram and optimize the beam resources in the form of variable resolution, and do not quantitatively measure the pros and cons of imaging quality from the perspective of information theory. In summary, how to combine the characteristics and advantages of metamaterial antennas, and balance the contradiction between high resolution and wide swath under the condition of limited resources, and maximize the extraction of information from the scene, is the problem we are concerned about and solve. SUMMARY

[0004] The present application is to solve the problem of how to maximize the extraction of scene mutual information by radar system under the condition of limited transmission resources, and aims to provide a variable resolution radar imaging method.

[0005] The present application provides a variable resolution radar imaging method, which has the following characteristics and comprises the following steps:

[0006] Step 1, a variable resolution synthetic aperture radar (SAR) imaging method is proposed, the original scene is divided into a plurality of sub-scenes, and based on the relationship between the azimuth direction resolution and the synthetic aperture length in the traditional SAR imaging method, the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained;

[0007] Step 2, selecting the minimum pulse repetition period (PRF) required for current aperture sampling according to the azimuth span required to illuminate at each aperture, thereby reducing echo data storage and increasing slant range to swath width;

[0008] Step 3, obtaining scene information according to scene scattering amplitude distribution;

[0009] Step 4, obtaining mutual information between the scene and the obtained synthetic aperture radar image based on information theory;

[0010] Step 5, constructing a mathematical model of a variable resolution synthetic aperture radar imaging optimization problem according to the scene information and the mutual information between the scene and the obtained synthetic aperture radar image based on the maximum mutual information principle;

[0011] Step 6, constructing an equivalent mathematical model of the optimization problem model, and under the constraints of minimum signal-to-noise ratio and priority positive side view, maximizing the mutual information between the original scene and the synthetic aperture radar by optimizing the synthetic aperture length of each sub-scene, thereby obtaining more accurate imaging results.

[0012] In the variable resolution radar imaging method provided by the application, the step 1 can further have the following characteristics: wherein, according to the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method, the synthetic aperture length under a certain scene is obtained, and the formula is:

[0013]

[0014] δ a , wherein, represents the azimuth resolution, L represents the synthetic aperture length, R0 represents the reference slant range, and λ represents the wavelength. If the imaging scene is divided into M sub-scenes, different sub-scenes are allocated different azimuth resolutions, different synthetic aperture lengths and synthetic aperture positions are generated, and then each sub-aperture space is superimposed to generate a total synthetic aperture, and the formula is:

[0015]

[0016] X m , wherein, represents the azimuth center coordinate of the mth sub-scene, L m represents the synthetic aperture length of the mth sub-scene.

[0017] In the variable resolution radar imaging method provided by the application, the step 2 can further have the following characteristics: wherein, according to the azimuth span required to illuminate at each aperture, the minimum required pulse repetition period is calculated, and the formula is:

[0018]

[0019] n represents the number of azimuth sub-scene illuminated by the current aperture, V represents the velocity of the radar platform, r x represents the azimuth side length of a sub-scene, PRF n represents the pulse repetition period selected by the current aperture sampling.

[0020] In the variable resolution radar imaging method provided by the application, the method can further have the following characteristics: in step 3, it is assumed that the original scene obeys a generalized stationary Gaussian distribution, and the scene is grid divided, the scene is divided into N a spatial resolution units in the azimuth direction and N r spatial resolution units in the slant range direction. Based on information theory, the scene information in the wave number domain is defined, and the formula is as follows:

[0021]

[0022] S a (u,v) represents the covariance spectrum of the scene scattering amplitude, e is a natural constant, A represents the scattering amplitude distribution of the scene in the wave number domain, H e [A] represents the information of the scene in the wave number domain.

[0023] In the variable resolution radar imaging method provided by the application, the method can further have the following characteristics: in step 3, it is assumed that the original scene obeys a generalized stationary Gaussian distribution, and the scene is grid divided, the scene is divided into N a spatial resolution units in the azimuth direction and N r spatial resolution units in the slant range direction. Based on information theory, the scene information in the wave number domain is defined, and the formula is as follows:

[0024]

[0025] S a (u,v) represents the covariance spectrum of the scene scattering amplitude, e is a natural constant, A represents the scattering amplitude distribution of the scene in the wave number domain, H e [A] represents the information of the scene in the wave number domain.

[0026] In the variable resolution radar imaging method provided by the application, the method can further have the following characteristics: in step 4, based on information theory, the mutual information between the scene and the obtained synthetic aperture radar image is defined, and the formula is as follows:

[0027]

[0028] S n (u,v) represents the covariance spectrum of the system noise, H(u,v) represents the imaging system function, and B represents the obtained synthetic aperture radar image.

