Spaceborne synthetic aperture radar image simulation and image quality assessment method and device

Abstract

The present invention discloses a method and device for simulating and evaluating image quality of a spaceborne synthetic aperture radar image, the method comprising: obtaining target information and incident point source information; the target information comprises target shape structure information and target electromagnetic echo image; the target shape structure information comprises target three-dimensional size information, composition information, surface material information and shape information; the incident point source information comprises incident point source position information and incident point source waveform information; the target electromagnetic echo image is a target SAR image obtained at the incident point source position; the target shape structure information and the incident point source information are simulated to obtain a spaceborne synthetic aperture radar target simulation image; the spaceborne synthetic aperture radar target simulation image and the target electromagnetic echo image are evaluated to obtain image simulation quality evaluation result information. The present invention effectively solves the problem of difficulty in simulating synthetic aperture radar images, provides a method for generating a large amount of radar target data at a low cost, and designs a method for evaluating the accuracy of simulated images.

Description

Technical Field

[0001] The invention relates to the technical field of remote sensing imaging, and in particular to a method and a device for simulating and evaluating image quality of a spaceborne synthetic aperture radar. Background Art

[0002] Synthetic Aperture Radar (SAR) images can provide detailed ground mapping and images in all weather conditions and a penetrating field of view, which is critical for Earth observation applications compared to traditional optical sensors.

[0003] For SAR, due to its special imaging mechanism, slight changes in imaging parameters and scenes will cause large changes in the image. It is very difficult to obtain SAR image radar data under comprehensive parameters and scenes. Although the target radar echo data can be obtained through microwave darkroom simulation, and then SAR radar target images can be generated, the accuracy of this method is restricted by the accuracy of model making, and it cannot reflect the characteristics of radar target data in complex scenes. The generated samples are single and there are many constraints. In addition, there is currently no complete evaluation system to evaluate the simulation effect of SAR images.

[0004] Therefore, how to generate high-precision SAR image radar simulation data for various scenarios and parameters, and conduct accurate and comprehensive evaluation of it, is an urgent problem that needs to be solved. Summary of the invention

[0005] The purpose of the present invention is to provide a spaceborne synthetic aperture radar image simulation and image quality assessment method, which can be applied to electrically large-sized targets such as ships and vehicles, and can achieve full-incident-angle, omni-angle, and high-resolution image simulation, and can generate high-precision SAR image radar simulation data for various scenes and parameters, and perform accurate and comprehensive assessments thereof.

[0006] In a first aspect of an embodiment of the present invention, a method for simulating and evaluating image quality of a spaceborne synthetic aperture radar is disclosed, comprising:

[0007] S1, acquiring target information and incident point source information; the target information includes target shape structure information and target electromagnetic echo image; the target shape structure information includes target three-dimensional size information, composition information, surface material information and shape information; the incident point source information includes incident point source position information and incident point source waveform information; the target electromagnetic echo image is a target SAR image obtained at the incident point source position;

[0008] S2, simulating the target shape structure information and the incident point source information to obtain a simulated image of the spaceborne synthetic aperture radar target;

[0009] S3, evaluating and processing the spaceborne synthetic aperture radar target simulation image and the target electromagnetic echo image to obtain image simulation quality evaluation result information.

[0010] The target shape structure information and the incident point source information are simulated to obtain a spaceborne synthetic aperture radar target simulation image, including:

[0011] S21, modeling the target shape structure information to obtain target discrete grid information;

[0012] S22, constructing a virtual incident wave set using incident point source information and target discrete grid information;

[0013] S23, using the far-field scattering field calculation model, the virtual incident wave set and the target discrete grid information are calculated and processed to obtain a spaceborne synthetic aperture radar target simulation image.

[0014] The modeling process of the target shape structure information to obtain the target discrete grid information includes:

[0015] S211, using the acquired target shape and structure information, establishing a target three-dimensional model; the target three-dimensional model includes three-dimensional size information, structure distribution information, and medium information of the target;

[0016] S212, processing the target three-dimensional model to obtain target discrete grid information.

[0017] The processing of the target three-dimensional model to obtain target discrete grid information includes:

[0018] S2121, performing segmentation processing on the target three-dimensional model to obtain basic facet information of the target and calculation area grid information of the target;

[0019] S2122, using the target calculation area grid information, perform grid generation processing on the basic surface element to obtain the target discrete grid information.

[0020] The method of constructing a virtual incident wave set by using the incident point source information and the target discrete grid information includes:

[0021] S221, using the incident point source position information, connect the coordinates of each grid center point in the target discrete grid information to obtain the corresponding virtual incident wave;

[0022] S222, construct a virtual incident wave set using all virtual incident waves.

