CN114781152B - Pseudo-target radiation source distinguishing method and system based on array factor characteristics - Google Patents

Abstract

The invention discloses a pseudo target radiation source distinguishing method and system based on array factor characteristics, and belongs to the field of array signal processing and source positioning. Comprising the following steps: reconstructing an image of the target radiation source; constructing a distance matrix from each source point to other source points; establishing a multi-antenna array factor model diagram; determining a side lobe point with the maximum relative intensity in a certain error range on an array factor model diagram, taking the side lobe point as a suspicious source point, and determining the position information and the distribution characteristics of the suspicious source point in a model; and comparing the actual position information with the position information and the distribution characteristics obtained in the model diagram on the reconstructed target radiation source image, and finally determining the specific position of the pseudo target radiation source point on the reconstructed target radiation source image. The invention solves the detection and processing problems of false source points generated by large fluctuation range of the intensity of the interference source, reduces the false alarm probability of the radiation interference source detection in the discrete multi-target image, and has higher robustness.

Description

A method and system for distinguishing false target radiation sources based on array factor characteristics

Technical Field

The present invention belongs to the field of array signal processing and source positioning, and more specifically, relates to a pseudo target radiation source identification method and system based on array factor characteristics.

Background technique

Radio frequency interference is an important factor affecting the scientificity and practicality of satellite/airborne remote sensing data, and will restrict the application and development of remote sensing such as geophysical parameter inversion, microwave imaging, and target detection. Whether it is through the coordination between the ITU and local management departments to shut down illegal target radiation sources, or directly processing contaminated remote sensing data to mitigate the impact of radio frequency interference, it is usually necessary to use the location information of the radiation source.

In the prior art, the existing target radiation source detection methods have many defects, such as the inability to detect and locate mixed strong and weak sources, insufficient resolution of adjacent interference sources, and the inability to detect and locate position-aggregated sources. Especially in the case of mixed strong and weak sources, the positioning, discrimination and processing of pseudo-target radiation sources are urgent problems to be solved. For example, the existing RFI detection algorithm usually sets a threshold as the judgment standard. This method cannot detect false positive points and will cover up weak sources, and the precision of precise measurement is not high.

Summary of the invention

In view of the defects and improvement needs of the prior art, the present invention provides a method and system for distinguishing pseudo target radiation sources based on array factor characteristics, which aims to screen out pseudo target radiation sources in the case of a mixture of strong and weak sources and improve the detection accuracy.

To achieve the above object, according to one aspect of the present invention, a method for distinguishing a pseudo target radiation source based on array factor characteristics is provided, comprising:

reconstructing an image of the target radiation source;

Traverse each pixel point in the reconstructed image, determine the maximum value and position of each source point in the reconstructed image, regard the position of the maximum value as the position of the source point, and construct a distance matrix D from each source point to other source points;

Establish a multi-antenna array factor model diagram;

On the array factor model diagram, determine the side lobe point with the largest relative intensity within a certain error range, take the side lobe point as a suspected source point, determine the position information and distribution characteristics of the suspected source point in the model; and calculate the distance d _t from the suspected source point to the main lobe according to the position information and the position information of the main lobe in the model, where t is a positive integer;

On the reconstructed image, for each d _t , traverse each column in the distance matrix D, and mark the source point whose distance is within the error range of d _t as the suspicious source point of the source point;

The distribution characteristics of the suspicious source points corresponding to each source point are counted. When the distribution characteristics are consistent with the distribution characteristics in the array factor model, the suspicious source point is determined to be the pseudo target radiation source point corresponding to the source point, and the positioning result of the pseudo target radiation source point is obtained.

Furthermore, by obtaining the difference correlation array Difference Coarray of the original sparse array of multiple antennas, the multi-antenna array factor model graph is constructed according to the distribution of the difference correlation array Difference Coarray.

Furthermore, a convex optimization algorithm is used to obtain the reconstructed image of the target radiation source.

