Terahertz band large-size rotating target imaging method, system and device

Abstract

The application relates to a terahertz frequency band large-size rotating target imaging method, system and device. The method is designed based on local image maximum contrast, particle swarm optimization distance center offset and rotating speed estimation and focused imaging. After a first-order distance cell migration correction of a baseband echo signal is performed through KT correction, second-order distance cell migration correction is performed by using estimated parameters and a fast Gaussian grid non-uniform fast Fourier transform, then second-order space-varying phase compensation is performed, finally, an azimuth fast Fourier transform is performed, two-dimensional windowing is performed, and then a particle swarm optimization algorithm is used to iteratively estimate two to-be-estimated parameters. When the local image contrast corresponding to the iteration termination reaches the maximum, a distance Doppler imaging result with good focusing can be obtained. The method greatly reduces the operation amount of search, is simple and efficient, and finally achieves the purpose of effectively improving the performance of terahertz frequency band large-size target imaging.

Description

Technical Field

[0001] This invention belongs to the field of target imaging data processing technology, and relates to a method, system and device for imaging large-size rotating targets in the terahertz band. Background Technology

[0002] Terahertz radar, compared to traditional microwave bands, features a wider bandwidth and shorter wavelength, thus offering higher range and azimuth resolution when imaging targets using inverse synthetic aperture radar (ISAR). Due to this higher resolution, when terahertz radar performs range Doppler (RD) imaging on large, uniformly rotating targets, even with a small accumulation angle (e.g., 3°–5°), significant range cell migration (RCM) can still occur, potentially even leading to higher-order RCM and resulting in image defocusing. However, in the traditional range-Doppler imaging model, the Taylor expansion of the trigonometric function only retains the first-order terms, often ignoring the influence of second-order and higher-order terms. When performing range-Doppler imaging of larger targets in the terahertz band, it is not enough to simply perform the traditional Keystone Transform (KT transform), as the second-order terms cannot be ignored. Therefore, in the process of realizing this invention, the inventors discovered that the traditional method has a technical problem of insufficient imaging performance in imaging larger targets in the terahertz band. Summary of the Invention

[0003] To address the problems existing in the above-mentioned traditional methods, this invention proposes a large-size rotating target imaging method in the terahertz band, a large-size rotating target imaging system in the terahertz band, and a target imaging device, which can effectively improve the performance of large-size target imaging in the terahertz band.

[0004] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0005] On the one hand, a method for imaging large rotating targets in the terahertz band is provided, including the following steps:

[0006] The inverse synthetic aperture radar echo data of a uniformly rotating large target is acquired and subjected to line frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle.

[0007] KT transform is applied to the baseband echo signal to eliminate first-order cross-range cell migration, and fast Gaussian grid non-uniform fast Fourier transform is used to eliminate second-order cross-range cell migration.

[0008] After constructing a spatially variable phase compensation function and compensating for the phase error of the baseband echo signal after eliminating the migration of the distance cell, a fast Fourier transform in the azimuth direction is performed to obtain the two-dimensional imaging result.

[0009] After constructing a U-shaped two-dimensional window, a window is added to the two-dimensional imaging result and the image contrast of the local image of the two-dimensional imaging result is calculated;

[0010] Under the criterion of maximizing local image contrast, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated. When the iteration terminates, the local image contrast reaches its maximum, and the final estimated two parameters to be estimated and the well-focused imaging result of a large target rotating at a constant speed are obtained. The two parameters to be estimated include the distance-center offset and rotation speed of the large target rotating at a constant speed.

[0011] In one embodiment, the process of performing KT transform on the baseband echo signal to eliminate first-order cross-range cell migration includes:

[0012] A new slow-time variable is constructed and then approximated by the baseband echo signal to obtain a baseband echo signal with KT transform eliminating first-order cross-range cell migration; the slow-time variable τ m for:

[0013]

[0014] Among them, f c For radar carrier frequency, t m For azimuth slow time, γ is the frequency modulation. For the distance to fast time, f r For fast time frequency.

[0015] In one embodiment, during the process of eliminating second-order cross-range cell migration using a fast Gaussian grid non-uniform fast Fourier transform, a resampling factor k is defined. resample for:

[0016]

[0017] Where ω is the target's rotational speed, τ m To construct new slow-time variables.

