A Fast Simulation Method for Radar Images Based on the Target's 3D Structure

Abstract

The present invention proposes a method for rapid simulation of radar images based on the three-dimensional structure of a target, and the steps are as follows: Step 1, obtaining a three-dimensional structure data set of a radar target: Step 2, dividing the radar target into facets: Step 3, constructing a radar target scattering point model: Step 4, scattering point occlusion judgment: Step 5, using the scattering point model after occlusion to generate an echo signal model: Step 6, scanning the line of sight direction and setting the imaging angle: Step 7, imaging. The simulation method proposed in the present invention avoids the shortcomings of electromagnetic simulation calculations that consume a large amount of computer memory space, take a long time to calculate, and are difficult to meet real-time requirements. Compared with the measured data, the simulation data ensures the completeness of the data set. It has been verified experimentally that the target scattering point model established by the present invention is accurate, the target scattering point echo data is obtained quickly, and the radar image set obtained by imaging technology is consistent with the actual scattering characteristics of the target, which greatly improves the simulation speed.

Description

Technical Field

[0001] The present invention is a method for rapid simulation of radar images based on the three-dimensional structure of a target, belonging to the field of radar target recognition. The method constructs the target scattering point structure by using the three-dimensional structure of the target, and obtains the radar image data set required for radar target recognition through the radar imaging method. This method has the advantages of fast simulation speed and realistic effect.

Background Art

[0002] Radar Automatic Target Recognition (ATR) technology can effectively obtain target information and improve the precision strike ability of weapons and equipment. In recent years, artificial intelligence methods have set off a wave in the field of recognition, making target recognition more and more intelligent. Only by inputting the data set into the training network can the recognition scheme be automatically learned without too much manual participation.

[0003] Target recognition technology can be divided into two categories: feature-driven target recognition technology and data-driven target recognition technology. Feature-driven target recognition technology does not require a large amount of data sets but requires feature extraction; data-driven target recognition technology requires a large amount of data sets to extract features from the data, avoiding the complex process of manual feature extraction. Feature-driven target recognition methods can be divided into micro-Doppler target recognition, one-dimensional image target recognition, comprehensive target recognition, etc. Data-driven target recognition methods can be divided into classical statistical pattern recognition, knowledge-based automatic target recognition, model-based automatic target recognition, multi-sensor fusion-based automatic target recognition, artificial neural network and expert system-based automatic target recognition, etc. Data-driven target recognition methods use the original data set as the network input, avoiding the complex feature extraction process, and transforming the original data into a more abstract expression through a simple non-linear model, playing an increasingly important role in the field of automatic target recognition.

[0004] The acquisition of data is a crucial link for data-driven target recognition. There are two ways to acquire data. One is measured data. Due to the high confidentiality of space targets and the lack of a mature radar measurement system, it is difficult to carry out field tests; radar scattering field chamber measurement faces a series of problems such as polarization calibration, antenna coupling, and target alignment, with high technical difficulty. Moreover, the measured data is insufficient in completeness, unable to give the echo model in various different situations, and also has the disadvantages of high acquisition cost and even some data that cannot be obtained. The other is simulation data, with flexible parameter settings, which can theoretically simulate the electromagnetic scattering characteristics of the target and the environment at any observation angle, greatly reducing the data acquisition cost and improving the research efficiency. Radar echo simulation has become a cheap, efficient and necessary technical means.

[0005] There are two main radar echo simulation methods: electromagnetic calculation and scattering point model simulation. Radar echoes simulated by electromagnetic calculation data can reflect the real scattering characteristics of the target, but electromagnetic simulation data processing consumes computer memory extremely and takes a long time to calculate, which is often unbearable. If every complex scene is simulated at every moment, it is almost impossible to meet the requirements of engineering applications using electromagnetic simulation methods.

[0006] Based on this, the present invention proposes a rapid simulation method of radar images based on the three-dimensional structure of the target, establishes a target scattering point model, quickly constructs the target scattering point radar echo data, and efficiently obtains the radar image data set through radar imaging technology, which greatly improves the simulation speed. [Summary of the invention]

[0007] The technical problem to be solved by the present invention is: in view of the problems that the current radar image simulation method is highly complex, time-consuming, and has a high cost for obtaining measurement data, the present invention constructs a target scattering point model by using a target three-dimensional structure model (division facets), then constructs a target echo by using a signal model, and then uses imaging technology to quickly simulate the two-dimensional image of the radar target. Finally, through simulation experiments, it is fully verified that the present invention can accurately obtain a radar simulation image that conforms to the target scattering mechanism, and the simulation speed is greatly improved, which can effectively solve the problem of difficult data acquisition in the field of radar target recognition.