Furthermore, the step of reconstructing the image of the target radiation source includes:

Step S1, obtaining a covariance matrix of an original sparse array of multiple antennas, redundantly averaging and vectorizing the covariance matrix to construct a difference coarray signal receiving model of the original sparse array;

Step S2, using the difference correlation array Difference Coarray of the original sparse array and the spatial sparse characteristics of the target radiation source as constraints, dividing the incoming wave direction of the target radiation source into networks to obtain an over-complete dictionary;

Step S3, based on the overcomplete dictionary, expanding the model in step S1 into a sparse reconstruction model of the target radiation source based on DifferenceCoarray;

Step S4: Use the reweighted l ₁ norm algorithm to solve the above model to obtain an image of the reconstructed target radiation source.

Furthermore, in the reweighted l ₁ norm algorithm, the calculation weight of the mth element of the k+1th solution vector is:

in, represents the partitioned network, k is the number of iterations, and ∈ represents a positive parameter for the robustness of the algorithm.

Furthermore, after obtaining the positioning result of the pseudo target radiation source point, the pseudo target radiation source point is completely attenuated or partially attenuated.

According to another aspect of the present invention, a pseudo target radiation source discrimination system based on array factor characteristics is provided, comprising:

An image reconstruction module of a target radiation source, used for reconstructing an image of the target radiation source;

A distance matrix construction module is used to traverse each pixel point in the reconstructed image, determine the maximum value and the position of each source point in the reconstructed image, regard the position of the maximum value as the position of the source point, and construct a distance matrix D from each source point to other source points;

A multi-antenna array factor model building module, used to build a multi-antenna array factor model diagram;

The position confirmation module of the suspected source point on the model diagram is used to determine the side lobe point with the largest relative intensity within a certain error range on the array factor model diagram, take the side lobe point as the suspected source point, determine the position information and distribution characteristics of the suspected source point in the model; and calculate the distance d _t from the suspected source point to the main lobe according to the position information and the position information of the main lobe in the model, where t is a positive integer;

A position confirmation module for a suspicious source point on a reconstructed image, for traversing each column of the distance matrix D for each d _t in the reconstructed image, and marking a source point whose distance is within the error range of d _t as a suspicious source point of the source point;

The pseudo target radiation source point determination module counts the distribution characteristics of the suspicious source points corresponding to each source point. When the distribution characteristics are consistent with the distribution characteristics in the array factor model, the suspicious source point is determined to be the pseudo target radiation source point corresponding to the source point, and the positioning result of the pseudo target radiation source point is obtained.

Furthermore, the multi-antenna array factor model construction module obtains a difference correlation array Difference Coarray of the original sparse array of the multi-antenna, and constructs a multi-antenna array factor model graph according to the distribution of the difference correlation array Difference Coarray.

Furthermore, the image reconstruction module of the target radiation source adopts a convex optimization algorithm to obtain a reconstructed image of the target radiation source.

Furthermore, it also includes an image processing module for completely or partially attenuating the pseudo target radiation source point.

In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

(1) The present invention establishes a multi-antenna array factor model diagram, finds a type of sidelobe point with the highest height on the array factor model diagram, takes it as a suspicious source point, determines the position information and characteristic distribution of this type of sidelobe point on the array factor model diagram, and compares the actual position information with the position information and distribution characteristics obtained in the model diagram on the reconstructed target radiation source image, and finally determines the specific position of the pseudo-target radiation source point on the reconstructed target radiation source image. Compared with the existing detection method, the present invention solves the detection and processing problems of false source points caused by the large fluctuation range of the interference source intensity, reduces the false alarm probability of radiation interference source detection in discrete multi-target images, improves the observability of the inversion image, and has high application value.

(2) Furthermore, when reconstructing the image of the target radiation source, the present invention adopts a reweighted l ₁ norm algorithm, which can enhance the signal strength of the radiation source and suppress background noise, further improving the detection and positioning accuracy. It can be applied to the field of signal detection to solve the problem of insufficient resolution of neighboring interference sources and the inability to detect and locate position-aggregated sources.

In summary, the present invention improves the robustness of the target radiation source detection method, screens out pseudo target radiation sources in the case of a mixture of strong and weak sources, improves the detection accuracy, and has high application value.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of a pseudo target radiation source identification method based on array factor characteristics provided by the present invention.

FIG. 2 is a structure of an original sparse array in an embodiment of the present invention.

FIG. 3 is a schematic diagram of a Difference Coarray of an original sparse array according to an embodiment of the present invention.