[0018] On the other hand, a large-size rotating target imaging system in the terahertz band is also provided, including:

[0019] The echo acquisition module is used to acquire the inverse synthetic aperture radar echo data of a uniformly rotating large target and perform line frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle.

[0020] The migration cancellation module is used to perform KT transform on the baseband echo signal to eliminate first-order cross-range cell migration, and to use fast Gaussian grid non-uniform fast Fourier transform to eliminate second-order cross-range cell migration.

[0021] The spatial variation compensation module is used to construct the spatial variation phase compensation function and perform phase error compensation on the baseband echo signal after eliminating the migration of the distance cell, and then perform azimuth fast Fourier transform to obtain the two-dimensional imaging result.

[0022] The windowing module is used to construct a "U"-shaped two-dimensional window, then window the two-dimensional imaging result and calculate the image contrast of the local image of the two-dimensional imaging result;

[0023] The imaging module is used to iteratively estimate two parameters to be estimated using a particle swarm optimization algorithm under the criterion of maximizing local image contrast. When the iteration terminates, the local image contrast reaches its maximum, and the final estimated two parameters to be estimated and the well-focused imaging result of a large-size target rotating at a constant speed are obtained. The two parameters to be estimated include the distance-to-center offset and rotation speed of the large-size target rotating at a constant speed.

[0024] In one embodiment, the migration cancellation module, during the process of performing KT transform on the baseband echo signal to cancel first-order cross-range cell migration, includes:

[0025] A new slow-time variable is constructed and then approximated by the baseband echo signal to obtain a baseband echo signal with KT transform eliminating first-order cross-range cell migration; the slow-time variable τ m for:

[0026]

[0027] Among them, f c For radar carrier frequency, t m y is the azimuth slow time, γ is the frequency modulation, t^ is the range fast time, and f r For fast time frequency.

[0028] In one embodiment, the migration elimination module defines a resampling factor k during the process of eliminating second-order cross-range cell migration using a fast Gaussian grid non-uniform fast Fourier transform. resample for:

[0029]

[0030] Where ω is the target's rotational speed, τ m To construct new slow-time variables.

[0031] On another front, a target imaging device is also provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described terahertz band large-size rotating target imaging method.

[0032] One of the above technical solutions has the following advantages and beneficial effects:

[0033] The aforementioned terahertz band large-size rotating target imaging method, system, and device, through the design concept of local image maximum contrast and particle swarm optimization for range center offset and rotational speed estimation, as well as focused imaging, fully leverages the high efficiency of the particle swarm optimization algorithm. Combined with the defocusing characteristics caused by target rotation, a rotation parameter estimation method based on local image maximum contrast is proposed. After performing KT correction on the baseband echo signal for first-order cross-range cell migration, second-order cross-range cell migration correction is performed using the estimated parameters and fast Gaussian grid non-uniform fast Fourier transform. Then, second-order spatially variable phase compensation is performed, followed by azimuth-direction fast Fourier transform. After two-dimensional windowing, the particle swarm optimization algorithm is used to iteratively estimate the two parameters. When the iteration terminates, the corresponding local image contrast reaches its maximum, ultimately yielding well-focused range Doppler imaging results.

[0034] Compared to traditional methods, the proposed technical solution accurately corrects the migration of first-order and second-order cross-range units. After performing second-order spatially variable phase compensation and azimuth fast Fourier transform, the image can be focused and imaged. Finally, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated, which greatly reduces the amount of computation in the search. It has the characteristics of simplicity and efficiency, and ultimately achieves the goal of effectively improving the performance of large-size target imaging in the terahertz band. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 A flowchart of a method for imaging a large rotating target in the terahertz band in one embodiment;

[0037] Figure 2 This is a flowchart of parameter estimation and focusing imaging for a large rotating target imaging method in the terahertz band in one embodiment;

[0038] Figure 3 This is a schematic diagram of range Doppler imaging of the baseband echo signal in one embodiment;

[0039] Figure 4 This is a schematic diagram of a two-dimensional imaging result and a local magnification obtained after KT transformation in one embodiment, wherein (a) is the imaging result after KT transformation, and (b) is a local magnification of the imaging result after KT transformation;

[0040] Figure 5 This is a schematic diagram of a two-dimensional window and the imaging result after windowing in one embodiment, where (a) is the two-dimensional window and (b) is the range Doppler imaging result after windowing;

[0041] Figure 6 The image shows the imaging result and a magnified view of a large rotating target imaging method in the terahertz band in one embodiment. (a) is the obtained imaging result, and (b) is a magnified view of the imaging result.