[0023] The expression of the far-zone scattering field calculation model is:

[0024]

[0025] in, represents the virtual incident wave The far-field electromagnetic scattering intensity information at time t, t represents the time variable, N represents the integrated electromagnetic constant vector, τ represents the incident pulse width in the incident point source waveform information, Q1 and Q2 represent the first integral function and the second integral function respectively, and the expression is:

[0026]

[0027] Among them, [L1, L2] are the upper and lower limits of the integration range of the x-axis; using the virtual incident wave and the two virtual incident waves closest to it, the three intersection points of the three virtual incident waves and the grid in the target discrete grid information are obtained, and the two connecting lines between the three intersection points are used to determine the corresponding two straight line equations, whose expressions are y=q1x+w1 and y=q2x+w2, respectively, and q1, w1, q2, and w2 are the coefficients of the two straight line equations respectively; using the intersection point of the virtual incident wave and the grid, and connecting with the integral point on the grid, the first straight line equation is obtained, and the expression of the first straight line equation is y=qx+w, q and w represent the coefficients of the first straight line equation; τ1 represents the time taken by the virtual incident wave from the incident point source through the intersection point of the virtual incident wave and the grid to reach the integral point of the grid, and t0 is the incident pulse time width control coefficient of the incident point source; the first integral function and the second integral function both integrate the variable t; the expression of the integral electromagnetic constant vector N is:

[0028]

[0029] in, Represents the unit external normal vector of the basic surface element corresponding to the mesh, represents the magnetic field unit vector in the virtual incident wave, Z0 is the free space wave impedance, c is the vacuum wave velocity, Represents the virtual incident wave corresponding vector The modulus of Modulo value, we get is the virtual incident wave The far-field electromagnetic scattering intensity information of the corresponding grid.

[0030] The step of evaluating and processing the spaceborne synthetic aperture radar target simulation image and the target electromagnetic echo image to obtain image simulation quality evaluation result information includes:

[0031] The spaceborne synthetic aperture radar target simulation image is represented as a matrix A, and the target electromagnetic echo image is represented as a matrix B;

[0032] The absolute value of the difference between matrix A and matrix B is calculated and expressed as matrix Q, Q = |AB|;

[0033] The matrix Q is subjected to deviation quantization calculation processing to obtain a deviation matrix H; the calculation expression of the deviation quantization calculation processing is:

[0034]

[0035] In the formula, h ki represents the element that deviates from the kth row and ith column of the matrix H, q ki represents the i-th element of the k-th row of the matrix q, and the dimension of the matrix H is m×n;

[0036] Decomposing the matrix A to obtain a feature matrix Y;

[0037] The decomposition process is calculated as follows:

[0038] A=UYV,

[0039] Among them, U is the left decomposition matrix, Y is the characteristic matrix, V is the right decomposition matrix, and both U and V are orthogonal matrices;

[0040] Extract the diagonal elements of the feature matrix to obtain a feature vector; the feature vector is represented by r;

[0041] Performing time-frequency transformation on each row vector of the matrix B to obtain a corresponding time-frequency vector; constructing a time-frequency matrix S using all the time-frequency vectors;

[0042] The time-frequency matrix S and the eigenvector r are subjected to correlation calculation processing to obtain a correlation matrix V; the calculation expression of the correlation calculation processing is:

[0043]

[0044] Among them, v ki represents the element in the kth row and ith column of the incidence matrix. The dimension of the incidence matrix V is m×n, r k represents the kth element of the eigenvector r, s ki is the element in the kth row and ith column of the time-frequency matrix S, and ρ is the correlation calculation factor;

[0045] Using the deviation matrix H, logarithmic weighting is performed on each element of the correlation matrix V to obtain image simulation quality assessment result information;

[0046] The calculation expression of the logarithmic weighting process is:

[0047]

[0048] Wherein, H0 is the information value of the image simulation quality assessment result.

[0049] In a second aspect of an embodiment of the present invention, a device for simulating and evaluating image quality of a spaceborne synthetic aperture radar is disclosed, the device comprising:

[0050] A memory storing executable program code;

[0051] a processor coupled to the memory;

[0052] The processor calls the executable program code stored in the memory to execute the spaceborne synthetic aperture radar image simulation and image quality assessment method.

[0053] In a third aspect of an embodiment of the present invention, a computer storable medium is disclosed, wherein the computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the space-borne synthetic aperture radar image simulation and image quality assessment method.

[0054] According to a fourth aspect of the embodiments of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the spaceborne synthetic aperture radar image simulation and image quality assessment method.

[0055] The beneficial effects of the present invention are:

[0056] The present invention is aimed at the SAR image simulation field of remote sensing imaging technology, and establishes a relatively complete SAR image simulation method, which is conducive to overcoming the high cost and long period of obtaining SAR images with different parameters by real SAR systems, as well as the shortcomings of slow speed and poor generalization ability of traditional ray tracing calculation methods. A method for evaluating the accuracy of simulated images is also provided.

[0057] The present invention can realize multi-incident angle, multi-azimuth, high-resolution SAR image simulation. Compared with the traditional simulation method, the present invention has a fast simulation calculation data speed, is consistent with the physical process of actual imaging, and has high simulation image accuracy and good simulation process flexibility, which can well meet the current demand for data volume and target identification and interpretation requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0059] Figure 1 is a flow chart of the method of the present invention.