[0029] In the variable resolution radar imaging method provided by the application, the step 5 can further have the following characteristics: the step 5 includes the following steps:

[0030] Step 5-1, the imaging scene is divided into M sub-scenes, the scattering amplitude A of each sub-scene is modeled according to prior knowledge, and the covariance spectrum S a (u,v) is used to represent the imaging system function of each sub-scene;

[0031] Step 5-2, the mutual information between the synthetic aperture radar image and the original scene is calculated according to the covariance spectrum of the scattering amplitude of all sub-scenes, and the formula is as follows:

[0032]

[0033] S m (u,v) represents the covariance spectrum of the mth sub-scene, H m (u,v) represents the imaging system function of the mth sub-scene, the mth sub-scene is divided into space resolution units in the azimuth direction, and space resolution units in the slant range direction;

[0034] Step 5-3, the two-dimensional matrix imaging system function H m (u,v) of each scene can be divided into one-dimensional matrix imaging system functions in the slant range direction and the azimuth direction, and the formula is as follows:

[0035]

[0036] H r (v) represents the one-dimensional matrix imaging system function in the slant range direction of the mth sub-scene. According to the relationship between the point spread function and the resolution, the formula is as follows:

[0037]

[0038] B r represents the bandwidth, and c represents the speed of light;

[0039] Step 5-4, the two-dimensional covariance spectrum of the scattering amplitude obeying the Gaussian distribution is constructed, and the formula is as follows:

[0040]

[0041] l x represents the correlation length of the scattering amplitude in the azimuth direction, l y represents the correlation length of the scattering amplitude in the slant range direction, represents the mean square error of the scattering amplitude;

[0042] Step 5-5, construct the objective function of the optimization problem, formula is:

[0043]

[0044] The correlation length of the scattering amplitude in the azimuth direction under the mth sub-scene is represented by L The correlation length of the scattering amplitude in the slant range direction under the mth sub-scene is represented by L

[0045] In the variable resolution radar imaging method provided by the application, the step 6 can further have the following characteristics: the specific implementation steps in step 6 include:

[0046] Step 6-1, calculate the growth slope of the mutual information of each sub-scene with the growth of the spectrum width according to formula (6), and construct the slope vector K

[0047] K=[k1,k2,…,k M ] (12)

[0048] Step 6-2, construct the illumination matrix P M×T , and define the synthetic aperture vector as

[0049] L=[l1,l2,…,l M ] Τ

[0050]

[0051] T represents the total number of azimuth samples;

[0052] Step 6-3, construct the optimization objective function equivalent to formula (11), so that the problem can be solved by using the CVX toolbox of matlab, formula is

[0053] arg max KL (14)

[0054] Step 6-4, set the constraint of the minimum signal-to-noise ratio SNR min According to the relationship between the signal-to-noise ratio and the antenna gain in the synthetic aperture radar image, formula is:

[0055]

[0056] N A represents the number of azimuth samples, G ave represents the average antenna gain of N A sampling, n0 represents the power spectral density of noise, P t represents the peak transmit power, T r represents the pulse width;

[0057] Step 6-5, set the aperture position interval of each sub-scene orientation direction, and the aperture number corresponding to the mth sub-scene is The formula for constructing the priority forward-looking restriction condition is:

[0058]

[0059] Step 6-6, based on formula (14), formula (15), and formula (16), an equivalent mathematical model of the optimization problem is constructed, and the formula is:

[0060] argmax KL

[0061]

[0062] Step 6-7, the optimization problem of formula (17) is solved by using the CVX toolbox of MATLAB, and the optimal illumination matrix P is obtained M×T And the directional diagram distribution.

[0063] In the variable resolution radar imaging method provided by the application, the variable resolution radar imaging system can also have the following features:

[0064] The variable resolution synthetic aperture radar imaging module proposes a variable resolution synthetic aperture radar imaging method, divides the original scene into multiple sub-scenes, and obtains the synthetic aperture length and synthetic aperture position of each sub-scene based on the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method.

[0065] The variable pulse repetition period module obtains the minimum pulse repetition period required for each aperture sampling according to the azimuth span required for illumination at each aperture, thereby reducing echo data storage and increasing slant range strip width.

[0066] The scene information quantity module defines the scene information quantity according to the scene scattering amplitude distribution.

[0067] The mutual information module defines the mutual information between the scene and the synthetic aperture radar image based on information theory.

[0068] The model construction module calculates the mutual information between the original scene and the obtained synthetic aperture radar based on the maximum mutual information principle, according to the scene information quantity and the mutual information between the scene and the obtained synthetic aperture radar image, and constructs a mathematical model of the variable resolution synthetic aperture radar imaging optimization problem.

[0069] The optimization model module constructs an equivalent mathematical model of the optimization problem model, and under the restriction conditions of the minimum signal-to-noise ratio and the priority forward-looking, the mutual information between the original scene and the synthetic aperture radar is maximized by optimizing the synthetic aperture length of each sub-scene, so as to maximize the acquisition and characterization of scene information.