[0042] Figure 7 This is a block diagram of a terahertz band large-size rotating target imaging system in one embodiment. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0045] It should be noted that, in this document, the reference to "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The presentation of this phrase in various locations throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments.

[0046] Those skilled in the art will understand that the embodiments described herein can be combined with other embodiments. The term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items, and all possible combinations thereof.

[0047] In realizing this invention, the inventors discovered that for inverse synthetic aperture radar (ISAR) imaging of large targets, range migration correction caused by rotation and high-order phase compensation are necessary. Traditional methods typically employ KT transform for parameterless range migration correction, followed by direct two-dimensional imaging of the KT-transformed data. When the target's longitudinal dimension is large, the influence of the second-order Taylor expansion term on imaging needs to be considered, including second-order range migration correction and second-order spatially variable phase compensation. The unknown rotation parameters include rotational speed information and the longitudinal offset distance of the rotation center, requiring two-dimensional search estimation of these parameters. Particle swarm optimization (PSO) is a relatively efficient parameter search method. Furthermore, when performing second-order range migration correction, the radially non-uniform data needs to be transformed. Theoretically, Fast Gaussian Gridding Nonuniform Fast Fourier Transform (FGG NUFFT) and Chirp-Z transform are both feasible.

[0048] Currently, traditional rotational velocity estimation methods can be broadly categorized into three types: The first type estimates parameters based on the correlation coefficient of the signal phase, including demodulation estimation, Radon-Wigner transform, and discrete polynomial phase transform. The accuracy of this type of algorithm largely depends on the quality of the extracted scattering points; its accuracy decreases significantly when the signal-to-noise ratio is low or there are many signal components. The second type estimates rotational velocity based on the correlation between adjacent frames of the target's rotation, including correlation methods, affine projection matrix methods, and image feature registration methods. These algorithms require long observation times and involve significant computational demands for image processing. The third type prioritizes optimal overall image quality, proposing methods based on minimum image entropy and maximum image sharpness. These methods fully utilize overall image information, are suitable for target rotational velocity estimation under low signal-to-noise ratio conditions, and offer high estimation accuracy.

[0049] The most similar traditional method to this invention is to iterate through and search for relevant parameters of second-order phase compensation based on the criteria of minimum image entropy or maximum image contrast. The method based on maximum image contrast first constructs a quadratic compensation function to compensate the range pulse compression signal, then performs a Fourier transform on the compensated signal along the azimuth direction to obtain the range-Doppler imaging result. Next, it calculates the contrast of the range-Doppler imaging result. When the contrast of the resulting image is maximized, the rotational speed corresponding to the quadratic compensation function is the final estimated value of the target's rotational speed.

[0050] The aforementioned traditional methods fail to consider the impact of higher-order cross-range cell migration. When such migration exists, directly performing phase compensation on the echo before cross-range cell migration correction is inaccurate. Furthermore, these methods do not account for the influence of target scattering intensity on parameter search. Since the degree of defocusing caused by rotation is proportional to the distance from the point target to the rotation center, when scattering points near the rotation center are strong, using global image contrast for parameter search will make it difficult to accurately focus weaker scattering points farther from the rotation center, leading to errors. Moreover, this invention uses image contrast rather than image entropy as the criterion function because image entropy values ​​are relatively small and change very little during parameter search, which is detrimental to accurate estimation of rotation parameters.

[0051] Directly solving for the two variables by maximizing image contrast is computationally intensive and inefficient. Therefore, this invention employs a method that first uses a KT transform on the original echo data to eliminate the most significant first-order range cell migration, then uses a fast Gaussian grid non-uniform fast Fourier transform for second-order range cell migration correction, achieving precise range focusing followed by precise second-order spatially varying phase compensation. Finally, a fast Fourier transform in the azimuth direction yields a well-focused two-dimensional inverse synthetic aperture radar (ISSAR) image. In parameter estimation, a particle swarm optimization method is used to construct a range resampling factor and a second-order spatially varying phase compensation function. The two-dimensional parameter search is performed based on the maximum contrast of the local image of the ISAR image, where the local image is the outer portion of the original ISAR image after adding a U-shaped window. This method significantly reduces computational cost and avoids the impact of strong scattering points near the rotation center on parameter estimation.

