Near-field three-dimensional imaging system and method for millimeter-wave synthetic aperture radiometer

Abstract

The present invention discloses a near-field three-dimensional imaging system and method for a millimeter wave synthetic aperture radiometer in the field of millimeter wave near-field imaging technology, comprising: obtaining a millimeter wave signal of a target scene; constructing a visibility function based on the millimeter wave signal; inverting a two-dimensional scene graph under different distance parameters based on the visibility function; constructing an image-distance data set based on the two-dimensional scene graph; performing partition calibration on the image-distance data set to construct a distance function for targets in different regions; constructing an imaging model based on the distance function, and performing adaptive reconstruction inversion on the target scene in different regions to obtain a millimeter wave image. The present invention can improve the accuracy of imaging inversion images of complex close-range target scenes.

Description

Technical Field

[0001] The invention relates to a near-field three-dimensional imaging system and method for a millimeter wave synthetic aperture radiometer, and belongs to the technical field of millimeter wave near-field imaging. Background Art

[0002] The millimeter-wave synthetic aperture radiometer uses synthetic aperture technology to synthesize a large-aperture antenna using a small-aperture array element antenna to achieve higher-resolution imaging detection. It also has the characteristics of both infrared and microwave imaging, and can penetrate clothing, plastic, wood, etc. to achieve hidden detection. It has strong application prospects in remote sensing monitoring, military, navigation, medical and security inspection and other fields.

[0003] Currently, due to the limitation of the system array scale, the number of visibility points measured by the aperture radiometer is relatively small; and due to the interference of background noise and system errors, the millimeter wave images measured by the aperture radiometer have large errors; in near-field applications, affected by spherical waves, there is a large modeling error between the traditional far-field imaging model and the actual near-field imaging process, making it difficult for traditional imaging methods to achieve accurate near-field imaging inversion. In order to achieve accurate near-field aperture radiometer imaging inversion, many near-field aperture radiometer algorithms have been proposed. According to the implementation principle of the algorithm, they can be divided into two categories: one is the MFFT algorithm based on Fourier transform, which performs near-field phase correction on the visibility function and then performs FFT operation on the corrected visibility function to obtain the brightness temperature image of the target scene. This type of algorithm is simple to operate, but it cannot eliminate the near-field modeling error of the imaging model, and the inverted image has large reconstruction errors and noise interference. The other type is a regularization algorithm based on the near-field G matrix. It reduces the modeling error of the imaging model by accurately mathematically modeling the near-field imaging process. It also uses the regularization algorithm to numerically invert the target image based on the prior information of the imaging process and the brightness temperature image, thereby achieving more accurate image reconstruction. This type of algorithm has a high accuracy in reconstructing the model and a relatively accurate reconstruction result. It is currently a commonly used near-field imaging method. However, it has extremely high requirements for the accuracy of the model parameters. For targets with multiple distance phases in the same scene, it is difficult to achieve accurate reconstruction and inversion, and there are certain limitations in practical applications. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a near-field three-dimensional imaging system and method for a millimeter-wave synthetic aperture radiometer, which can improve the imaging inversion image accuracy of complex close-range target scenes.

[0005] To achieve the above object, the present invention is implemented by adopting the following technical solutions:

[0006] In a first aspect, the present invention provides a near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer, comprising:

[0007] Acquire the millimeter wave signal of the target scene;

[0008] Construct visibility function based on millimeter wave signals;

[0009] Based on the visibility function, a two-dimensional scene graph with different distance parameters is inverted;

[0010] Construct an image-distance dataset based on a two-dimensional scene graph;

[0011] Perform partition calibration on the image-distance dataset and construct distance functions for targets in different regions;

[0012] An imaging model is constructed based on the distance function, and the target scene is adaptively reconstructed and inverted in different regions to obtain a millimeter wave image.