[0060] The present invention utilizes the radar cross section (RCS) of the target calculated by magnetic scattering to generate echo data for imaging, which is not only consistent with the actual physical process, but also has high simulation image accuracy and good simulation process flexibility. It also provides the change trend and formation mechanism of the target electromagnetic scattering characteristics, thereby increasing the interpretability of the imaging mechanism.

[0061] Embodiment 1

[0062] In a first aspect of an embodiment of the present invention, a method for simulating and evaluating image quality of a spaceborne synthetic aperture radar is disclosed, comprising:

[0063] S1, acquiring target information and incident point source information; the target information includes target shape structure information and target electromagnetic echo image; the target shape structure information includes target three-dimensional size information, composition information, surface material information and shape information; the incident point source information includes incident point source position information and incident point source waveform information; the target electromagnetic echo image is a target SAR image obtained at the incident point source position.

[0064] S2, simulating the target shape structure information and the incident point source information to obtain a simulated image of the spaceborne synthetic aperture radar target;

[0065] S3, evaluating and processing the spaceborne synthetic aperture radar target simulation image and the target electromagnetic echo image to obtain image simulation quality evaluation result information.

[0066] The target shape structure information and the incident point source information are simulated to obtain a spaceborne synthetic aperture radar target simulation image, including:

[0067] S21, modeling the target shape structure information to obtain target discrete grid information.

[0068] S22, constructing a virtual incident wave set using incident point source information and target discrete grid information;

[0069] S23, using a far-field scattering field calculation model, calculating and processing a virtual incident wave set and target discrete grid information to obtain a spaceborne synthetic aperture radar target simulation image;

[0070] The modeling process of the target shape structure information to obtain the target discrete grid information includes:

[0071] S211, using the acquired target shape and structure information, establish a target three-dimensional model; the target three-dimensional model includes three-dimensional size information, structure distribution information, medium information, etc. of the target.

[0072] S212, processing the target three-dimensional model to obtain target discrete grid information.

[0073] The processing of the target three-dimensional model to obtain target discrete grid information includes:

[0074] S2121, performing segmentation processing on the target three-dimensional model to obtain basic facet information of the target and calculation area grid information of the target;

[0075] S2122, using the target calculation area grid information, perform grid generation processing on the basic surface element to obtain the target discrete grid information.

[0076] The method of constructing a virtual incident wave set by using the incident point source information and the target discrete grid information includes:

[0077] S221, using the incident point source position information, connect the coordinates of each grid center point in the target discrete grid information to obtain the corresponding virtual incident wave;

[0078] S222, construct a virtual incident wave set using all virtual incident waves.

[0079] The expression of the far-zone scattering field calculation model is:

[0080]

[0081] in, represents the virtual incident wave The far-field electromagnetic scattering intensity information at time t, t represents the time variable, N represents the integrated electromagnetic constant vector, τ represents the incident pulse width in the incident point source waveform information, Q1 and Q2 represent the first integral function and the second integral function respectively, and the expression is:

[0082]

[0083] Among them, [L1, L2] are the upper and lower limits of the integration range of the x-axis; using the virtual incident wave and the two virtual incident waves closest to it, the three intersection points of the three virtual incident waves and the grid in the target discrete grid information are obtained, and the two connecting lines between the three intersection points are used to determine the corresponding two straight line equations, whose expressions are y=q1x+w1 and y=q2x+w2, respectively, and q1, w1, q2, and w2 are the coefficients of the two straight line equations respectively; using the intersection point of the virtual incident wave and the grid, and connecting with the integral point on the grid, the first straight line equation is obtained, and the expression of the first straight line equation is y=qx+w, q and w represent the coefficients of the first straight line equation; τ1 represents the time taken by the virtual incident wave from the incident point source through the intersection point of the virtual incident wave and the grid to reach the integral point of the grid, and t0 is the incident pulse time width control coefficient of the incident point source; the first integral function and the second integral function both integrate the variable t; the expression of the integral electromagnetic constant vector N is:

[0084]

[0085] in, Represents the unit external normal vector of the basic surface element corresponding to the mesh, represents the magnetic field unit vector in the virtual incident wave, Z0 is the free space wave impedance, c is the vacuum wave velocity, Represents the virtual incident wave corresponding vector The modulus of Modulo value, we get is the virtual incident wave The far-field electromagnetic scattering intensity information of the corresponding grid.

[0086] Wherein, [L1, L2] also represents the value range of the vertex on the X-axis when the local discrete grid is projected onto the XOY plane of the target rectangular coordinate system, and is usually set to t0 = 0.8τ. The intersection of the virtual incident wave and the grid can be determined by the method of S212211 to S212218.