[0070] Effects of the Invention

[0071] According to the variable resolution radar imaging method and system, the original scene is divided into multiple sub-scenes, and based on the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method, the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained, then the minimum pulse repetition period required for the current aperture sampling is obtained according to the azimuth span required for irradiation at each aperture, the scene information quantity is obtained according to the scene scattering amplitude distribution, the mutual information between the original scene and the obtained synthetic aperture radar image is obtained based on information theory, and based on the maximum mutual information principle, a mathematical model of the variable resolution synthetic aperture radar imaging optimization problem is constructed, and under the restriction conditions of the minimum signal-to-noise ratio and the priority of the forward-looking, the synthetic aperture length of each sub-scene is optimized to maximize the mutual information between the original scene and the synthetic aperture radar, so as to maximize the acquisition and characterization of scene information. The above process can balance the contradiction between high resolution and large range width, and can take into account the high resolution of part of the scene while having a larger range width.

[0072] Further, by adopting the variable PRF method, the storage space of the original echo data is reduced without causing azimuth ambiguity, so that the imaging process is more efficient. BRIEF DESCRIPTION OF DRAWINGS

[0073] Figure 1 is a flowchart of the variable resolution radar imaging method in embodiment 1 of the present application;

[0074] Figure 2 is a schematic diagram of an application scene in embodiment 2 of the present application;

[0075] Figure 3 is a variable pulse repetition period allocation schematic diagram in embodiment 2 of the present application;

[0076] Figure 4 is a point target variable resolution imaging result diagram in embodiment 2 of the present application;

[0077] Figure 5 is a point target azimuth point spread function diagram in embodiment 2 of the present application;

[0078] Figure 6 is an equivalent experiment flowchart in embodiment 3 of the present application based on measured data;

[0079] Figure 7 is an azimuth filtering algorithm flowchart in embodiment 3 of the present application;

[0080] Figure 8is the equivalent experiment original data imaging result figure in embodiment 3 of the present application;

[0081] Figure 9 is the equivalent experiment variable resolution imaging result figure in embodiment 3 of the present application;

[0082] Figure 10 is the equivalent experiment variable resolution imaging result enlarged contrast figure in embodiment 3 of the present application;

[0083] Figure 11 is the variable resolution SAR one-dimensional imaging result figure based on the maximum mutual information principle in embodiment 3 of the present application;

[0084] Figure 12 is the variable resolution SAR two-dimensional imaging result figure based on the maximum mutual information principle in embodiment 3 of the present application. DETAILED DESCRIPTION

[0085] In order to make the technical means, creative features, purposes and effects of the present application easy to understand, the following embodiments will be specifically described in combination with the drawings.

[0086] <Embodiment 1>

[0087] In embodiment 1, a variable resolution radar imaging method and system are provided.

[0088] Figure 1 is the flow chart of the variable resolution radar imaging method in the embodiment of the present application;

[0089] As shown in Figure 1 , the variable resolution radar imaging method involved in the present embodiment comprises the following steps:

[0090] Step S1, a variable resolution synthetic aperture radar imaging method is proposed, the original scene is divided into multiple sub-scenes, and based on the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method, the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained.

[0091] The specific implementation of step S1 is: according to the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method, the synthetic aperture length under a certain scene is obtained, and the formula is:

[0092]

[0093] δ awhere L represents the synthetic aperture length, R0 represents the reference slant range, and λ represents the wavelength. If the imaging scene is divided into M sub-scenes, different sub-scenes are assigned different azimuthal resolutions, different synthetic aperture lengths and synthetic aperture positions are correspondingly generated, and each sub-aperture is spatially superimposed to generate a total synthetic aperture, and the formula is:

[0094]

[0095] X m where L m represents the azimuthal center coordinate of the mth sub-scene, and L x represents the synthetic aperture length of the mth sub-scene.

[0096] Step S2: The minimum pulse repetition period required for sampling each aperture is obtained according to the azimuthal span required for illumination at each aperture, thereby reducing echo data storage and increasing slant range strip width.

[0097] The specific implementation of step S2 is to calculate the minimum required pulse repetition period according to the azimuthal span required for illumination at each aperture, and the formula is:

[0098]

[0099] where n represents the number of azimuthal sub-scenes illuminated by the current aperture, V represents the radar platform motion speed, and r x represents the azimuthal side length of a sub-scene, PRF n represents the selected pulse repetition period for sampling the current aperture. The variable pulse repetition period method can reduce echo data storage, improve signal processing efficiency, and increase slant range strip width.

[0100] Step S3: Scene information is defined according to the scene scattering amplitude distribution.

[0101] The specific implementation of step S3 is to assume that the original scene obeys a generalized stationary Gaussian distribution, and to divide the scene into a grid. The scene is divided into N a spatial resolution units in the azimuthal direction and N r spatial resolution units in the slant range direction. Based on information theory, the scene information is defined in the wave number domain, and the formula is:

[0102]

[0103] S a where (u, v) represents the covariance spectrum of the scene scattering amplitude, e is a natural constant, A represents the scattering amplitude distribution of the scene in the wave number domain, and H e [A] represents the information of the scene in the wave number domain.

[0104] Step S4: Define the mutual information between the scene and the synthetic aperture radar image based on information theory;

[0105] The specific implementation of step S4 is as follows: Based on information theory, the mutual information between the scene and the obtained synthetic aperture radar image is defined by the formula:

[0106]

[0107] S n (u,v) represents the covariance spectrum of the system noise, H(u,v) represents the imaging system function, and B represents the obtained synthetic aperture radar image.