[0013] Furthermore, a visibility function is constructed based on the millimeter wave signal, including:

[0014] After accumulating the millimeter wave signals received by the array antennas, the visibility function is obtained by complex correlation calculation between the antennas. The sample points of the visibility function are expressed as:

[0015]

[0016] Among them, V c,l represents the visibility function sample point, Indicates the distance from antenna c to radiation source S i The distance Represents the distance from antenna l to radiation source S i distance, t represents the propagation time, represents the electromagnetic signal received by the array element antenna c, represents the conjugate of the electromagnetic signal received by the array element antenna l, <·> represents the time integration, τ is the integration time, N represents the number of discrete point radiation sources, (x i ,y i ) is the radiation source S i Coordinate, T(x i ,y i ) is the normalized brightness temperature, F c (x i ,y i ) represents the antenna pattern of antenna c, represents the conjugate of the antenna pattern of antenna l, r c,l is the stripe elimination function, represents the phase difference between the antenna pair (c,l).

[0017] 4. Further, based on the visibility function, the two-dimensional scene graph under different distance parameters is inverted, including:

[0018] The distance from the antenna to the radiation source is expressed as:

[0019]

[0020]

[0021] In the formula, Indicates the distance from antenna c to radiation source S i The distance Represents the distance from antenna l to radiation source S i The distance (X c ,Y c ) and (X l ,Y l ) represents the coordinates of antennas c and l, (x i ,y i ) represents the radiation source S i Coordinates, R represents the distance from the radiation source plane to the antenna plane;

[0022] Substituting the distance expression from the antenna to the radiation source into the visibility function, we get the matrix equation:

[0023] V M×1 =G M×N ·T N×1 (4)

[0024]

[0025] Where V M×1 Represents the vector form of the visibility matrix, G M×N Represents the vector form of the sensor matrix, T N×1 represents the vector form of the brightness temperature image to be obtained, G(m,n) represents the expression form of the sensor matrix, and F mc (xi,yi) represents the antenna pattern of antenna c corresponding to the mth visibility function sample point, represents the conjugate of the antenna pattern of the lth antenna corresponding to the mth visibility function sample point, (X ml ,Y ml ) and (X mc ,Y mc ) represents the coordinates of the two antennas corresponding to the mth visibility function sample point;

[0026] The calculated sensor matrix is ​​iteratively inverted and solved by a regularization method to obtain a two-dimensional scene image under different distance parameters. The solution model is:

[0027]

[0028]

[0029] In the formula, minT J(T) represents the minimum cost function of the optimal parameters, G R represents the sensing matrix G with distance parameter R, T represents the reconstructed image, V represents the visibility function, T(i,j) represents the current pixel of the reconstructed image, T(i,j+1) represents the next column of pixels of the current pixel, and T(i+1,j) represents the next row of pixels of the current pixel;

[0030] The model is solved by the fast projected gradient FGP iterative inversion algorithm to obtain the two-dimensional scene images under different distance parameters.

[0031] Furthermore, constructing a three-dimensional image-distance dataset based on the two-dimensional scene graph includes: arranging the two-dimensional target scene graph according to its distance parameter to construct the three-dimensional image-distance dataset.

[0032] Furthermore, the three-dimensional image-distance dataset is partitioned and calibrated to construct distance functions for targets in different regions, including:

[0033] The distance parameter of the three-dimensional data set is corrected by region according to the image clarity. For the ki-th partition, the image clarity f in the region ki (R) is:

[0034] f ki (R) = EOG[T(x,y,R).*B ki (x,y)] (8)

[0035] Where EOG is the energy gradient function, which is used to characterize the self-definition of the image; B ki (x,y) is the position label matrix of the ki-th partition, and the expression is:

[0036]

[0037] In the formula, (x ki ,y ki ) is the center coordinate of the ki-th partition, N s is the size of the partition block, usually set to 1 / 10 of the image dimension;

[0038] According to the clarity of each area ki (R) Determine max[f ki (R)] corresponds to the distance R ki , construct the distance function R(x,y):

[0039] R(x,y)=mean(R ki ),if (x,y)∈B ki (10)

[0040] In the formula, mean(R ki) means (x,y) in all partitions R ki The average value of B ki Represents the position label matrix of the ki-th partition.