[0087] The step of evaluating and processing the spaceborne synthetic aperture radar target simulation image and the target electromagnetic echo image to obtain image simulation quality evaluation result information includes:

[0088] The spaceborne synthetic aperture radar target simulation image is represented as a matrix A, and the target electromagnetic echo image is represented as a matrix B;

[0089] The absolute value of the difference between matrix A and matrix B is calculated and expressed as matrix Q, Q = |AB|;

[0090] The matrix Q is subjected to deviation quantization calculation processing to obtain a deviation matrix H; the calculation expression of the deviation quantization calculation processing is:

[0091]

[0092] In the formula, h ki represents the element that deviates from the kth row and ith column of the matrix H, q ki represents the i-th element of the k-th row of the matrix q, and the dimension of the matrix H is m×n;

[0093] Decomposing the matrix A to obtain a feature matrix Y;

[0094] The decomposition process is calculated as follows:

[0095] A=UYV,

[0096] Among them, U is the left decomposition matrix, Y is the characteristic matrix, V is the right decomposition matrix, and both U and V are orthogonal matrices;

[0097] Extract the diagonal elements of the feature matrix to obtain a feature vector; the feature vector is represented by r;

[0098] Performing time-frequency transformation on each row vector of the matrix B to obtain a corresponding time-frequency vector; constructing a time-frequency matrix S using all the time-frequency vectors;

[0099] The time-frequency matrix S and the eigenvector r are subjected to correlation calculation processing to obtain a correlation matrix V; the calculation expression of the correlation calculation processing is:

[0100]

[0101] Among them, v ki represents the element in the kth row and ith column of the incidence matrix. The dimension of the incidence matrix V is m×n, r k represents the kth element of the eigenvector r, s ki is the element in the kth row and ith column of the time-frequency matrix S, and ρ is the correlation calculation factor;

[0102] Using the deviation matrix H, logarithmic weighting is performed on each element of the correlation matrix V to obtain image simulation quality assessment result information;

[0103] The calculation expression of the logarithmic weighting process is:

[0104]

[0105] Among them, H0 is the image simulation quality evaluation result information value, and the smaller its value is, the better the image simulation quality is.

[0106] The decomposition process may adopt an eigenvalue decomposition algorithm or a singular value decomposition algorithm.

[0107] The electromagnetic scattering intensity calculation model uses the basic surface element intersecting with the virtual incident wave as the integral area for integral calculation. The two lines between the three intersections can be selected from any two lines among all the lines of the three intersections. The incident point source waveform can be a Gaussian pulse.

[0108] The method uses a far-field scattering field calculation model to calculate and process a virtual incident wave set and target discrete grid information to obtain a space-borne synthetic aperture radar target simulation image, wherein the far-field scattering field calculation model is used to calculate the far-field electromagnetic scattering intensity information of each virtual incident wave, and the far-field electromagnetic scattering intensity information of all virtual incident waves of the virtual incident wave set is integrated to obtain a space-borne synthetic aperture radar target simulation image.

[0109] The S2121 includes:

[0110] S21211, establish a three-dimensional rectangular coordinate system with the center point of the target three-dimensional model as the origin; use basic surface elements to segment the target three-dimensional model to obtain basic surface element information of the target; the basic surface element information includes the number of basic surface elements, the normal vector information of each basic surface element, the vertex coordinate information of each basic surface element, and the medium information of each basic surface element; the medium information includes dielectric constant, electrical conductivity, magnetic permeability and magnetic permeability.

[0111] The step S21211 may be to use 3ds Max software to implement segmentation processing on the target 3D model. From 3ds Max software, export the basic facet metadata file of the target 3D model in stl format; set the number of processes n, use the main process of the computing server to read the stl file, and determine the number of basic facets N by counting the number of times "endfacet" appears in the stl file. tri , get the normal vector information and vertex coordinate information of each basic surface element from the statistical stl file.

[0112] The step S21211 may also be implemented using HyperMesh software.

[0113] S21212, using vertex coordinate information of the basic face element to determine the region range of the basic face element, and using the region ranges of all the basic face elements to determine the region range of the target three-dimensional model;

[0114] S21213, using the preset highest calculation frequency and the target fine structure size L min , determine the grid discrete size value;

[0115] The S21213 includes:

[0116] S212131, based on the highest calculation frequency a max and the conductivity of air σ, the minimum wavelength b of the detection wave in air is calculated min , its calculation expression is:

[0117]

[0118] Wherein, μ0 is the magnetic permeability coefficient in vacuum; the detection wave may be an electromagnetic wave.

[0119] S212132, select L min and λ min The minimum value among / 20 is the grid discrete size value;

[0120] S21214, determining the target calculation area grid information according to the grid discrete size value and the area range of the target three-dimensional model;

[0121] The S21214 includes:

[0122] S212141, using the area range of the target three-dimensional model, divided by the grid discrete size value, to obtain a calculation area of ​​the target three-dimensional model;

[0123] S212142, perform unit discretization processing on the calculation area of ​​the target three-dimensional model to obtain the calculation area grid information of the target; the calculation area grid includes a plurality of area grid lines; the calculation area grid information includes equations of the area grid lines and range information of the calculation area grid; the range information of the calculation area grid is expressed in the three-dimensional rectangular coordinate system as (Xmin: 1: Xmax, Ymin: 1: Ymax, Zmin: 1: Zmax); the area grid line is a combination of two variables selected from the three variables of the X-axis, Y-axis and Z-axis of the three-dimensional rectangular coordinate system, and the two variables are selected from the area range (Xmin: 1: Xmax, Ymin: 1: Ymax, Zmin: 1: Zmax) according to the coordinate axes where they are located, and the resulting straight line is constructed; the value set of the endpoints of the area grid line on the three axes constitutes the range information of the calculation area grid.