[0052] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0053] The main design idea of ​​this invention is as follows: KT and range-oriented non-uniform fast Fourier transform are applied to the original echo to accurately correct the migration of first-order and second-order range cells. After second-order spatially variable phase compensation and azimuth-oriented fast Fourier transform, focused imaging can be achieved. During the imaging process, unknown rotational parameters, namely rotational speed and longitudinal distance center offset (RCD), are involved. Combined with the defocusing characteristics caused by the migration of range cells, the accuracy of the estimated parameters and the optimal focusing effect are determined by the local image contrast of the imaging result. The parameter search method employs a particle swarm optimization algorithm.

[0054] Please see Figure 1In one embodiment, a method for imaging a large rotating target in the terahertz band is provided, comprising the following processing steps S12 to S20:

[0055] S12: Acquire the inverse synthetic aperture radar echo data of a uniformly rotating large target and perform line frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle.

[0056] S14 performs KT transform on the baseband echo signal to eliminate first-order cross-range cell migration, and uses fast Gaussian grid non-uniform fast Fourier transform to eliminate second-order cross-range cell migration.

[0057] S16. After constructing the spatially variable phase compensation function and compensating for the phase error of the baseband echo signal after eliminating the migration of the distance cell, the azimuth fast Fourier transform is performed to obtain the two-dimensional imaging result.

[0058] S18: After constructing a "U"-shaped two-dimensional window, add a window to the two-dimensional imaging result and calculate the image contrast of the local image of the two-dimensional imaging result.

[0059] S20, under the criterion of maximizing local image contrast, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated. When the iteration terminates, the local image contrast reaches its maximum, and the final estimated two parameters to be estimated and the well-focused imaging result of the uniformly rotating large target are obtained. The two parameters to be estimated include the distance center offset and rotation speed of the uniformly rotating large target.

[0060] The aforementioned terahertz band large-size rotating target imaging method, through the design concept of local image maximum contrast and particle swarm optimization for range center offset and rotational speed estimation, as well as focused imaging, fully leverages the high efficiency of the particle swarm optimization algorithm. Combined with the defocusing characteristics caused by target rotation, a rotation parameter estimation method based on local image maximum contrast is proposed. After performing KT correction on the baseband echo signal for first-order range cell migration, second-order range cell migration correction is performed using the estimated parameters and fast Gaussian grid non-uniform fast Fourier transform. Then, second-order spatially variable phase compensation is performed, followed by azimuth-directed fast Fourier transform. After two-dimensional windowing, the particle swarm optimization algorithm is used to iteratively estimate the two parameters. When the iteration terminates, the corresponding local image contrast reaches its maximum, ultimately yielding well-focused range Doppler imaging results.

[0061] Compared to traditional methods, the proposed technical solution accurately corrects the migration of first-order and second-order cross-range units. After performing second-order spatially variable phase compensation and azimuth fast Fourier transform, the image can be focused and imaged. Finally, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated, which greatly reduces the amount of computation in the search. It has the characteristics of simplicity and efficiency, and ultimately achieves the goal of effectively improving the performance of large-size target imaging in the terahertz band.

[0062] It should be noted that, specifically, step S12 involves acquiring the inverse synthetic aperture radar echo data of a large target rotating at a constant speed, simplifying the inverse synthetic aperture radar far-field imaging scene of the target into a turntable imaging, and approximating it using a high-order Taylor expansion.

[0063] First, broadband echo acquisition. This embodiment is for a terahertz broadband radar system, whose transmitted signal is shown in the following formula.

[0064]

[0065] in, For the distance to fast time, t m For the direction of slow time, T p f is the pulse width of a pulse radar or the frequency sweep period of a frequency-modulated continuous wave radar. c γ is the radar carrier frequency, and γ is the frequency modulation frequency.