[0008] The present invention provides a method for rapid simulation of radar images based on target three-dimensional structure, and the technical solution adopted comprises the following steps:

[0009] The first step is to obtain the radar target 3D structure dataset

[0010] The radar target three-dimensional structure data set is obtained through modeling software (such as AutoCAD), and its three-dimensional file format is iges / step.

[0011] The second step is to divide the radar target into surface elements.

[0012] The triangular face element set of the three-dimensional model is obtained by dividing the face element according to the side length of the triangular face element of the electromagnetic wave wavelength λ / 8 to λ / 10.

[0013] The third step is to construct a radar target scattering point model

[0014] The total electromagnetic scattering of the target in the optical zone is synthesized by the electromagnetic scattering at certain local positions, and these local scattering sources are called scattering centers. In the high-frequency zone, the total electromagnetic scattering of the target can be considered to be synthesized by the electromagnetic scattering at certain local positions, and these local scattering sources are usually called equivalent multi-scattering centers. Therefore, the present invention regards the centers of the subdivided triangular facets as equivalent to scattering points.

[0015] Step 4: Scattering point occlusion judgment

[0016] Identify the occluded triangular facets from the target 3D model, and at the same time remove the scattering centers from the occluded facets, which can make the radar target characteristics consistent with the real situation.

[0017] Step 5: Generate the echo signal model using the occluded scattering point model

[0018] According to the radar waveform (such as the chirp signal), generate the echo waveform using the set of scattering centers after occlusion judgment;

[0019] Step 6: Scan the line-of-sight direction and set the imaging angle

[0020] Set waveform parameters such as the frequency step Δf and bandwidth B of the signal, as well as the electromagnetic wave irradiation angle step Δδ and the angle transformation range δ to obtain a broadband data matrix in a certain airspace.

[0021] Step 7: Imaging

[0022] Use radar imaging technology to perform coherent processing on the echo data matrix to obtain a simulated radar image.

[0023] The advantages and beneficial effects of the present invention are as follows: The present invention is applied in the field of radar target recognition. Through the fast simulation method of radar images based on the target 3D structure proposed by the present invention, it avoids the deficiencies of large computer memory consumption, long calculation time, and difficulty in meeting real-time performance in electromagnetic simulation calculations. Compared with measured data, the simulation data ensures the completeness of the data set. Verified by experiments, the target scattering point model established by the present invention is accurate, the target scattering point echo data can be obtained quickly, and the radar image set obtained through imaging technology is consistent with the real scattering characteristics of the target, greatly improving the simulation speed.

Description of the Drawings

[0024] Figure 1 It is a flowchart of the fast simulation method of radar images based on the target 3D structure.

[0025] Figure 2 It is a 3D structure diagram of the radar target.

[0026] Figure 3 It is a schematic diagram of the conversion from the radar target facet model to the scattering point model.

[0027] Figure 4 It is the converted radar target scattering point model.

[0028] Figure 5 It is the line-of-sight direction and the definition schematic diagram of θ, and.

[0029] Figure 6 It is the occluded radar target scattering point model.

[0030] Figure 7 It is a schematic diagram of the radar target imaging angle scanning method.

[0031] Figure 8a is the radar target observation angle Imaging results when θ=150°.

[0032] Figure 8b is the radar target observation angle Imaging results when θ=140°.

[0033] Figure 8c is the radar target observation angle Imaging results for θ=130°.

[0034] Figure 8d is the radar target observation angle Imaging results when θ=120°. [Specific implementation method]

[0035] The following is combined with Figure 1 -8 further describes the present invention. The present invention is a rapid simulation method of radar images based on target three-dimensional structure, and its implementation flow chart is as follows: Figure 1 As shown, the steps are as follows:

[0036] The first step is to obtain the radar target 3D structure dataset

[0037] Through the target 3D modeling, the 3D structure of the radar target is obtained, and its 3D file format can be iges / step. For example, the 3D structure model of the radar target is as follows Figure 2 shown.