FIG. 4 is a schematic diagram of the distribution of element values of a redundant marking matrix according to an embodiment of the present invention.

FIG. 5 is an image of a target radiation source reconstructed in an embodiment of the present invention.

FIG. 6 is a diagram of an array factor model in an embodiment of the present invention.

FIG. 7 is a schematic diagram of the position of the highest side lobe in an embodiment of the present invention.

FIG8 is a diagram showing the detection and positioning results of the actual target radiation source after image feature matching processing in an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

As shown in FIG1 , the pseudo target radiation source identification method based on array factor characteristics provided by the present invention includes:

reconstructing an image of the target radiation source;

Traverse each pixel point in the reconstructed target radiation source image, find the maximum value and position of each source point in the image, regard the position of the maximum value as the position of the source point (x _i , y _i ), and construct the distance matrix D from each source point to other source points:

Wherein, _Dij represents the coordinate distance between the i-th and j-th source points, i=1,2,3..., j=1,2,3...

Establish an antenna array factor model for a multi-sensor system: obtain the difference correlation array Difference Coarray of the original sparse array of multiple antennas, and construct an antenna array factor model diagram according to the distribution of the difference correlation array Difference Coarray;

Determine the position information of the suspected source point on the array factor model diagram: find the sidelobe point with the highest height on the array factor model diagram, which is also the sidelobe point with the largest relative intensity of the radiation source on the array factor model diagram, take the sidelobe point as the suspected source point, and determine the position information and distribution characteristics of the suspected source point in the model; calculate the distance from the suspected source point to the main lobe based on the position information of the suspected source point and the position information of the main lobe; according to the required accuracy, there may be multiple sidelobe points with the largest relative intensity within a certain error range, and the distance from this type of suspected source point to the main lobe is recorded as d _t , where t is a positive integer.

Determine the suspicious source points on the image of the target radiation source: for each d _t , t is a positive integer, traverse each column in the distance matrix D, that is, the distance from each source point to other source points, and mark the source points whose distances are within the error range of d _t as suspicious source points of the source point, until each column in the distance matrix D is compared with t reference distances respectively, and obtain the positioning result of the suspicious source point corresponding to each source point; wherein, according to the required accuracy, d _t takes a value within a certain error range.

Determine the false target radiation source point: Count the distribution characteristics of the suspicious source points corresponding to each source point. When the distribution characteristics are consistent with the distribution characteristics in the array factor model, determine these suspicious source points that meet the consistent distribution characteristics as false target radiation source points, and obtain the positioning results of the false target radiation source points. In this embodiment, the distribution characteristics of the suspicious source points are 6-point hexagonal distribution.

Processing the pseudo target radiation source points: Calculate the number n of pseudo target radiation source points of each source point separately, and according to the specific accuracy, completely attenuate the pseudo target radiation source points whose n is greater than a certain value, and partially attenuate the pseudo target radiation source points whose n is less than the value. In this embodiment, when n≥3, the pseudo source points of the source point are completely attenuated, and when n＜3, the pseudo source points of the source point are partially attenuated.

Specifically, in the present invention, the step of reconstructing the image of the target radiation source includes:

Step S1: using a multi-sensor system to obtain the covariance matrix of the original sparse array of multiple antennas, redundantly averaging and vectorizing the covariance matrix to construct a difference coarray signal receiving model of the original sparse array, as follows:

z′＝D′(C)q+E

Where D′(C) represents the Difference Coarray flow type, the dimension is _Nu ×K, _Nu represents the number of array antennas, K represents the number of target radiation sources, C represents the set of cosine functions of the incoming waves of K target radiation sources, and E represents the measured noise matrix. in, and e′ ₁ represent the autocorrelation output power and the corresponding unit vector after redundant averaging of the covariance matrix, respectively, and q represents the position vector of the target radiation source.

Step S2: construct an overcomplete dictionary based on the difference correlation array Difference Coarray of the original sparse array and the spatial sparse characteristics of the target radiation source, and divide the incoming wave direction of the target radiation source into a network M _g ×M _g , where M _g >>K, and the overcomplete dictionary is D°(C°), Represents the divided grid, and expands the signal receiving model in step S1 into a sparse reconstruction model of the target radiation source based on Difference Coarray:

z′＝D°(C°)q+E

Step S3: Use the reweighted l ₁ norm algorithm to estimate the angle of arrival and solve the above model to obtain the image of the reconstructed target radiation source.