[0041] Furthermore, an imaging model is constructed based on the distance function, and adaptive reconstruction and inversion are performed on the target scene in different regions to obtain a near-field millimeter wave image, including:

[0042] Substituting the distance function into the sensor matrix, we get the matrix imaging equation:

[0043] V=G 3D ·T (11)

[0044]

[0045] In the formula, G 3D represents the sensor matrix G after parameter calibration, T represents the reconstructed image, V represents the visibility function, F mc (xi,yi) represents the antenna pattern of antenna c corresponding to the mth visibility function sample point, represents the conjugate of the antenna pattern of the lth antenna corresponding to the mth visibility function sample point, (X ml ,Y ml ) and (X mc ,Y mc ) represents the coordinates of the two antennas corresponding to the mth visibility function sample point, 2π / λ is the circular wave number, and λ is the central wavelength of the system;

[0046] A solution model is constructed for the matrix imaging equation through a regularization method, and then a fast gradient projection FGP iterative inversion algorithm is used to solve the equation to obtain a millimeter wave image.

[0047] Furthermore, the distance function is smoothed and filtered before the matrix imaging equation is reconstructed and inverted.

[0048] In a second aspect, the present invention provides a near-field three-dimensional imaging system for a millimeter-wave synthetic aperture radiometer, comprising:

[0049] Sampling module: used to obtain the millimeter wave signal of the target scene;

[0050] Function building module: used to build visibility function based on millimeter wave signals;

[0051] Inversion module: used to invert the two-dimensional scene graph under different distance parameters based on the visibility function;

[0052] Dataset construction module: used to construct image-distance dataset based on two-dimensional scene graph;

[0053] Calibration module: used to calibrate the image-distance dataset in different regions and construct distance functions for targets in different regions;

[0054] Reconstruction and inversion module: used to build an imaging model based on the distance function, perform adaptive reconstruction and inversion on the target scene in different regions, and obtain millimeter wave images.

[0055] In a third aspect, the present invention provides a near-field three-dimensional imaging device for a millimeter-wave synthetic aperture radiometer, including a processor and a storage medium;

[0056] The storage medium is used to store instructions;

[0057] The processor is used to operate according to the instructions to execute the steps of any of the methods described above.

[0058] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of any of the above methods when executed by a processor.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] The near-field imaging method based on adaptive correction proposed in the present invention can perform partition correction on targets of different distance phases according to the clarity of the reconstructed image, and achieve more accurate target reconstruction based on the corrected parameter information, so as to better realize the description of the actual scene by the imaging model. Compared with the traditional synthetic aperture imaging method, it can achieve more sophisticated three-dimensional imaging inversion for complex close-range target scenes. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A two-dimensional sparse "T" array distribution diagram provided in Example 1 of the present invention;

[0062] Figure 2 A schematic diagram of the imaging principle of a synthetic aperture radiometer provided in the first embodiment of the present invention;

[0063] Figure 3 A schematic diagram of a target scene provided in the first embodiment of the present invention;

[0064] Figure 4 A three-dimensional distribution map of ship targets provided in the first embodiment of the present invention;

[0065] Figure 5 A distribution diagram of the corrected distance R (x, y) of a ship target provided in the first embodiment of the present invention;

[0066] FIG6( a ) is an image result of a ship target adaptive correction provided by Embodiment 1 of the present invention;

[0067] FIG6( b ) is a conventional imaging result of a ship target provided by the first embodiment of the present invention (R=7 m);

[0068] FIG6( c ) is a normalized target scene diagram of a ship target provided in Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0069] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0070] Embodiment 1:

[0071] The present embodiment discloses a near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer, which mainly includes: firstly, constructing a high-precision synthetic aperture imaging model based on the near-field imaging principle, and performing preliminary scene image reconstruction; then, partitioning and calibrating the model sensitive parameter R according to the clarity of the reconstructed image, and determining the model parameter values ​​of different target areas; finally, using a regularization method to perform adaptive reconstruction and inversion of the target scene in different regions, and obtaining an accurate near-field millimeter-wave image.