[0108] Step S5: Based on the principle of maximum mutual information, calculate the mutual information between the original scene and the obtained synthetic aperture radar based on the scene information and the mutual information between the scene and the obtained synthetic aperture radar image, and construct a mathematical model for the variable resolution synthetic aperture radar imaging optimization problem.

[0109] The specific implementation steps of step S5 include:

[0110] Step S5-1: Divide the imaging scene into M sub-scenes. Based on prior knowledge, perform electromagnetic scattering modeling on the scattering amplitude A of each sub-scene, and use the covariance spectrum S a (u,v) representation;

[0111] Step S5-2: Calculate the mutual information between the synthetic aperture radar image and the original scene based on the covariance spectrum of the scattering amplitudes of all sub-scenes. The formula is:

[0112]

[0113] H represents the covariance spectrum of the m-th sub-scene. m (u,v) represents the imaging system function of the m-th sub-scene, which is divided into the following directions in the azimuth direction: Each spatial resolution unit is divided into [number] units along the slant range direction. One spatial resolution unit;

[0114] Step S5-3, the two-dimensional rectangular imaging system function H for each scene m (u,v) can be divided into one-dimensional rectangular imaging system functions in the slant range and azimuth directions, and the formula is:

[0115]

[0116] H represents the one-dimensional rectangular imaging system function representing the orientation of the m-th sub-scene. r(v) represents a one-dimensional matrix imaging system function of the slant range of the mth sub-scene. According to the relationship between the point spread function and the resolution, the formula is:

[0117]

[0118] B r represents the bandwidth, and c represents the speed of light;

[0119] Step S5-4, a two-dimensional covariance spectrum of the scattering amplitude conforming to the Gaussian distribution is constructed, and the formula is:

[0120]

[0121] l x represents the correlation length of the scattering amplitude in the azimuth direction, and l y represents the correlation length of the scattering amplitude in the slant range direction, represents the mean square error of the scattering amplitude;

[0122] Step S5-5, an objective function of an optimization problem is constructed, and the formula is:

[0123]

[0124] represents the correlation length of the scattering amplitude in the azimuth direction under the mth sub-scene, represents the correlation length of the scattering amplitude in the slant range direction under the mth sub-scene.

[0125] Step S6, an equivalent mathematical model of the optimization problem model is constructed, and under the restriction conditions of the lowest signal-to-noise ratio and the priority of the side-looking, the synthetic aperture length of each sub-scene is optimized, so that the mutual information between the original scene synthetic aperture radars is maximized, thereby maximizing the acquisition and characterization of scene information.

[0126] The specific implementation steps in step S6 include:

[0127] Step S6-1, the growth slope of the mutual information of each sub-scene with the growth of the spectrum width is calculated according to formula (6), and a slope vector K

[0128] K = [k1, k2, …, k M ] (12)

[0129] Step S6-2, an illumination matrix P M×T is constructed, and the synthetic aperture vector is defined as

[0130] L = [l1, l2, …, l M ] Τ

[0131]

[0132] T represents the total number of azimuth samples;

[0133] Step S6-3: Construct an optimization objective function equivalent to formula (11), so that the problem can be solved using the CVX toolbox in MATLAB. The formula is as follows:

[0134] arg max KL (14)

[0135] Step S6-4: Set the minimum signal-to-noise ratio (SNR) min Based on the constraints and the relationship between signal-to-noise ratio and antenna gain in synthetic aperture radar images, the formula is:

[0136]

[0137] N A G represents the number of azimuth samples. ave N represents A The average antenna gain of the samples, n0 represents the power spectral density of the noise, P t T represents the peak transmit power. r Indicates the pulse width;

[0138] Step S6-5: Set the aperture position range directly opposite each sub-scene in the orientation direction. The aperture number corresponding to the m-th sub-scene is... The constraint for prioritizing frontal and side views is constructed using the following formula:

[0139]

[0140] Step S6-6: Based on formulas (14), (15), and (16), construct the equivalent mathematical model of the optimization problem. The formula is as follows:

[0141] argmaxKL

[0142]

[0143] Steps S6-7: Use the CVX toolbox in MATLAB to solve the optimization problem of formula (17) and obtain the optimal illumination matrix P. M×T And the pattern assignment maximizes the mutual information between the SAR image and the original scene.

[0144] The variable resolution radar imaging system includes:

[0145] The variable resolution synthetic aperture radar imaging module divides the original scene by using the method of step 1, and obtains the synthetic aperture length and the synthetic aperture position of each sub-scene; the variable pulse repetition period module obtains the minimum pulse repetition period required for each aperture sampling by using the method of step 2; the scene information quantity module defines the scene information quantity according to the scene scattering amplitude distribution by using the method of step 3; the mutual information module defines the mutual information between the scene and the synthetic aperture radar image based on information theory by using the method of step 4; the model construction module constructs the mathematical model of the variable resolution synthetic aperture radar imaging optimization problem by using the method of step 5; the optimization model module maximizes the mutual information between the original scene and the synthetic aperture radar under certain constraints by optimizing the corresponding parameters by using the method of step 6; and the control module controls the above modules as a whole.