[0124] The S212, after completing S21214 and before performing the grid generation process, further includes:

[0125] S21215, dividing the calculation area of ​​the target three-dimensional model to obtain three-dimensional virtual topological structure information of the target characteristics; the three-dimensional virtual topological structure information includes range information of the sub-areas obtained by dividing; according to the three-dimensional virtual topological structure information, each sub-area is allocated to a corresponding calculation process of the calculation server to achieve optimal communication and load balancing;

[0126] S21216, according to the sub-region corresponding to each calculation process and the vertex coordinate information of the basic surface element, the basic surface elements included in each calculation process are statistically obtained.

[0127] The S21216 includes: calculating the center coordinate information of the basic face element according to the vertex coordinate information of the basic face element; judging the calculation process to which the basic face element belongs according to the center coordinate information of the basic face element and the corresponding sub-area of ​​each calculation process, and obtaining the basic face element information contained in each calculation process by counting, which is specifically the position and quantity.

[0128] The S2122 includes:

[0129] S21221, using the target calculation area grid information, performing intersection judgment processing on the basic face elements to obtain the intersection points of all basic face elements and the regional grid lines of the calculation area grid;

[0130] S21222, using the intersection of the basic facet and the regional grid line of the calculation regional grid, numbering each regional grid line to obtain local discrete grid information of each regional grid line; the local discrete grid information includes grid position information and grid center point coordinate information;

[0131] Specifically, step S2122 may be, in each calculation process, performing mesh generation processing on the included basic surface elements to obtain a local discrete mesh of the target;

[0132] S21223, integrating the local discrete grid information of all regional grid lines to obtain target discrete grid information.

[0133] The S21223 may be to use the local discrete grids of all regional grid lines as the target discrete grid information.

[0134] The S21223 may be that each computing process sends the local discrete grid of the target to the main process of the computing server, and the main process integrates all collected grid data to achieve rapid generation of the target discrete grid.

[0135] The method of using the target calculation area grid information to perform intersection judgment processing on the basic face elements to obtain the intersection points of all basic face elements and the regional grid lines of the calculation area grid includes:

[0136] S212211, determining the absolute values ​​of the components of the normal vector of the basic face element on the three coordinate axes of the three-dimensional rectangular coordinate system, and determining the coordinate axes corresponding to the two components with the smallest absolute values; constructing a projection plane using the coordinate axes corresponding to the two components; and projecting the basic face element onto the projection plane of the three-dimensional rectangular coordinate system to obtain a projection basic shape;

[0137] S212212, constructing a projected rectangular area using the vertex coordinates of the projected basic shape within the value range of the coordinate axes corresponding to the two components;

[0138] S212213, judging whether the projected rectangular area contains regional grid lines, and obtaining a judgment result;

[0139] If the result of the judgment is yes, execute S212214; if the result of the judgment is no, execute S212218;

[0140] The projection rectangular area contains regional grid lines, which means that the regional grid lines pass through the projection rectangular area.

[0141] S212214, determining the intersection of the regional grid lines contained in the projected rectangular area with the projection plane;

[0142] S212215, using the intersection point as a vector starting point and using the vertices of the projection basic shape as vector end points respectively, to construct a projection vector set; the projection vector set includes a plurality of projection vectors;

[0143] S212216, constructing a projection vector combination set by using all combinations of two projection vectors in the projection vector set;

[0144] S212217, for each projection vector combination in the projection vector combination set, respectively calculate the corresponding vector cross product value to obtain a vector cross product value set;

[0145] Determine whether the vector cross product value set meets the cross discrimination condition, and obtain a cross discrimination result;

[0146] When the intersection determination result satisfies the intersection determination condition, determining that the intersection point of the regional grid line and the basic bin is inside the basic bin;

[0147] When the intersection determination result is that the intersection determination condition is not satisfied, determining that the intersection point between the regional grid line and the basic bin is not inside the basic bin;

[0148] The crossover discrimination condition is that each vector cross product value in the vector cross product value set is a positive number, or each vector cross product value in the vector cross product value set is a negative number, or one vector cross product value in the vector cross product value set is 0, or two vector cross product values ​​in the vector cross product value set are 0.

[0149] S212218, if the basic bin is not the last basic bin in the preset sequence, execute S212211 for the next basic bin in the preset sequence;

[0150] If the basic surface element is the last basic surface element according to a preset sequence, the intersection points of all basic surface elements and the regional grid lines of the calculation regional grid are obtained.

[0151] The preset order may be the spatial distribution order of the basic face elements in the target three-dimensional model, for example, in an order from top to bottom and from left to right.