Specifically, by convexly relaxing the above sparse reconstruction problem, we can obtain:

In the formula, δ represents the regularization parameter.

Using the reweighted _l1 norm model to replace the above:

In the formula, is a weight coefficient vector with dimension M _g ×M _g .

Among them, the calculation weight of the mth element of the k+1th solution vector is:

In the formula, It represents the mth element of the kth reweighted solution vector, k is the number of iterations, and ∈ represents a positive parameter of the robustness of the algorithm.

The fast convergence threshold convergence algorithm and neighborhood weighting strategy are used to solve the replaced model; specifically:

Transform the above replaced model into an unconstrained solution model:

in, and W ^(k) represent the solution and weight matrix of the kth solution process, and μ represents the regularization parameter.

Using soft threshold convergence to solve the above unconstrained condition solution model:

in, represents the p+1th iteration solution vector in the kth reweighting process. ^{w (k)} = diag(W ^(k) ) represents the diagonalized weight matrix, and the parameter α is determined by the Lipschitz constant L, that is, α = 1/L. Represents a vector shrinkage operator that performs element-wise soft thresholding.

The image of the target radiation source is reconstructed by using the obtained solution, that is, the position vector q of the target radiation source.

In the present invention, by using the weighting coefficient W to weight the position vector q, it is possible to ensure that the strength of the signal source is enhanced and the background noise is suppressed.

In the present invention, the antenna array factor model is determined according to the antenna array arrangement of the multi-sensor system:

Where C(u,v) is a part of the two-dimensional comb function in the rectangular window, (l,m) is the cosine coordinate, and is the position vector of the rth and sth antennas. AF represents the array factor, and λ represents the wavelength of the radiation source.

In this embodiment, 400 points are uniformly selected on (-0.2, 0.2) for l and m, respectively, and the results are calculated by plotting, thereby obtaining an antenna array factor model diagram.

The present invention firstly uses the Difference Coarray of the original sparse array of the multi-sensor system to obtain a virtual array with a larger scale, constructs the covariance matrix of the original sparse array, and constructs the Difference Coarray signal receiving model by redundantly averaging and vectorizing the current covariance matrix; an over-complete dictionary is constructed according to the DifferenceCoarray of the original sparse array and the spatial sparse characteristics of the target radiation source, and a sparse reconstruction model of the target radiation source based on the virtual array is established; the sparse reconstruction model of the target radiation source based on the over-complete dictionary adopts a reweighted l ₁ norm algorithm to estimate the arrival angle, detects and locates the target radiation source, and obtains the image of the reconstructed target radiation source.

When the real source point is strong and there are weaker source points in the image, the decision threshold needs to be lowered, and false source points that do not actually exist will appear at the sidelobe position of the strong source. Based on the principle that the measured brightness temperature can be regarded as the convolution of the brightness temperature of the real target and the array factor, the antenna array factor model of the multi-sensor system is established.

In this embodiment, after obtaining the image of the reconstructed target radiation source, a type of false source points generated at the sidelobe position due to the excessive intensity of the real source points in the inversion image (i.e., the image of the reconstructed target radiation source) is found according to the array factor characteristics of the microwave radiometer antenna, and then this type of source points are processed on the image of the reconstructed target radiation source.

In this embodiment, the "Y"-shaped array carried by the European Space Agency SMOS (Soil Moisture and Ocean Salinity) satellite is selected as the original sparse array, the number of array elements is n′=69, and its array structure and Difference Coarray (virtual array) are shown in Figures 2 and 3.

According to the structural redundancy characteristics of Difference Coarray, a redundant marking matrix is determined to mark the redundant sampling data in the current covariance matrix. The matrix dimension is 931×931, and its element distribution is shown in FIG4 .

The covariance matrix after redundant averaging is vectorized to obtain a virtual array signal receiving model, namely, a Difference Coarray signal receiving model.

An over-complete dictionary is constructed according to the virtual signal receiving model and the spatial sparse characteristics of the target radiation source, and the sparse reconstruction model of the target radiation source is obtained using the over-complete dictionary.