[0146] <Embodiment 2>

[0147] In embodiment 2, a specific application scenario of steps 1-2 in embodiment 1 is provided, specifically, the practical application of the point target variable resolution imaging method.

[0148] The specific implementation process of the embodiment includes:

[0149] Step S1, a variable resolution synthetic aperture radar imaging method is proposed, the original scene is divided into multiple sub-scenes, and based on the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method, the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained;

[0150] Figure 2 is the application scenario diagram in embodiment 2 of the present application.

[0151] As shown in Figure 2 , the specific implementation process of the application scenario in the embodiment includes:

[0152] Step S1-1, the scene is divided into two sub-scenes in the azimuth direction, and different azimuth resolutions are set, the synthetic aperture length and the synthetic aperture position corresponding to the scene U and the scene V are calculated according to the set resolution, scene coordinates, wavelength, reference slant range and other parameters;

[0153] Step S1-2, when the radar platform moves to aperture P along the flight line, the beam only irradiates scene 1, as shown by beam E in Figure 1 ;

[0154] Step S1-3, when the radar platform moves to aperture Q, the beam only irradiates scene 2, as shown by beam F in Figure 1 ;

[0155] Step S1-4, when the radar platform moves to the aperture R, which is the common aperture of the two sub-scenes, the beam needs to irradiate scene 1 and scene 2 at the same time, as shown by the beam G in Figure 1 ;

[0156] Step S1-5, the original echo data is processed based on the BP algorithm to form an image.

[0157] Step S2, the minimum pulse repetition period required for sampling at each aperture is obtained according to the azimuth span required to be irradiated at each aperture, so as to reduce echo data storage and increase slant range to strip width;

[0158] Figure 3 is a variable pulse repetition period allocation diagram in embodiment 2 of the present application.

[0159] As shown in Figure 3 , the variable pulse repetition period (PRF) allocation method in the embodiment, the specific implementation process includes:

[0160] Step S2-1, the scene is divided into three sub-scenes in the azimuth direction, and different azimuth resolutions are set, the synthetic aperture length and the synthetic aperture position corresponding to scene U, scene V and scene W are calculated according to the set resolution, scene coordinates, wavelength, reference slant range and other parameters, and are represented by the apertures P, Q and R on the flight line respectively; Figure 2 ;

[0161] Step S2-2, when the radar platform moves along the flight line to the aperture of a certain scene, the beam only irradiates the corresponding sub-scene at the current aperture, as shown by the beams E and F in Figure 2 , and PRF1 is taken at the current aperture for azimuth sampling;

[0162] Step S2-3, when the radar platform moves along the flight line to the common aperture of two scenes, the beam simultaneously irradiates the corresponding two sub-scenes, as shown by the beam G in Figure 2 , and PRF2 is taken at the current aperture for azimuth sampling;

[0163] Step S2-4, when the radar platform moves along the flight line to the common aperture of three scenes, the beam simultaneously irradiates the three sub-scenes, as shown by the beam H in Figure 2 , and PRF3 is taken at the current aperture for azimuth sampling.

[0164] Figure 4 is a point target variable resolution imaging result diagram in embodiment 2 of the present application.

[0165] Figure 5 is a point target azimuth point spread function diagram in embodiment 2 of the present application.

[0166] AsFigure 4 As shown, in this embodiment, the imaging scene is divided into three sub-scenes. The radar parameters include: carrier frequency 9.60GHz, bandwidth 700MHz, pulse duration 10μs, platform movement speed 50m / s, and reference slant range 9.60km. Three pulse repetition cycles are used for sampling, namely 118.66Hz, 79.11Hz and 39.55Hz. The azimuth resolution is set to 0.6m, 0.2m and 1.1m from left to right.

[0167] like Figure 5 As shown, the widths of the point spread function from left to right are 0.46 meters, 0.24 meters, and 0.95 meters, respectively. Compared to strip SAR, the echo data storage space is reduced by approximately 76.30%.

[0168] <Example 3>

[0169] In Example 3, a specific application scenario of steps 1-6 in Example 1 is provided, specifically the practical application of the variable resolution SAR imaging method based on satellite RADARSAT-1 measured data.

[0170] The specific process of this embodiment includes:

[0171] Step S1 proposes a variable resolution synthetic aperture radar imaging method, which divides the original scene into multiple sub-scenes and obtains the synthetic aperture length and synthetic aperture position of each sub-scene based on the relationship between the azimuth resolution and synthetic aperture length in the traditional synthetic aperture radar imaging method.

[0172] Step S2: Based on the azimuth span that needs to be irradiated at each aperture, the minimum pulse repetition period required for sampling at each aperture is obtained, thereby reducing the data storage of echo data and increasing the slant range strip width.

[0173] Figure 6 This is an equivalent experimental flowchart based on measured data in Embodiment 3 of the present invention.