[0072] like Figure 1 As shown, the present invention adopts a two-dimensional sparse array layout, and a "T" antenna array is composed of 59 array element antennas, which are fixed on the platform.

[0073] Step 1: Use the sparse array of millimeter-wave comprehensive aperture radiation to passively detect the target scene, obtain the millimeter-wave signal of the target scene, and construct the visibility function of the target scene. The specific implementation method is as follows:

[0074] The schematic diagram of the imaging principle of the comprehensive aperture radiometer of the present invention is as follows Figure 2 As shown in the figure, the array antenna (c, l) is located on the OXY plane, and the target S is located near the oxy plane. The visibility function is obtained by accumulating the millimeter wave signals received by the array antenna and performing complex correlation operations between the antennas. According to the principle of synthetic aperture interferometry, the sample points of the visibility function measured by the array antenna pair (c, l) can be expressed as:

[0075]

[0076] Among them, V c,l represents the visibility function sample point, Indicates the distance from antenna c to radiation source S i The distance Represents the distance from antenna l to radiation source S i distance, t represents the propagation time, represents the electromagnetic signal received by the array element antenna c, represents the conjugate of the electromagnetic signal received by the array element antenna l, <·> represents the time integration, τ is the integration time, N represents the number of discrete point radiation sources, (x i ,y i ) is the radiation source S i Coordinate, T(x i ,y i ) is the normalized brightness temperature, F c (x i ,y i ) represents the antenna pattern of antenna c, represents the conjugate of the antenna pattern of antenna l, r c,l is the stripe elimination function, represents the phase difference between the antenna pair (c,l).

[0077] The present invention verifies and analyzes the proposed near-field three-dimensional imaging method by experimental simulation. The main parameters used in the simulation are shown in Table 1; the array used in the simulation is "T" shaped, and its array elements are arranged as follows Figure 1 As shown, ● is the real element antenna and ○ is the visibility sampling point.

[0078]

[0079] Table 1. Main inversion parameters

[0080] The brightness temperature distribution of the target scene is as follows: Figure 3 As shown in the figure, the dimension is 100×100. In order to simulate the actual millimeter wave comprehensive aperture radiometer imaging process, the simulation Figure 3 The grayscale value of each pixel in is taken as the radiation intensity of the point source here, and the distance between the radiation point sources is set to half of the system spatial resolution; the distances between the four target ships and the array antenna are [5.06.78.39.5] m respectively, and their three-dimensional distribution diagram is shown in Figure 4 As shown: The received signal of the array element antenna is simulated by accumulating the millimeter wave signals generated by all point sources, and then the visibility function is obtained by complex correlation operation between antennas.

[0081] Step 2: With the help of the high-precision G matrix imaging algorithm, the two-dimensional scene graph under different distance parameters is inverted from the visibility function in step 1 to construct a three-dimensional image-distance data set T(x, y, R). The specific implementation method is as follows:

[0082] The distance from the antenna to the radiation source is expressed as:

[0083]

[0084]

[0085] In the formula, Indicates the distance from antenna c to radiation source Si The distance Represents the distance from antenna l to radiation source S i The distance (X c ,Y c ) and (X l ,Y l ) represents the coordinates of antennas c and l, (x i ,y i ) represents the radiation source S i The coordinates of the antenna are shown in Figure 2, and R represents the distance from the radiation source plane to the antenna plane. Substituting the distance expression from the antenna to the radiation source into the visibility function, we get the matrix equation:

[0086] V M×1 =G M×N ·T N×1 (4)

[0087]

[0088] Where V M×1 Represents the vector form of the visibility matrix, G M×N Represents the vector form of the sensor matrix, T N×1 represents the vector form of the brightness temperature image to be obtained, G(m,n) represents the expression form of the sensor matrix, and F mc (xi,yi) represents the antenna pattern of antenna c corresponding to the mth visibility function sample point, represents the conjugate of the antenna pattern of the lth antenna corresponding to the mth visibility function sample point, (X ml ,Y ml ) and (X mc ,Y mc ) represents the coordinates of the two antennas corresponding to the mth visibility function sample point.