[0066] Assume the distance from the radar to the center of the turntable is R. ref The reference echo can then be written as:

[0067]

[0068] Where c is the speed of light, rect(·) represents a square pulse, exp represents an exponential function with base e, j is the imaginary unit, and the target scattering coefficient is 1 by default.

[0069] Inverse synthetic aperture radar (ISAR) imaging in the far field can be simplified to turntable imaging. Assuming the distance between the rotating target and the radar is R, a Cartesian coordinate system (xoy) is established with the turntable center as the origin. In turntable mode, the distance from the target point P(x0, y0) to the radar can be expressed as R = R0 + x0sinθ + y0cosθ, where R0 is the initial distance from the radar to the turntable center, and θ is the target rotation accumulation angle. The received echo can be written as:

[0070]

[0071] Assuming the reference distance and the turntable center distance are equal, after demodulating the line frequency and compensating for the residual video phase (RVP), the expression for the received baseband target echo signal is:

[0072]

[0073] Secondly, a second-order Taylor expansion of the sine and cosine of the accumulation angle is performed to obtain a more accurate turntable imaging model. In traditional range-Doppler imaging, the accumulation angle is generally small (e.g., 3°–5°), and the approximation often only retains the first order of the Taylor expansion, i.e.:

[0074]

[0075] Where ω is the rotational speed, since the resolution of traditional frequency band radar is not high, the influence of second-order and higher-order expansion terms on imaging can be ignored. However, in the terahertz band, especially for large rotating targets, the influence of the second-order terms cannot be ignored. If the second-order expansion is retained, the above equation (5) can be expressed as:

[0076]

[0077] Therefore, a more accurate turntable echo model can be expressed as:

[0078]

[0079] Furthermore, the baseband echo signal of the target is subjected to KT transform and then to range-oriented non-uniform Fourier transform. First, the KT transform is performed to eliminate the influence of first-order cross-range cell migration. Further, regarding step S14 above, the process of eliminating first-order cross-range cell migration by performing KT transform on the baseband echo signal can include the following processing steps:

[0080] A new slow-time variable is constructed and then approximated by the baseband echo signal to obtain a baseband echo signal with KT transform eliminating first-order cross-range cell migration; the slow-time variable τ m for:

[0081]

[0082] Among them, f c For radar carrier frequency, t m For azimuth slow time, γ is the frequency modulation. For the distance to fast time, f r For the fast time frequency, substituting the newly constructed slow time variable into the baseband echo signal and making appropriate approximations yields:

[0083]

[0084] After KT transformation, the above equation shows that the three exponential terms correspond to azimuth imaging, range imaging, and second-order spatially variable phase error, respectively.

[0085] In one embodiment, further, regarding the above step S14, in the process of eliminating the second-order cross-range unit migration using the fast Gaussian gridding non-uniform fast Fourier transform, it can be observed that the data corresponding to range direction imaging is non-uniform, so the defined resampling factor k resample is:

[0086]

[0087] where ω is the rotation speed of the target, τ m is the newly constructed slow time variable. The fast Gaussian gridding non-uniform fast Fourier transform can realize the Fourier transform of non-uniform data, thereby eliminating the influence of the second-order cross-range unit migration and achieving precise focusing in the range direction.

[0088] Regarding the above step S16, it can be understood that constructing the space-variant phase error compensation function where Δy and y0 are the RCD and the longitudinal range information without range offset respectively, phase compensation is performed on the echo signal after KT transform and fast Gaussian gridding non-uniform fast Fourier transform, and then the azimuth direction fast Fourier transform is performed to obtain the focused range-Doppler imaging result. In summary, the expression of the imaging algorithm can be expressed as follows:

[0089]

[0090] where represent the estimated RCD and rotation speed respectively.

[0091] Regarding the above step S18, it can be understood that a "hui"-shaped two-dimensional window is constructed first where (m, n) represent the row and column indices of the two-dimensional matrix respectively, ns and ne, ms and me represent the start and end of the number of columns, the start and end of the number of rows of the zero-set part of the "hui"-shaped two-dimensional window respectively, M and N represent the total number of rows and columns of the two-dimensional image matrix respectively. When ns≤n≤ne and ms≤m≤me, the middle part of the "hui"-shaped window is 0, and the rest part is 1, that is, the peripheral part of the two-dimensional image is retained, and the windowed two-dimensional image can be expressed as follows:

[0092] S W (y,x)=S r (y,x)·W M×N (10)

[0093] Then calculate the contrast of the local image: assuming that the local image contrast after windowing is C W , then its expression is:

[0094]

[0095] where

[0096] Regarding step S20 above, under the criterion of maximizing local image contrast, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated, namely the distance-to-center offset and rotational speed of the uniformly rotating large-sized target. The expression is as follows:

[0097]

[0098] The specific steps are as follows:

[0099] Initialize particle position That is, assign initial values ​​to the parameters to be estimated.