[0174] like Figure 6 As shown, the specific implementation process of raw data processing based on RADARSAT-1 measured data in this embodiment includes:

[0175] Step S1-1: Extract the raw data and divide it into sub-scenes based on the real scene corresponding to the raw data;

[0176] Step S1-2: Calculate the optimal azimuth resolution for each sub-scene based on the geometric relationship of the flight path scene during RADARSAT-1 runtime.

[0177] Steps S1-3: Set the resolution for each sub-scene;

[0178] Step S1-4, calculating a transmitting matrix according to the resolution of each sub-scene;

[0179] Step S1-5, performing phase compensation, range compression and azimuth fast Fourier transform to the RD domain on the original data;

[0180] Step S1-6, calculating an azimuth filter of each sub-scene according to the transmitting matrix, and performing azimuth filtering on the echo data in the RD domain;

[0181] The azimuth filtering algorithm flowchart in step S3-1-6 is shown in Figure 7 The specific steps include:

[0182] Step S1-6-1, inputting the RD domain RADARSAT-1 original echo data after range compression, and selecting a certain aperture;

[0183] Step S1-6-2, calculating the illumination matrix P M×T Designing an antenna pattern at the current aperture, and designing an azimuth filter corresponding to the current aperture according to the pattern;

[0184] Step S1-6-3, azimuth fast inverse Fourier transform;

[0185] Step S1-6-4, extracting the azimuth compressed signal at the current aperture, and replacing the original signal;

[0186] Step S1-6-5, selecting the next aperture until all the signals at the apertures are updated.

[0187] Step S1-7, calculating projection parameters according to the geometric relationship of the track scene, and performing BP imaging.

[0188] Figure 8 is the equivalent experimental original data imaging result graph in embodiment 3 of the application.

[0189] As shown in Figure 8 , the BP imaging result based on the RADARSAT-1 original data in the embodiment is shown, the original data is collected by the squint strip SAR, there are 2000 azimuth samples, the beam squint angle is calculated to be from 1.4607° to 1.6914° through data processing, and the full image resolution is 7.23 meters.

[0190] Figure 9 is the equivalent experimental variable resolution imaging result graph in embodiment 3 of the application.

[0191] As shown in Figure 8 , the BP imaging result based on the RADARSAT-1 variable resolution in the embodiment is shown, the imaging scene is divided into 8 sub-scenes along the azimuth direction, Figure 8The scene resolution from top to bottom is set as 7.23 m, 7.23 m, 20 m, 20 m, 12 m, 12 m, 20 m and 20 m, respectively, and there are 1819 azimuth samples, and the data space is reduced by about 9.05%.

[0192] Figure 10 is an equivalent experimental variable resolution imaging result magnification comparison chart in embodiment 3 of the application.

[0193] As Figure 10 shown, the local magnification comparison results of the strip and variable resolution BP imaging results based on RADARSAT-1 in the embodiment are shown, wherein, Figure 10 (a)-(c) are strip imaging results, Figure 10 (d)-(f) are variable resolution imaging results.

[0194] Step S3, defining scene information quantity according to scene scattering amplitude distribution;

[0195] Step S4, defining mutual information between the scene and the synthetic aperture radar image based on information theory;

[0196] Step S5, based on the maximum mutual information principle, calculating the mutual information between the original scene and the obtained synthetic aperture radar according to the scene information quantity and the mutual information between the scene and the obtained synthetic aperture radar image, and constructing a mathematical model of the variable resolution synthetic aperture radar imaging optimization problem;

[0197] Step S6, constructing an equivalent mathematical model of the optimization problem model, and under the restriction conditions of the lowest signal-to-noise ratio and priority positive side view, maximizing the mutual information between the original scene and the synthetic aperture radar by optimizing the synthetic aperture length of each sub-scene, so as to maximize the acquisition and characterization of scene information.

[0198] Figure 11 is a variable resolution SAR one-dimensional imaging result chart based on the maximum mutual information principle in embodiment 3 of the application.

[0199] As Figure 11As shown in the figure, the one-dimensional imaging scene is divided into five sub-scenes, each sub-scene has an azimuthal side length of 100 meters, each sub-scene is composed of 250 resolution units, the scattering coefficient correlation lengths are 1 meter, 50 meters, 50 meters, 80 meters and 100 meters respectively, and the scattering coefficient mean square deviations are 10, 8, 5, 1 and 0.1 respectively. Under the constraint condition of the same number of azimuthal samples, the resolution of the strip SAR is 0.75 meters, and the resolutions of the optimized variable resolution SAR are 0.29 meters, 0.50 meters, 0.62 meters, 0.86 meters and 1.50 meters respectively. After calculation, when the noise power spectral density is 0.02, the mutual information of the variable resolution SAR is 1837.21 nats, and the mutual information of the strip SAR is 1390.46 nats, which verifies that the variable resolution SAR mode can obtain more and more accurate original scene texture information than the strip SAR under the same resource constraint.

[0200] Figure 12 It is a two-dimensional imaging result figure of the variable resolution SAR based on the maximum mutual information principle in Embodiment 3 of the application.