[0152] The method of using the intersection of the basic facet and the regional grid line of the calculation regional grid to perform numbering processing on each regional grid line to obtain a local discrete grid of each regional grid line includes:

[0153] S212221, for each regional grid line, numbering the intersections in order from small to large absolute values ​​of the coordinates of the intersections contained therein to obtain numbered intersections;

[0154] S212222, construct a directed line segment using all two numbered intersections with adjacent numbers; in the directed line segment, the intersection with a smaller number among the two numbered intersections is used as the starting point of the directed line segment, and the intersection with a larger number among the two numbered intersections is used as the end point of the directed line segment;

[0155] S212223, determining all directed line segments whose starting points are odd-numbered and whose end points are even-numbered as local discrete grids of the regional grid lines.

[0156] The S21222 may be: for each regional grid line, in the order of the absolute values ​​of the coordinates of the intersections it contains from small to large, n numbered intersections are obtained, 1, 2, 3, 4, ..., n, the directed line segment between intersection 1 and intersection 2 is Ⅰ, the directed line segment between intersection 2 and intersection 3 is Ⅱ, the directed line segment between intersection 3 and intersection 4 is Ⅲ, and so on. The directed line segments numbered Ⅰ, Ⅲ, Ⅴ, ... to the last odd number less than or equal to n are the local discrete grids of the regional grid line; for the local discrete grids, the corresponding medium parameters are marked.

[0157] The acquisition of the target shape and structure information may be achieved through an Internet database, data collection on the target, or a method of generating simulation data.

[0158] The target data collection may be achieved by radar measurement, laser measurement or acoustic measurement.

[0159] The method for generating the simulation data may be to use mechanical simulation software.

[0160] The three-dimensional size information of the target may be the length, width and height; the component information includes the power part, load part, peripheral part and fine structure part; the surface material information includes the type, parameters and thickness of the surface material, etc. The shape information includes the specific three-dimensional size information of each component.

[0161] The fine structure may be a mast, an antenna, a sensor, etc.

[0162] The step S2 can be implemented by three-dimensional mechanical simulation software; the three-dimensional mechanical simulation software can be 3ds Max software.

[0163] The target electromagnetic echo image is obtained by actual measurement using SAR radar.

[0164] The basic face element may be a triangle face element, a rhombus face element, a regular pentagon face element, etc.

[0165] In a second aspect of an embodiment of the present invention, a device for simulating and evaluating image quality of a spaceborne synthetic aperture radar is disclosed, the device comprising:

[0166] A memory storing executable program code;

[0167] a processor coupled to the memory;

[0168] The processor calls the executable program code stored in the memory to execute the spaceborne synthetic aperture radar image simulation and image quality assessment method.

[0169] In a third aspect of an embodiment of the present invention, a computer storable medium is disclosed, wherein the computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the space-borne synthetic aperture radar image simulation and image quality assessment method.

[0170] According to a fourth aspect of the embodiments of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the spaceborne synthetic aperture radar image simulation and image quality assessment method.

[0171] Embodiment 2 Specific implementation method one:

[0173] The electromagnetic simulation radar image data and image quality assessment method described in this embodiment are implemented as follows:

[0174] Step 1: Processing the 3D model of the simulation target: determine a target for which SAR image simulation is required, and use 3D modeling software to build a specific 3D model for it; divide the established 3D model into triangular meshes, and the division size should not exceed one-sixth of the wavelength in principle. The simulation wavelength is calculated from the simulation frequency;

[0175] The relationship between wavelength, frequency and speed of light is as shown in formula (1):

[0176]

[0177] Among them, λ is the wavelength, c is the speed of light, f is the center frequency, and the grid division should follow the following basic principles:

[0178] The model's mesh size is usually between 1 / 8 and 1 / 10 wavelengths. However, in the actual calculation process, if the model's electrical size is too large, a 1 / 4 wavelength mesh can be used to achieve better calculation accuracy. The mesh size will affect the convergence speed of the solution. A too coarse mesh size will cause the convergence speed to decrease or even diverge. For media with larger dielectric constants or larger losses, the mesh size can be finer. MOM, MLFMM and PO (physical optics) all use triangular units for meshing. The mesh size must be less than 1 / 3 wavelength, and 1 / 8 to 1 / 10 wavelength is usually recommended.

[0179] Step 2: Set electromagnetic simulation parameters: import the model divided into triangular mesh into electromagnetic simulation software; set the azimuth resolution, distance resolution, simulation range, simulation frequency, angle and sampling points, and select a specific and appropriate solution method; run the electromagnetic simulation software to solve the three-dimensional model;

[0180] The calculation formula of azimuth resolution R is as follows:

[0181]

[0182] Among them, Δθ is the angle of rotation of the target during the coherent accumulation time;

[0183] Distance resolution R res As shown in formula (3):

[0184]

[0185] Where, B is the simulation bandwidth;

[0186] Step 3: Simulation data imaging processing: Obtain the data after electromagnetic simulation calculation, calculate the target radar cross-section data matrix; set the corresponding simulation parameters according to the spaceborne radar, and use the imaging algorithm to obtain the target echo data; add coherent speckle noise to obtain the target simulation SAR image;

[0187] Step 4: Calculate the accuracy of the simulated image: Calculate the histogram similarity, hash algorithm similarity, structural similarity, and cosine similarity between the simulated image and the real image. Calculate the average simulated image accuracy of the values ​​calculated by the four methods and normalize them. If it is greater than or equal to 0.8, the two images are considered similar; if it is less than 0.8, the two images are considered dissimilar.