The concept of neighborhood reweighting is used to construct a reweighted l ₁ norm sparse reconstruction model of the target radiation source, and the fast iterative threshold convergence algorithm is used to solve it. The detection and positioning results are shown in Figure 5, where ζ and η are cosine coordinates.

The array factor model established using the “Y”-shaped array carried by the SMOS satellite is shown in FIG6 , where the position of the highest type of sidelobe is shown in FIG7 .

The suspicious false positive source points in the positioning result map are processed according to the position information, and the measured target radiation source detection and positioning results after processing are shown in Figure 8. It can be seen from the figure that after processing the suspicious source points, a type of false source points generated in the sidelobe position due to the excessive intensity of the real source points are obviously removed, with good detection and positioning effects, and the observability of the image is improved.

In other embodiments, other convex optimization algorithms may also be used to obtain the reconstructed image of the target radiation source.

In the embodiment provided by the present invention, the detection results obtained indicate that the above-mentioned pseudo-target radiation source discrimination and image processing method based on array factor characteristics has good spatial resolution and detection accuracy performance, and can solve the detection and positioning of target radiation sources with weaker intensity under the condition of large intensity dynamics (i.e., strong sources and weak sources exist at the same time), and has high robustness in actual target radiation source detection and positioning applications.

The present invention also provides a pseudo target radiation source identification system based on array factor characteristics, comprising:

An image reconstruction module of a target radiation source, used for reconstructing an image of the target radiation source;

The distance matrix construction module is used to traverse each pixel point in the reconstructed image, determine the maximum value and position of each source point in the reconstructed image, regard the position of the maximum value as the position of the source point, and construct the distance matrix D from each source point to other source points;

A multi-antenna array factor model building module, used to build a multi-antenna array factor model diagram;

The position confirmation module of the suspected source point on the model diagram is used to determine the side lobe point with the largest relative intensity within a certain error range on the array factor model diagram, take the side lobe point as the suspected source point, determine the position information and distribution characteristics of the suspected source point in the model; and calculate the distance d _t from the suspected source point to the main lobe according to the position information and the position information of the main lobe in the model, where t is a positive integer;

A position confirmation module of a suspicious source point on a reconstructed image is used to traverse each column of the distance matrix D for each d _t in the reconstructed image, and mark all source points whose distances are within the error range of d _t as suspicious source points of the source point;

The pseudo target radiation source point determination module counts the distribution characteristics of the suspicious source points corresponding to each source point. When the distribution characteristics are consistent with the distribution characteristics in the array factor model, the suspicious source point is determined to be the pseudo target radiation source point corresponding to the source point, and the positioning result of the pseudo target radiation source point is obtained.

Specifically, the multi-antenna array factor model building module obtains the difference correlation array Difference Coarray of the original sparse array of the multi-antenna, and builds the multi-antenna array factor model graph according to the distribution of the difference correlation array Difference Coarray.

Specifically, the image reconstruction module of the target radiation source adopts a convex optimization algorithm to obtain a reconstructed image of the target radiation source.

The steps of reconstructing the image of the target radiation source include:

Step S1, obtaining the covariance matrix of the original sparse array of multiple antennas, and constructing a difference coarray signal receiving model of the original sparse array by redundantly averaging and vectorizing the covariance matrix;

Step S2, using the difference correlation array Difference Coarray of the original sparse array and the spatial sparse characteristics of the target radiation source as constraints, dividing the incoming wave direction of the target radiation source into networks to obtain an over-complete dictionary;

Step S3, based on the overcomplete dictionary, the model in step S1 is expanded into a sparse reconstruction model of the target radiation source based on Difference Coarray;

Step S4: Use the reweighted l ₁ norm algorithm to solve the above model to obtain an image of the reconstructed target radiation source.

Specifically, it also includes an image processing module, which is used to completely attenuate or partially attenuate the pseudo target radiation source point.