[0201] As Figure 12 shown, wherein Figure 11 (a) is the scattering coefficient distribution of the original scene, the scattering coefficient correlation lengths of the four sub-scenes from left to right and from bottom to top are 9 meters, 18 meters, 45 meters and 80 meters respectively, and the mean square deviations are 3.5, 3.2, 3.1 and 3.0 respectively. Figure 11 (b) is the strip SAR imaging result, Figure 11 (c) is the variable resolution SAR imaging result. Under the constraint condition of the same number of azimuthal samples and the lowest signal-to-noise ratio of-5dB, the resolution of the strip SAR is 10 meters, the resolutions of the optimized variable resolution SAR are 1.8 meters, 1.8 meters, 0.75 meters and 0.75 meters respectively, the strip SAR obtains mutual information 1159.6 nats, and the variable resolution SAR obtains mutual information 1187.1 nats. This embodiment verifies the feasibility of the variable resolution SAR system optimized based on the maximum mutual information principle and the advantage in information acquisition.

[0202] Effects of the embodiment

[0203] According to the variable resolution radar imaging method related to embodiments 1-3, because the variable resolution synthetic aperture radar imaging method is proposed, the original scene is divided into multiple sub-scenes, and after the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained based on the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method, the minimum pulse repetition period required for the current aperture sampling is obtained according to the azimuth span required to be irradiated at each aperture, the scene information quantity is obtained according to the scene scattering amplitude distribution, the mutual information between the original scene and the obtained synthetic aperture radar image is obtained based on information theory, a mathematical model of the variable resolution synthetic aperture radar imaging optimization problem is constructed based on the maximum mutual information principle, and under the restriction conditions of the minimum signal-to-noise ratio and the priority of the forward-looking, the synthetic aperture length of each sub-scene is optimized, so that the mutual information between the original scene and the synthetic aperture radar is maximized, thereby the scene information is maximizedly obtained and characterized. The above process can balance the contradiction between high resolution and large swath, and can take into account the high resolution of part of the scene of interest while having a larger range in the whole.

[0204] Further, by adopting the variable PRF method, the storage space of the original echo data is reduced without producing azimuth ambiguity, so that the imaging process is more efficient.

[0205] The above embodiments are preferred cases of the present application and do not limit the protection scope of the present application.

Claims

1. A variable resolution radar imaging method, characterized by, The method comprises the following steps: Step 1, a variable resolution synthetic aperture radar imaging method is proposed, the original scene is divided into multiple sub-scenes, and the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained based on the relationship between the azimuth resolution and the synthetic aperture length in a traditional synthetic aperture radar imaging method; Step 2, the minimum pulse repetition period required by each aperture sampling is obtained according to the azimuth span required to be irradiated at each aperture, so as to reduce echo data storage and increase the slant range strip width; Step 3, scene information is defined according to the scene scattering amplitude distribution; Step 4, mutual information between the scene and the obtained synthetic aperture radar image is defined based on information theory; Step 5, based on the maximum mutual information principle, the mutual information between the original scene and the obtained synthetic aperture radar is calculated according to the scene information and the mutual information between the scene and the obtained synthetic aperture radar image, and a mathematical model of a variable resolution synthetic aperture radar imaging optimization problem is constructed; Step 6, an equivalent mathematical model of the optimization problem model is constructed, and under the restriction conditions of the minimum signal-to-noise ratio and the priority forward-looking, the mutual information between the original scene and the synthetic aperture radar is maximized by optimizing the synthetic aperture length of each sub-scene, so as to maximize the acquisition and characterization of scene information, In step 3, it is assumed that the original scene obeys a generalized stationary Gaussian distribution, and the scene is divided into a grid, which is divided into spatial resolution units in the azimuth direction and spatial resolution units in the slant range direction, and the scene information is defined in the wave number domain based on information theory, and the formula is as follows: a covariance spectrum representing the scene scattering amplitudes, is a natural constant, represents the scene scattering amplitude distribution in the wave number domain, represents the scene information content in the wave number domain, In step 4, the mutual information between the scene and the obtained synthetic aperture radar image is defined based on information theory, and the formula is: The covariance spectrum represents the system noise. Represents the imaging system function. This represents the resulting synthetic aperture radar image. In step 5, the specific implementation steps include: Step 5-1, divide the imaging scene into Each sub-scene, based on prior knowledge, determines the scattering amplitude of each sub-scene. Electromagnetic scattering modeling was performed, and covariance spectrum was used. Characterization; Step 5-2, the mutual information between the synthetic aperture radar image and the original scene is calculated according to the covariance spectrum of the scattering amplitude of all sub-scenes, and the formula is: Indicates the first Covariance spectrum of each sub-scene Indicates the first The imaging system function for each sub-scene, the first Each sub-scene is divided into the following directions: Each spatial resolution unit is divided into [number] units along the slant range direction. One spatial resolution unit; Step 5-3, two-dimensional matrix formation imaging system function of each scene The one-dimensional matrix formation imaging system function can be divided into slant range direction and azimuth direction, and the formula is: one-dimensional matrix imaging system function representing the azimuth direction of the one-dimensional matrix imaging system function representing the azimuth direction of the one-dimensional matrix imaging system function representing the slant range direction of thesub-scene, derived from the relationship between the point spread function and the resolution, and is given by denotes the bandwidth, denotes the speed of light; Step 5-4, a two-dimensional covariance spectrum of the scattering amplitude conforming to the Gaussian distribution is constructed, and the formula is: denotes the correlation length of the scattering amplitude in azimuth, denotes the correlation length of the scattering amplitude in slant range, denotes the mean square deviation of the scattering amplitude; Step 5-5, the objective function of the optimization problem is constructed, and the formula is: denotes the correlation length of the scattering amplitude in the azimuth direction under the mth sub-scene, denotes the correlation length of the scattering amplitude in the slant range direction under the mth sub-scene.