[0188] The similarity measure of the histogram Euclidean distance is formula (4):

[0189]

[0190] Where 0≤s i ≤1 and 0≤mi≤1 are normalized values. After normalization, the similarity measure of the image histogram is formula (5):

[0191]

[0192] When calculating the image similarity using the hash algorithm, the discrete cosine transform (DCT) calculation formula is formula (6):

[0193]

[0194] Among them, f(i,j) is the original signal, F(u,v) is the coefficient after DCT transformation, N is the number of points of the original signal, c(u)c(v) is the compensation coefficient, so that the DCT transformation matrix is ​​an orthogonal matrix. Select the difference value hash algorithm, reduce the image size, keep only the structure brightness and darkness and convert it into a grayscale image; calculate DCT; reduce DCT; calculate the average value; further reduce DCT; calculate the hash value; compare the simulated SAR image with the real SAR image. If the number of identical data bits exceeds 59 (the ratio is greater than 0.92), it means that the two images are similar; if it is less than 54 (the ratio is less than 0.84), it means that these two images are not similar.

[0195] The formula for calculating structural similarity is formula (7):

[0196]

[0197] Among them, x, y are the two input images, μ x and μ y are the average values ​​of the two images x and y, σ x and σ yare the standard deviations of the two images x and y respectively, c1 and c2 are constants.

[0198] The cosine similarity calculation formula is formula (8):

[0199]

[0200] Among them, A and B are n-dimensional vectors.

[0201] Finally, the accuracy formula of the simulation image is obtained as formula (9):

[0202]

[0203] Specific implementation method 2: The difference between this implementation method and specific implementation method 1 is that in step 4, the hash algorithm can select the average hash algorithm, scale the image and grayscale it; calculate the average value and binarize each pixel according to the average value; compare the grayscale of each pixel with the average value; if it is greater than or equal to the average value, it is recorded as 1; if it is less than the average value, it is recorded as 0, and the calculation results are combined to obtain the hash value of the image. Calculate the hash values ​​of the simulated SAR image and the real SAR image respectively, and calculate the Hamming distance. The larger the Hamming distance, the more inconsistent the images are. Conversely, the smaller the Hamming distance, the more similar the images are. When the distance is 0, it means they are exactly the same.

[0204] The other steps are the same as those in the first embodiment.

[0205] Specific implementation method three: The difference between this implementation method and specific implementation method one is that in step four, the hash algorithm can select the perceptual hash algorithm, scale the image and grayscale it; calculate DCT, but only take the 8*8 matrix in the upper left corner, which presents the lowest frequency in the image; calculate the average value of the DCT coefficient; compare the grayscale of each pixel with the average value of the DCT coefficient, if it is greater than or equal to the average value, it is recorded as 1, and if it is less than the average value, it is recorded as 0. The results are combined to form a 64-bit binary integer, which is the hash value of the image. Calculate the hash values ​​of the simulated SAR image and the real SAR image respectively, calculate the Hamming distance and compare them. If the number of different data bits does not exceed 5, it means that the two images are very similar. If it is greater than 10, it means that they are two different images.

[0206] The other steps are the same as those in the first and second embodiments.

[0207] The present invention is not limited to the contents of the above-mentioned embodiments. For example, when the SAR image of the real target is difficult to obtain, the adversarial network can be used to generate the target SAR image and replace the real image to calculate the image similarity; or the similarity of the target simulated SAR image and the target real SAR image and the GAN generated SAR image can be calculated respectively and then averaged.

Claims

1. A method for simulating and evaluating image quality of a spaceborne synthetic aperture radar image, characterized in that: include: S1, obtain target information and incident point source information; The target information includes target shape and structure information and target electromagnetic echo image; The target shape and structure information includes three-dimensional size information, composition information, surface material information and shape information of the target; The incident point source information includes the incident point source position information and the incident point source waveform information; The target electromagnetic echo image is a target SAR image obtained at the incident point source position; S2, simulating the target shape structure information and the incident point source information to obtain a simulated image of the spaceborne synthetic aperture radar target; S3, evaluating and processing the spaceborne synthetic aperture radar target simulation image and the target electromagnetic echo image to obtain image simulation quality evaluation result information; The step of evaluating and processing the spaceborne synthetic aperture radar target simulation image and the target electromagnetic echo image to obtain image simulation quality evaluation result information includes: The spaceborne synthetic aperture radar target simulation image is represented as a matrix A, and the target electromagnetic echo image is represented as a matrix B; The absolute value of the difference between matrix A and matrix B is calculated and expressed as matrix Q, Q = |AB|; The matrix Q is subjected to deviation quantization calculation processing to obtain a deviation matrix H; the calculation expression of the deviation quantization calculation processing is: In the formula, h ki represents the element that deviates from the kth row and ith column of the matrix H, q ki represents the i-th element of the k-th row of the matrix q, and the dimension of the matrix H is m×n; Decomposing the matrix A to obtain a feature matrix Y; The decomposition process is calculated as follows: A=UYV, Among them, U is the left decomposition matrix, Y is the characteristic matrix, V is the right decomposition matrix, and both U and V are orthogonal matrices; Extract the diagonal elements of the feature matrix to obtain a feature vector; the feature vector is represented by r; Performing time-frequency transformation on each row vector of the matrix B to obtain a corresponding time-frequency vector; constructing a time-frequency matrix S using all the time-frequency vectors; The time-frequency matrix S and the eigenvector r are subjected to correlation calculation processing to obtain a correlation matrix V; the calculation expression of the correlation calculation processing is: Among them, v ki represents the element in the kth row and ith column of the incidence matrix. The dimension of the incidence matrix V is m×n, r k represents the kth element of the eigenvector r, s ki is the element in the kth row and ith column of the time-frequency matrix S, and ρ is the correlation calculation factor; Using the deviation matrix H, logarithmic weighting is performed on each element of the correlation matrix V to obtain image simulation quality assessment result information; The calculation expression of the logarithmic weighting process is: Wherein, H0 is the information value of the image simulation quality assessment result.