[0038] The second step is to divide the radar target into surface elements.

[0039] Using the existing 3D modeling software, the 3D structure of the radar target is divided into facets. The facets are divided according to the side length of the triangular facets of the electromagnetic wave wavelength λ / 8 to λ / 10 to obtain a set of triangular facets of the 3D model. That is, the 3D model of the target is divided into a series of triangular facets, and these facet sets are used to represent the target. The method of moments (MoM), one of the three classic algorithms of computational electromagnetics, is a commonly used electromagnetic precision calculation method. It usually requires the target 3D model to be divided into a size between 1 / 8 and 1 / 10 of the wavelength. However, in the actual calculation process, if the electrical size of the 3D model is too large, it is divided into 1 / 4 of the wavelength.

[0040] After being divided into surface elements, the electromagnetic simulation software is not used to calculate the target electromagnetic field at this time, which would be very time-consuming. Instead, the surface element model is exported and prepared to be transformed into a scattering point model using this method, and then fast simulation is carried out. The method of transforming into a scattering point model will be introduced in detail in the next step.

[0041] In the third step, the center of the divided surface element is used as the scattering point to construct the radar target scattering point model.

[0042] In the high-frequency region, the target can be equivalent to discrete scattering centers. The center of the divided triangular surface element is equivalent to the scattering point. Physically speaking, the scattering center theory holds that the total electromagnetic scattering of the target in the optical region is synthesized by the electromagnetic scattering at some local positions, and these local scattering sources are called scattering centers, which reflect the local structural characteristic information of the target; Mathematically speaking, the scattering center corresponds to the upper and lower limits, singularities or stationary phase points of the integral term of the scattering field. In the optical region, the scattering centers on the extended target objectively exist and are independent of the sensing methods such as the waveform and bandwidth of the radar.

[0043] In a three-dimensional structure, the data of each divided surface element is composed of the three vertex coordinates of the triangular surface element, and each vertex can be represented by three coordinates X, Y, and Z. The average value of the three vertex coordinate values of each surface element data is taken to obtain a corresponding scattering point coordinate data. That is to say, the average value of the X coordinates of the three vertices is the X coordinate of the equivalent scattering point, the average value of the Y coordinates of the three vertices is the Y coordinate of the equivalent scattering point, and the average value of the Z coordinates of the three vertices is the Z coordinate of the equivalent scattering point. Finally, a new set of X, Y, and Z coordinates is obtained as the coordinate data of the equivalent scattering point. These sets of scattering points constitute the scattering point model of the radar target. The schematic diagram of the transformation from the radar target surface element model to the scattering point model is as Figure 3 shown. The scattering point model is as Figure 4 shown.

[0044] In the fourth step, occlusion judgment of the scattering point model.

[0045] When a 2D picture is taken from a 3D scene in the real world, occlusion will inevitably occur, that is, the object closer to the camera will block the object behind it, making it partially invisible. Identifying occlusion from a three-dimensional model and simultaneously inferring the occlusion and occluded relationships between objects can make the imaging approach the real situation.

[0046] The Hidden Points Remove (HPR) algorithm is an algorithm for 3D model visualization. Its main function is to eliminate occluded points and improve rendering efficiency. Based on the camera view, it is commonly used in real-time rendering in computer graphics. The main principle of this algorithm is based on the depth buffer (Z-buffer) technology and adopts the following steps: Traverse all equivalent scattering points and calculate the depth values of all scattering points according to the distance from the scattering points to the observation point. Compare the depth value of each scattering point with the corresponding depth value in the depth buffer. If the current scattering point is closer to the observer, update the depth value in the depth buffer and record this scattering point to form a set of "visible" scattering points. The scattered point model after occlusion judgment is as Figure 6 shown. HPR processing eliminates the problem of front and back point occlusion, and the generated scattered point model is more in line with the actual situation of radar observation.

[0047] Step 5: Generate an echo signal model using the scattered point model after occlusion

[0048] The radar echo signal model is

[0049]

[0050] where N is the number of scattering points in the set of visible scattering points of the radar target recorded in Step 4, f is the frequency vector, f 0 is the center frequency, B is the signal bandwidth, and Δf is the frequency step. The time delay where c is the speed of light.