2. The variable resolution radar imaging method according to claim 1, wherein: wherein In step 1, the synthetic aperture length under a certain scene is obtained according to the relationship between the azimuth resolution and the synthetic aperture length in the traditional synthetic aperture radar imaging method, and the formula is: ， denotes an azimuth resolution, denotes a synthetic aperture length, denotes a reference slant range, denotes a wavelength, if the imaging scene is divided into different sub-scenes are assigned different azimuth resolutions, then different synthetic aperture lengths and synthetic aperture positions are generated correspondingly, and each sub-aperture space is superimposed to generate a total synthetic aperture, the formula is: ， a center coordinate of an azimuth of a first sub-scene, a center coordinate of an azimuth of a first sub-scene, a center coordinate of an azimuth of a first sub-scene, a center coordinate of an azimuth of a first sub-scene, 3. The variable resolution radar imaging method according to claim 2, wherein: wherein, In step 2, the minimum required pulse repetition period is calculated according to the azimuth span required to be irradiated at each aperture, and the formula is: ， the number of azimuth sub-scene illuminated by the current aperture, the velocity of radar platform motion, the azimuth side length of a sub-scene, the pulse repetition period selected by the current aperture sampling, the method of changing pulse repetition period can reduce echo data storage, improve signal processing efficiency, and increase the width of range-to-bin.

4. The variable resolution radar imaging method according to claim 3, wherein: wherein, In step 6, the specific implementation steps include: Step 6-1, calculate the growth slope of the mutual information of each sub-scene with the growth of the spectrum width according to formula (6), and construct a slope vector , ， Step 6-2, Constructing the illumination matrix , defining the synthetic aperture vector as ， represents the total number of azimuth samples; Step 6-3, the optimization objective function equivalent to formula (11) is constructed, so that the problem can be solved by using the CVX tool box of matlab, and the formula is ， Step 6-4, set minimum signal-to-noise ratio According to the relationship between the signal-to-noise ratio and the antenna gain in the synthetic aperture radar image, the formula is: ， denotes the number of azimuth samples, denotes the average antenna gain of the n-th sample, denotes the power spectral density of the noise, denotes the peak transmit power, denotes the pulse width; Step 6-5, set the aperture position interval of each sub-scene orientation, the aperture number corresponding to the first sub-scene is , the aperture number corresponding to the second sub-scene is , and a restriction condition of preferentially looking at the front is constructed, and the formula is: ， Step 6-6, an equivalent mathematical model of the optimization problem is constructed based on formula (14), formula (15) and formula (16), and the formula is: ， Step 6-7, the optimal problem of formula (17) is solved by using the CVX toolbox of MATLAB and a directivity pattern allocation, such that the mutual information between the SAR image and the original scene is maximized.

5. A system for variational resolution radar imaging according to any one of claims 1 to 4, characterized in that, It comprises: A variable resolution synthetic aperture radar imaging module, a variable resolution synthetic aperture radar imaging method is proposed, the original scene is divided into multiple sub-scenes, and the synthetic aperture length and the synthetic aperture position of each sub-scene are obtained based on the relationship between the azimuth resolution and the synthetic aperture length in a traditional synthetic aperture radar imaging method; a variable pulse repetition period module, which obtains the minimum pulse repetition period required by each aperture sampling according to the azimuth span required to be illuminated at each aperture, thereby reducing echo data storage and increasing slant range to strip width; a scene information amount module, which defines scene information amount according to scene scattering amplitude distribution; a mutual information module, which defines mutual information between a scene and a synthetic aperture radar image based on information theory; a model construction module, which calculates mutual information between an original scene and a synthetic aperture radar based on the maximum mutual information principle, according to the scene information amount and the mutual information between the scene and the obtained synthetic aperture radar image, and constructs a mathematical model of a variable resolution synthetic aperture radar imaging optimization problem; an optimization model module, which constructs an equivalent mathematical model of the optimization problem model, and maximizes the mutual information between the original scene and the synthetic aperture radar by optimizing the synthetic aperture length of each sub-scene under the constraints of minimum signal-to-noise ratio and priority to the forward-looking, so as to maximize the acquisition and characterization of scene information.

Images