[0089] Different distance R i The following sensing matrix G satisfies:

[0090] V=G Ri ·T Ri ,i=1,2,3,... (6)

[0091] Where V represents the visibility function, G Ri Represents R i Sensing matrix G, T at distance Ri Represents the reconstructed image at Ri distance.

[0092] Take the distance R from the radiation source plane to the antenna plane i is the estimated distance R 0 The confidence interval is 0.5R 0 ~1.5R 0, the visibility function under different distance parameters is reconstructed and inverted respectively to obtain the two-dimensional scene images under different distances.

[0093] The sensor matrix G of equation (6) is inverted and solved by using the regularization method, and the solution model is:

[0094]

[0095] In the formula, min T J(T) represents the minimum cost function of the optimal parameters, G R represents the sensing matrix G with distance parameter R, T represents the reconstructed image, V represents the visibility function, α is the regularization parameter, ||GT-V|| 2 2 is the fidelity term, which is used to ensure that the difference between the reconstructed image and the true millimeter wave image is small enough. P(T) is the regularization term of the total variation function, and its expression is:

[0096]

[0097] Where T(i, j) represents the current pixel of the reconstructed image, T(i, j+1) represents the next column of pixels of the current pixel, and T(i+1, j) represents the next row of pixels of the current pixel.

[0098] The two-dimensional scene images under different distance parameters are obtained through the fast gradient projection (FGP) iterative inversion algorithm. The two-dimensional target scene images are arranged according to their distance parameter R to construct a three-dimensional image-distance dataset T(x, y, R).

[0099] Step 3: Based on the image clarity, the image-distance dataset T(x, y, R) is partitioned and calibrated to construct the distance function R(x, y) of targets in different regions. The specific implementation method is as follows:

[0100] The present invention calibrates the distance parameter R(x, y) of targets in different regions based on the image-distance data set T(x, y, R), and the specific method is as follows:

[0101] (1) The distance parameter of the three-dimensional image-distance data set is calibrated in different regions according to the image clarity. For the ki-th partition, the image clarity f in the region ki (R) is:

[0102] f ki (R) = EOG[T(x,y,R).*B ki (x,y)] (9)

[0103] Where EOG is the energy gradient function, which is used to characterize the self-definition of the image; B ki(x,y) is the position label matrix of the ki-th partition, and the expression is:

[0104]

[0105] In the formula, (x ki ,y ki ) is the center coordinate of the ki-th partition, N s is the size of the partition block, usually set to 1 / 10 of the image dimension.

[0106] (2) Based on the clarity of each area ki (R) Determine its accurate distance parameter, that is, find max[f ki (R)] corresponds to the distance R ki . Then construct the accurate distance function R(x,y):

[0107] R(x,y)=mean(R ki ),if(x,y)∈B ki (11)

[0108] In the formula, mean(R ki ) means (x,y) in all partitions R ki The average value of B ki Represents the position label matrix of the ki-th partition.

[0109] The distribution of the corrected distance R(x,y) of the ship scene is as follows Figure 5 As shown in the figure, it can be seen that for the main target (grey value greater than 20), the calibration of its distance parameter is relatively accurate, but due to the influence of noise, the faint target is submerged in the noise, and it is difficult to accurately determine its distance parameter.

[0110] Step 4: Based on the distance function in step 3, an accurate 3D-G imaging model is constructed to achieve more accurate image reconstruction for complex close-range target scenes. The specific implementation method is as follows:

[0111] (1) After obtaining the accurate distance parameter R(x, y), substitute it into equations (2) and (3) to obtain the accurate 3D-G matrix imaging equation:

[0112] V=G 3D ·T (12)

[0113]

[0114] (2) Using the aforementioned regularization algorithm for reconstruction inversion;

[0115] (3) To overcome the image break caused by the distance function R(x,y), a smoothing filter is first performed in the iterative inversion;

[0116] (4) The fast gradient projection (FGP) iterative inversion algorithm is used to solve the problem, with the maximum iteration step k set to 20 and the parameter α set to 0.07.