[0100] Constructing the spatially variable phase error compensation function P com (ω,Δy) and resampling factor k resample (ω), then for each iteration, the fast Gaussian grid non-uniform fast Fourier transform in step S14 will be repeated once in step S18, and the error compensation function and resampling factor will also be updated once.

[0101] Assuming the total number of iterations is L, where l represents the number of iterations, the final particle position is obtained when the number of iterations reaches the total number of iterations L. This is the final estimated result, at which point the corresponding local image contrast will also reach its maximum. For example... Figure 2 The diagram shown is a flowchart of parameter estimation and focusing imaging for the large-size rotating target imaging method in the terahertz band described above. In the flowchart, O1 is the PSO parameter estimation algorithm and O2 is the focusing imaging algorithm.

[0102] Through the detailed description of the above embodiments, a complete set of new technical solutions is provided for obtaining longitudinal range center offset and rotational velocity estimates of rotating targets from the echoes of large-sized rotating targets in terahertz broadband radar, as well as focusing range Doppler images. Specifically, the longitudinal range center offset and rotational velocity estimation and focusing imaging method based on the combination of local image maximum contrast and particle swarm optimization, that is, performing KT transform and range-direction non-uniform fast Fourier transform on the original echo, can accurately correct the first-order and second-order range cell migrations. After performing second-order spatially varied phase compensation and azimuth-direction fast Fourier transform, focused imaging of the image can be achieved.

[0103] In the imaging process, the unknown rotation speed and the estimation accuracy of the RCD are crucial to the imaging effect. Since there are two parameters to be estimated, directly using the traditional two-dimensional search under the image optimization criterion would result in a very high computational load and very low efficiency. Therefore, the particle swarm optimization method can significantly improve the parameter estimation efficiency. Furthermore, considering the defocusing characteristics caused by the migration of units with increasing rotation distance—that is, scattering points farther from the rotation center exhibit more pronounced defocusing during imaging—when the scattering intensity of scattering points closer to the rotation center is high, the parameter estimation accuracy of the image optimization-based criterion method, which requires calculating the scattering intensity, will be greatly reduced, easily leading to iteratively erroneous estimation results. Therefore, a new local image maximum contrast method is proposed, which adds a "U"-shaped window to the two-dimensional image, retaining only the more significantly defocused outer portion, thus avoiding the influence of scattering points closer to the rotation center on parameter estimation. Additionally, image contrast is chosen as the criterion for particle swarm optimization parameter search estimation because image entropy is insensitive to changes in the parameters to be estimated, while image contrast becomes more pronounced with changes in the parameters to be estimated, which is more conducive to the accurate estimation of rotation parameters.

[0104] Compared to traditional two-dimensional search, the technical solution of this invention fully leverages the speed and efficiency of particle swarm optimization (PSO) algorithms. It employs the maximum contrast criterion, which, compared to image entropy, shows a more pronounced change with parameter variations, thus improving parameter estimation accuracy and noise robustness. Furthermore, it fully utilizes the image defocusing characteristics caused by the migration of rotational distance units. By adding a "U"-shaped window to the image and using the contrast of the peripheral local image as the parameter estimation criterion, it avoids the problem of decreased parameter estimation accuracy when strong scattering points are located close to the rotation center, thereby improving estimation accuracy to a certain extent.

[0105] In the process of parameter estimation using particle swarm optimization algorithm, the resampling factor and spatially variable phase compensation function are updated and iterated synchronously. By combining parameter estimation with high-precision imaging algorithm, when the iteration terminates and the maximum contrast of the local image is obtained, well-focused two-dimensional imaging results and accurately estimated rotation parameters can be obtained at the same time.