2. The method for simulating and evaluating the image quality of a spaceborne synthetic aperture radar according to claim 1, wherein: The target shape structure information and the incident point source information are simulated to obtain a spaceborne synthetic aperture radar target simulation image, including: S21, modeling the target shape structure information to obtain target discrete grid information; S22, constructing a virtual incident wave set using incident point source information and target discrete grid information; S23, using the far-field scattering field calculation model, the virtual incident wave set and the target discrete grid information are calculated and processed to obtain a spaceborne synthetic aperture radar target simulation image.

3. The method for simulating and evaluating the image quality of a spaceborne synthetic aperture radar according to claim 2, wherein: The modeling process of the target shape structure information to obtain the target discrete grid information includes: S211, using the acquired target shape and structure information, establishing a target three-dimensional model; the target three-dimensional model includes three-dimensional size information, structure distribution information, and medium information of the target; S212, processing the target three-dimensional model to obtain target discrete grid information.

4. The method for simulating and evaluating the image quality of a spaceborne synthetic aperture radar according to claim 3, wherein: The processing of the target three-dimensional model to obtain target discrete grid information includes: S2121, performing segmentation processing on the target three-dimensional model to obtain basic facet information of the target and calculation area grid information of the target; S2122, using the target calculation area grid information, perform grid generation processing on the basic surface element to obtain the target discrete grid information.

5. The method for simulating and evaluating the image quality of a spaceborne synthetic aperture radar according to claim 2, wherein: The method of constructing a virtual incident wave set by using the incident point source information and the target discrete grid information includes: S221, using the incident point source position information, connect the coordinates of each grid center point in the target discrete grid information to obtain the corresponding virtual incident wave; S222, construct a virtual incident wave set using all virtual incident waves.

6. The method for simulating and evaluating the image quality of a spaceborne synthetic aperture radar according to claim 2, wherein: The expression of the far-zone scattering field calculation model is: in, represents the virtual incident wave The far-field electromagnetic scattering intensity information at time t, t represents the time variable, N represents the integrated electromagnetic constant vector, τ represents the incident pulse width in the incident point source waveform information, Q1 and Q2 represent the first integral function and the second integral function respectively, and the expression is: Among them, [L1, L2] are the upper and lower limits of the integration range of the x-axis; using the virtual incident wave and the two virtual incident waves closest to it, the three intersection points of the three virtual incident waves and the grid in the target discrete grid information are obtained, and the two connecting lines between the three intersection points are used to determine the corresponding two straight line equations, whose expressions are y=q1x+w1 and y=q2x+w2, respectively, and q1, w1, q2, and w2 are the coefficients of the two straight line equations respectively; using the intersection point of the virtual incident wave and the grid, and connecting with the integral point on the grid, the first straight line equation is obtained, and the expression of the first straight line equation is y=qx+w, q and w represent the coefficients of the first straight line equation; τ1 represents the time taken by the virtual incident wave from the incident point source through the intersection point of the virtual incident wave and the grid to reach the integral point of the grid, and t0 is the incident pulse time width control coefficient of the incident point source; the first integral function and the second integral function both integrate the variable t; the expression of the integral electromagnetic constant vector N is: in, Represents the unit external normal vector of the basic surface element corresponding to the mesh, represents the magnetic field unit vector in the virtual incident wave, Z0 is the free space wave impedance, c is the vacuum wave velocity, Represents the virtual incident wave corresponding vector The modulus of Modulo value, we get is the virtual incident wave The far-field electromagnetic scattering intensity information of the corresponding grid.

7. A spaceborne synthetic aperture radar image simulation and image quality assessment device, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the spaceborne synthetic aperture radar image simulation and image quality assessment method according to any one of claims 1 to 6.

8. A computer storable medium, characterized in that: The computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the spaceborne synthetic aperture radar image simulation and image quality assessment method according to any one of claims 1 to 6.

9. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the spaceborne synthetic aperture radar image simulation and image quality assessment method as described in any one of claims 1 to 6.

Images