[0117] After obtaining a relatively accurate distance R(x,y), it is brought into the self-correction imaging model, and the millimeter wave image of the target scene is obtained by reconstruction and inversion as shown in Figure 6(a). By performing partitioned self-correction reconstruction and inversion on targets in different areas, a relatively accurate reconstructed image is obtained, and the target information of the four ships is relatively clear. For comparison, with the help of the traditional regularization algorithm, imaging inversion is performed under the same conditions (the distance R is set to the median value of 7m), and the reconstructed image result is shown in Figure 6(b). It can be seen that only the image of the second ship is clear (its distance is close to 7m); due to the large deviation between the distance parameter and the actual target distance, the images of the other three ships are relatively blurred and difficult to distinguish.

[0118] In order to objectively evaluate the accuracy of the reconstructed image, the root mean square error (RMSE), peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) of the inversion image of the traditional imaging method and the inversion image of the near-field 3D imaging method were calculated. The results are shown in Table 2:

[0119]

[0120] Table 2. Comparison of evaluation data of reconstructed images

[0121] Obviously, the RMSE of the inversion results of the near-field 3D imaging method based on adaptive correction is small, and the PSNR and SSIM are large. The results show that the proposed near-field 3D imaging method can effectively cope with complex close-range target scenes and achieve accurate near-field multi-target imaging inversion.

[0122] Embodiment 2:

[0123] A near-field three-dimensional imaging system for a millimeter-wave synthetic aperture radiometer can implement the near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer in Embodiment 1, including:

[0124] Sampling module: used to obtain the millimeter wave signal of the target scene;

[0125] Function building module: used to build visibility function based on millimeter wave signals;

[0126] Inversion module: used to invert the two-dimensional scene graph under different distance parameters based on the visibility function;

[0127] Dataset construction module: used to construct image-distance dataset based on two-dimensional scene graph;

[0128] Calibration module: used to calibrate the image-distance dataset in different regions and construct distance functions for targets in different regions;

[0129] Reconstruction and inversion module: used to build an imaging model based on the distance function, perform adaptive reconstruction and inversion on the target scene in different regions, and obtain millimeter wave images.

[0130] Embodiment three:

[0131] The embodiment of the present invention also provides a near-field three-dimensional imaging device for a millimeter wave comprehensive aperture radiometer, which can implement the near-field three-dimensional imaging method for a millimeter wave comprehensive aperture radiometer in the first embodiment, including a processor and a storage medium;

[0132] The storage medium is used to store instructions;

[0133] The processor is used to operate according to the instructions to perform the steps of the following method:

[0134] Acquire the millimeter wave signal of the target scene;

[0135] Construct visibility function based on millimeter wave signals;

[0136] Based on the visibility function, a two-dimensional scene graph with different distance parameters is inverted;

[0137] Construct an image-distance dataset based on a two-dimensional scene graph;

[0138] Perform partition calibration on the image-distance dataset and construct distance functions for targets in different regions;

[0139] An imaging model is constructed based on the distance function, and the target scene is adaptively reconstructed and inverted in different regions to obtain a millimeter wave image.

[0140] Embodiment 4:

[0141] The embodiment of the present invention further provides a computer-readable storage medium, which can implement the near-field three-dimensional imaging method for the millimeter-wave synthetic aperture radiometer in the first embodiment, and stores a computer program thereon, which implements the steps of the following method when the program is executed by a processor:

[0142] Acquire the millimeter wave signal of the target scene;

[0143] Construct visibility function based on millimeter wave signals;

[0144] Based on the visibility function, a two-dimensional scene graph with different distance parameters is inverted;

[0145] Construct an image-distance dataset based on a two-dimensional scene graph;

[0146] Perform partition calibration on the image-distance dataset and construct distance functions for targets in different regions;

[0147] An imaging model is constructed based on the distance function, and the target scene is adaptively reconstructed and inverted in different regions to obtain a millimeter wave image.