It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A method for distinguishing a pseudo target radiation source based on array factor characteristics, characterized by comprising: reconstructing an image of the target radiation source; Traverse each pixel point in the reconstructed image, determine the maximum value and position of each source point in the reconstructed image, regard the position of the maximum value as the position of the source point, and construct a distance matrix D from each source point to other source points; Establish a multi-antenna array factor model diagram; On the array factor model diagram, determine the side lobe point with the largest relative intensity within a certain error range, take the side lobe point as a suspected source point, determine the position information and distribution characteristics of the suspected source point in the model; and calculate the distance d _t from the suspected source point to the main lobe according to the position information and the position information of the main lobe in the model, where t is a positive integer; On the reconstructed image, for each d _t , traverse each column in the distance matrix D, and mark the source point whose distance is within the error range of d _t as the suspicious source point of the source point; The distribution characteristics of the suspicious source points corresponding to each source point are counted. When the distribution characteristics are consistent with the distribution characteristics in the array factor model, the suspicious source point is determined to be the pseudo target radiation source point corresponding to the source point, and the positioning result of the pseudo target radiation source point is obtained. 2. The method according to claim 1 is characterized in that the multi-antenna array factor model diagram is constructed according to the distribution of the difference correlation array Difference Coarray by obtaining the difference correlation array Difference Coarray of the multi-antenna original sparse array. 3. The method according to claim 2 is characterized in that a convex optimization algorithm is used to obtain the reconstructed image of the target radiation source. 4. The method according to claim 3, wherein the step of reconstructing the image of the target radiation source comprises: Step S1, obtaining a covariance matrix of an original sparse array of multiple antennas, redundantly averaging and vectorizing the covariance matrix to construct a difference coarray signal receiving model of the original sparse array; Step S2, using the difference correlation array Difference Coarray of the original sparse array and the spatial sparse characteristics of the target radiation source as constraints, dividing the incoming wave direction of the target radiation source into networks to obtain an over-complete dictionary; Step S3, based on the overcomplete dictionary, expanding the model in step S1 into a sparse reconstruction model of the target radiation source based on Difference Coarray; Step S4: Use the reweighted l ₁ norm algorithm to solve the above model to obtain an image of the reconstructed target radiation source. 5. The method according to claim 4, characterized in that, in the reweighted l ₁ norm algorithm, the calculation weight of the mth element of the k+1th solution vector is: in, represents the partitioned network, k is the number of iterations, and ∈ represents a positive parameter for the robustness of the algorithm. 6. The method according to any one of claims 1-5 is characterized in that after obtaining the positioning result of the pseudo target radiation source point, it also includes completely attenuating or partially attenuating the pseudo target radiation source point. 7. A pseudo target radiation source identification system based on array factor characteristics, characterized by comprising: An image reconstruction module of a target radiation source, used for reconstructing an image of the target radiation source; A distance matrix construction module is used to traverse each pixel point in the reconstructed image, determine the maximum value and the position of each source point in the reconstructed image, regard the position of the maximum value as the position of the source point, and construct a distance matrix D from each source point to other source points; A multi-antenna array factor model building module, used to build a multi-antenna array factor model diagram; The position confirmation module of the suspected source point on the model diagram is used to determine the side lobe point with the largest relative intensity within a certain error range on the array factor model diagram, take the side lobe point as the suspected source point, determine the position information and distribution characteristics of the suspected source point in the model; and calculate the distance d _t from the suspected source point to the main lobe according to the position information and the position information of the main lobe in the model, where t is a positive integer; A position confirmation module for a suspicious source point on a reconstructed image, for traversing each column of the distance matrix D for each d _t in the reconstructed image, and marking a source point whose distance is within the error range of d _t as a suspicious source point of the source point; The pseudo target radiation source point determination module counts the distribution characteristics of the suspicious source points corresponding to each source point. When the distribution characteristics are consistent with the distribution characteristics in the array factor model, the suspicious source point is determined to be the pseudo target radiation source point corresponding to the source point, and the positioning result of the pseudo target radiation source point is obtained. 8. The system according to claim 7 is characterized in that the multi-antenna array factor model construction module obtains the difference correlation array Difference Coarray of the original sparse array of the multi-antenna, and constructs the multi-antenna array factor model diagram according to the distribution of the difference correlation array Difference Coarray. 9. The system according to claim 8, characterized in that the image reconstruction module of the target radiation source adopts a convex optimization algorithm to obtain the reconstructed image of the target radiation source. 10. The system according to any one of claims 7 to 9, further comprising an image processing module for completely or partially attenuating a pseudo target radiation source point.