[0051] The determination of the bandwidth and frequency step parameters is determined according to Equation (2).

[0052]

[0053] where the maximum unambiguous range in the range direction is R max which should be greater than the length of the target, and the resolution ΔR should meet the actual imaging requirements.

[0054] Assume that a certain scattering point in the set of visible scattering points of the radar target is denoted as P n , and its distance R n can be obtained by taking the dot product of the scattering point coordinate vector and the radar line-of-sight direction vector, as shown in the following equation

[0055] R n =(x,y,z)·i (3)

[0056] where (x,y,z) represents the three-dimensional coordinates of the scattering point P n ; i represents the unit vector of the radar line-of-sight direction (line of sight, LOS), which can be obtained through the following calculation

[0057]

[0058] where r is the radial distance from the observation point to the origin of the coordinate system; θ is the elevation angle, representing the angle between the observation line of sight and the Z-axis; is the azimuth angle, representing the angle between the observation line of sight and the X-axis. The definition of the line of sight direction and θ, is as shown in Figure 5 The projection of the scattering point P n in the direction of the line of sight LOS is the distance R n .

[0059] Step 6: Scan the line of sight direction and set the imaging angle

[0060] To obtain the radar simulation image of the target, not only frequency scanning but also relative motion is required, so the irradiation direction of the electromagnetic wave needs to be changed. The relative motion can be achieved by continuously changing the elevation angle of radar observation or continuously changing the azimuth angle. The process of changing the observation angle is as shown in Figure 7 . Among them, continuously changing the elevation angle or continuously changing the azimuth angle respectively corresponds to different imaging planes (the imaging plane is determined by the direction of the line connecting the middle observation point and the target and the direction of the line of sight angle change).

[0061] The range of angle change and the step of angle change are determined according to Equation (5).

[0062]

[0063] where the unambiguous range x max in the azimuth direction should be greater than the transverse length of the target, and the azimuth resolution Δx should meet the actual imaging requirements.

[0064] Step 7: Imaging

[0065] After Step 6, the wideband echo data matrix is obtained, where a certain row corresponds to the target scattering echo data at different angles under a fixed frequency point. The radar simulation image can be finally obtained by performing imaging processing on the wideband echo data matrix. Imaging algorithms can select relatively mature range-Doppler algorithms, polar format algorithms, or convolution backprojection algorithms, etc.

[0066] For simulation verification, set the bandwidth B = 1.5 GHz, the starting frequency 10 GHz, the ending frequency 11.5 GHz, and the frequency interval 15 MHz. The observation azimuth angle is The pitch angles are 120°, 130°, 140° and 150° respectively, the angle step is 0.08°, and the azimuth angle variation range is 8°. The number of target dissection surface elements is 55635, and the number of triangular surface elements after occlusion judgment is directly related to the observation angle. The simulation experiment uses the polar coordinate format algorithm for radar imaging, and finally obtains the imaging result of the radar target as Figures 8a - 8d shown.

Claims

1. A rapid simulation method of radar images based on target three-dimensional structure, It is characterized in that Here are the steps: Step 1: Obtain the radar target 3D structure dataset: Obtain the radar target 3D structure data set through modeling software, and its 3D file format is iges or step; Step 2: Divide the radar target into triangular face elements: The triangular face element set of the three-dimensional model is obtained by dividing the face element according to the side length of the triangular face element of the electromagnetic wave wavelength λ / 8 to λ / 10; Step 3: Construct radar target scattering point model: The total electromagnetic scattering of the target in the optical zone is synthesized by the electromagnetic scattering at local positions, and these local scattering sources are called scattering centers; in the high-frequency zone, the total electromagnetic scattering of the target is synthesized by the electromagnetic scattering at local positions, and these local scattering sources are called equivalent multi-scattering centers; the centers of the subdivided triangular face elements are equivalent to scattering points; Step 4: Scattering point occlusion judgment: Determine the blocked triangular facets from the target 3D model and simultaneously remove the blocked triangular facets from the scattering center to make the radar target characteristics consistent with the actual situation; In step 4, the hidden point removal algorithm is used to traverse all equivalent scattering points, and the depth values ​​of all scattering points are calculated according to the distance from the scattering point to the observation point; Compare the depth value of each scattering point with the corresponding depth value in the depth buffer. If the current scattering point is closer to the observer, update the depth value in the depth buffer and record the scattering point to form a "visible" scattering point set. Use HPR processing to eliminate the problem of front and rear point occlusion. Step 5: Generate the echo signal model using the scattering point model after occlusion: According to the radar waveform, the echo waveform is generated using the scattering center set after the occlusion judgment; Step 6: Scan the sight direction and set the imaging angle: The frequency step Δf of the signal, the waveform parameters of the bandwidth B, the electromagnetic wave irradiation angle step Δδ, and the angle change range δ are set to obtain a broadband data matrix; Step 7, Imaging: Radar imaging technology is used to perform coherent processing on the echo data matrix to obtain a simulated radar image.