[0148] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0149] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0150] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0151] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0152] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. Near-field three-dimensional imaging method for millimeter-wave synthetic aperture radiometer, Its characteristics are: include: Acquire the millimeter wave signal of the target scene; Construct visibility function based on millimeter wave signals; Based on the visibility function, a two-dimensional scene graph with different distance parameters is inverted; Construct an image-distance dataset based on a two-dimensional scene graph; Perform partition calibration on the image-distance dataset and construct distance functions for targets in different regions; An imaging model is constructed based on the distance function, and adaptive reconstruction and inversion are performed on the target scene in different regions to obtain a millimeter wave image. The image-distance dataset is partitioned and calibrated to construct distance functions for targets in different regions, including: According to the image clarity, the distance parameter of the image-distance data set is corrected in different regions. For the ki-th partition, the image clarity f in the region ki (R) is: f ki (R)=EOG[T(x,y,R).*B ki (x,y)] (8) Where EOG is the energy gradient function, which is used to characterize the self-definition of the image; T(x, y, R) is the image-distance dataset; B ki (x,y) is the position label matrix of the ki-th partition, and the expression is: In the formula, (x ki ,y ki ) is the center coordinate of the ki-th partition, N s is the size of the partition block; According to the clarity of each area ki (R) Determine max[f ki (R)] corresponds to the distance R ki , construct the distance function R(x,y): R(x,y)=mean(R ki ),if(x,y)∈B ki (10) In the formula, mean(R ki ) means (x,y) in all partitions R ki The average value of B ki Represents the position label matrix of the ki-th partition.

2. The near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer according to claim 1, Its characteristics are: Construct visibility functions based on millimeter wave signals, including: After accumulating the millimeter wave signals received by the array antennas, the visibility function is obtained by complex correlation calculation between the antennas. The sample points of the visibility function are expressed as: Among them, V c,l represents the visibility function sample point, Represents the distance from antenna c to radiation source S i The distance Represents the distance from antenna l to radiation source S i distance, t represents the propagation time, represents the electromagnetic signal received by the array element antenna c, represents the conjugate of the electromagnetic signal received by the array element antenna l, <·> represents the time integration, τ is the integration time, N represents the number of discrete point radiation sources, (x i ,y i ) is the radiation source S i Coordinate, T(x i ,y i ) is the normalized brightness temperature, F c (x i ,y i ) represents the antenna pattern of antenna c, represents the conjugate of the antenna pattern of antenna l, r c,l is the stripe elimination function, represents the phase difference between the antenna pair (c,l).

3. The near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer according to claim 2, Its characteristics are: Based on the visibility function, the two-dimensional scene graph under different distance parameters is inverted, including: The distance from the antenna to the radiation source is expressed as: In the formula, Indicates the distance from antenna c to radiation source S i The distance Represents the distance from antenna l to radiation source S i The distance (X c ,Y c ) and (X l ,Y l ) represents the coordinates of antennas c and l, (x i ,y i ) represents the coordinates of the radiation source Si, and R represents the distance from the radiation source plane to the antenna plane; Substituting the distance expression from the antenna to the radiation source into the visibility function, we get the matrix equation: V M×1 =G M×N ·T N×1 (4) Where V M×1 Represents the vector form of the visibility matrix, G M×N Represents the vector form of the sensor matrix, T N×1 represents the vector form of the brightness temperature image to be obtained, G(m,n) represents the expression form of the sensor matrix, and F mc (x i ,y i ) represents the antenna pattern of antenna c corresponding to the mth visibility function sample point, represents the conjugate of the antenna pattern of the lth antenna corresponding to the mth visibility function sample point, (X ml ,Y ml ) and (X mc ,Y mc ) represents the coordinates of the two antennas corresponding to the mth visibility function sample point; 2π / λ is the circular wave number, λ is the central wavelength of the system; The calculated sensor matrix is ​​iteratively inverted and solved by a regularization method to obtain a two-dimensional scene image under different distance parameters. The solution model is: In the formula, min T J(T) represents the minimum cost function of the optimal parameters, G R represents the sensing matrix G with distance parameter R, T represents the reconstructed image, V represents the visibility function, P(T) is the regularization term of the total variation function, α is the regularization parameter, T(i,j) represents the current pixel of the reconstructed image, T(i,j+1) represents the next column of pixels of the current pixel, and T(i+1,j) represents the next row of pixels of the current pixel; The model is solved by the fast projected gradient FGP iterative inversion algorithm to obtain the two-dimensional scene images under different distance parameters.

4. The near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer according to claim 1, Its characteristics are: An image-distance dataset is constructed based on a two-dimensional scene graph, including: arranging the two-dimensional target scene graph according to its distance parameter to construct a three-dimensional image-distance dataset.

5. The near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer according to claim 1, Its characteristics are: An imaging model is constructed based on the distance function, and adaptive reconstruction and inversion are performed on the target scene in different regions to obtain millimeter wave images, including: Substituting the distance function into the sensor matrix, we get the matrix imaging equation: V=G 3D ·T (11) In the formula, G 3D represents the sensor matrix G after parameter calibration, T represents the reconstructed image, V represents the visibility function, F mc (x i ,y i ) represents the antenna pattern of antenna c corresponding to the mth visibility function sample point, represents the conjugate of the antenna pattern of the lth antenna corresponding to the mth visibility function sample point, (X ml ,Y ml ) and (X mc ,Y mc ) represents the coordinates of the two antennas corresponding to the mth visibility function sample point, (x i ,y i ) is the radiation source S i coordinates, 2π / λ is the circular wave number, and λ is the central wavelength of the system; A solution model is constructed for the matrix imaging equation through a regularization method, and then a fast gradient projection FGP iterative inversion algorithm is used to solve the equation to obtain a millimeter wave image.

6. The near-field three-dimensional imaging method for a millimeter-wave synthetic aperture radiometer according to claim 5, Its characteristics are: The distance function is smoothed and filtered once before the matrix imaging equation is reconstructed and inverted.

7. Near-field three-dimensional imaging system for millimeter-wave synthetic aperture radiometer, Its characteristics are: include: Sampling module: used to obtain the millimeter wave signal of the target scene; Function building module: used to build visibility function based on millimeter wave signals; Inversion module: used to invert the two-dimensional scene graph under different distance parameters based on the visibility function; Dataset construction module: used to construct image-distance dataset based on two-dimensional scene graph; Calibration module: used to calibrate the image-distance dataset in different regions and construct distance functions for targets in different regions; Reconstruction and inversion module: used to build an imaging model based on the distance function, perform adaptive reconstruction and inversion of the target scene in different regions, and obtain millimeter wave images; The calibration module is specifically used for: According to the image clarity, the distance parameter of the image-distance data set is corrected in different regions. For the ki-th partition, the image clarity f in the region ki (R) is: f ki (R)=EOG[T(x,y,R).*B ki (x,y)] (8) Where EOG is the energy gradient function, which is used to characterize the self-definition of the image; T(x, y, R) is the image-distance dataset; B ki (x,y) is the position label matrix of the ki-th partition, and the expression is: In the formula, (x ki ,y ki ) is the center coordinate of the ki-th partition, N s is the size of the partition block; According to the clarity of each area ki (R) Determine max[f ki (R)] corresponds to the distance R ki , construct the distance function R(x,y): R(x,y)=mean(R ki ),if(x,y)∈B ki (10) In the formula, mean(R ki ) means (x,y) in all partitions R ki The average value of B ki Represents the position label matrix of the ki-th partition.

8. Near-field three-dimensional imaging device for millimeter-wave synthetic aperture radiometer, Its characteristics are: including processor and storage medium; The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, Its characteristics are: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images