2. According to claim 1, a radar image rapid simulation method based on target three-dimensional structure, Features: In step 2, the three-dimensional model of the target is divided into a series of triangular surface elements, and the target is represented by a set of these triangular surface elements.

3. According to claim 1, a rapid simulation method of radar image based on target three-dimensional structure, Features: In step three, in the three-dimensional structure, each subdivided triangular face data is composed of three vertex coordinates of the triangular face unit, and each vertex can be represented by three coordinates of X, Y, and Z. The three vertex coordinate values ​​of each triangular face data are averaged to obtain a corresponding scattering point coordinate data; that is, the average value of the three vertex X coordinates is the X coordinate of the equivalent scattering point, the average value of the three vertex Y coordinates is the Y coordinate of the equivalent scattering point, and the average value of the three vertex Z coordinates is the Z coordinate of the equivalent scattering point. Finally, a new X, Y, and Z coordinates are obtained as the coordinate data of the equivalent scattering point; these scattering point sets constitute the scattering point model of the radar target.

4. A fast simulation method of radar images based on a target three-dimensional structure according to claim 1, characterized in that: In step five, the radar echo signal model is: Where N is the number of scattering points in the scattered point set where the radar target is visible recorded in Step 4, f is the frequency vector, f 0 is the center frequency, B is the signal bandwidth, and Δf is the frequency step; where the time delay c is the speed of light, and the projection of the scattering point in the line-of-sight (LOS) direction is the distance R n ; The determination of the bandwidth and frequency step parameters is determined according to Equation (2); where R max represents the maximum unambiguous range in the range direction, and ΔR represents the azimuth resolution.

5. A fast simulation method of radar images based on a target three-dimensional structure according to claim 4, characterized in that: The concentrated scattering points where the radar target is visible are denoted as P n , and its distance R n is obtained by taking the dot product of the scattering point coordinate vector and the unit vector in the radar line-of-sight direction, as shown in the following formula: R n = (x, y, z)·i (3) where (x, y, z) represents the three-dimensional coordinates of the scattering point P n ; i represents the unit vector in the radar line-of-sight direction, which is obtained through the following calculation where r is the radial distance from the observation point to the origin of the coordinate system; θ is the elevation angle, representing the angle between the observation line of sight and the Z-axis; is the azimuth angle, representing the angle between the observation line of sight and the X-axis; the projection of the scattering point P n in the direction of the line of sight LOS is the distance R n .

6. A fast simulation method of radar images based on a target three-dimensional structure according to claim 5, characterized in that: In step six, in order to obtain the radar simulation image of the target, not only frequency scanning but also relative motion is required, so the irradiation direction of the electromagnetic wave needs to be changed; the relative motion is achieved by continuously changing the elevation angle of the radar observation or continuously changing the azimuth angle; Among them, continuously changing the elevation angle or continuously changing the azimuth angle respectively corresponds to different imaging planes, that is, the imaging plane is determined by the direction of the line connecting the middle observation point and the target and the direction of the line of sight angle change; The range of angle change and the step of angle change are determined according to Equation (5); where x max represents the unambiguous distance, and Δx represents the resolution.

7. A fast simulation method of radar images based on a target three-dimensional structure according to claim 6, characterized in that: In step seven, in the wideband echo data matrix, a certain row corresponds to the target scattering echo data at different angles under a fixed frequency point; The wideband echo data matrix is finally processed by imaging to obtain the radar simulation image; the imaging algorithm selects the range-Doppler algorithm, the polar format algorithm or the convolution backprojection algorithm.

Images