[0106] Compared to traditional range-Doppler imaging which only performs a first-order Taylor expansion, the signal model of the present invention is more accurate and more suitable for high-resolution terahertz radar, especially for imaging large targets.

[0107] This invention is based on a combination of local image maximum contrast and particle swarm optimization for estimating distance-center offset and rotation velocity. Traditional parameter estimation methods based on minimum entropy show little variation in entropy values ​​with parameter changes, hindering high-precision parameter estimation. Furthermore, parameter estimation methods based on global image maximum contrast fail to consider the impact of scattering point intensity on estimation accuracy, especially when scattering points close to the rotation center have high scattering intensity, severely affecting accuracy. By combining KT with non-uniform fast Fourier transform, more precisely focused distance-image imaging can be obtained, facilitating accurate compensation for spatially varied phase errors. Moreover, by iteratively updating the resampling factor and spatially varied phase compensation function simultaneously with the particle swarm optimization algorithm, both well-focused imaging results and accurately estimated rotation parameters can be achieved.

[0108] In some embodiments, to more intuitively demonstrate the technical effects of the present invention, some simulation experimental verification examples are also provided here. For example, taking a broadband radar system with a carrier frequency of 220 GHz and a bandwidth of 10 GHz as an example, using a satellite model made of a good conductor material as the target, with a maximum target size range of [-1.8m, 1.8m], by processing the method steps proposed in the above embodiments, focused imaging is gradually achieved, and the distance center offset and rotation speed are finally estimated, illustrating the correctness and effectiveness of the technical solution of the present invention. In the electromagnetic simulation experiment, the uniform rotation speed of the target is set to 18° / s, and the RCD is set to 0m.

[0109] The range Doppler imaging obtained by directly performing a two-dimensional fast Fourier transform on the baseband echo signal is as follows: Figure 3 As shown in the imaging results, severe defocusing occurred due to significant cross-range cell migration of the target during rotation.

[0110] The two-dimensional imaging results and local magnified images obtained after performing KT transform on the baseband echo signal are as follows: Figure 4 As shown, (a) is the imaging result after KT transformation, and (b) is a magnified view of the imaging result after KT transformation. It can be seen that after KT transformation, the most serious first-order cross-range cell migration is corrected, and the cross-range cell migration phenomenon is greatly improved. The two-dimensional imaging result is also greatly improved compared with range Doppler imaging. However, from the magnified view, there is still defocusing in the azimuth direction, and the larger the longitudinal coordinate, the more severe the defocusing. Therefore, second-order cross-range cell migration correction and spatially varied phase compensation are still needed. The key parameters for second-order cross-range cell migration correction and spatially varied phase compensation are the rotation speed and the longitudinal distance center offset. Therefore, the particle swarm optimization method is used for parameter estimation and extraction.

[0111] According to the method proposed in the foregoing invention, a two-dimensional window is first constructed as follows Figure 5 As shown in the figure, wherein (a) is a two-dimensional window, the middle rectangular part is 0, and the rest part is 1, that is, the covered part of the "loop"-shaped window is 1, and (b) is the range-Doppler imaging result after windowing.

[0112] The image contrast of the windowed imaging result is calculated, and particle swarm optimization parameter estimation is carried out with the maximum contrast as the criterion. The estimation results are 17.95° / s and 0.0027m, and the obtained well-focused imaging result and partial enlarged view are as follows Figure 6 As shown in the figure, (a) is the imaging result obtained by applying the above-mentioned imaging method for large-sized rotating targets in the terahertz band of the present invention, and (b) is a partial enlarged view of the imaging result. It can be seen that after the second-order cross-range unit migration correction and accurate space-variant phase compensation, compared with the imaging only using the KT transform, the focusing effect at the position far from the rotation center is greatly improved. On the contrary, if the particle swarm optimization parameter estimation is carried out with the maximum contrast of the global image as the criterion, the obtained estimation results are 10.92° / s and -0.44m, and the parameter estimation error is very large. The simulation results prove the correctness and effectiveness of the proposed algorithm.

[0113] It should be understood that although Figure 1 The respective steps in the flowchart of are displayed in sequence according to the indication of the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and these steps may be executed in other orders. Moreover Figure 1 At least a part of the steps may include a plurality of sub-steps or a plurality of stages. These sub-steps or stages are not necessarily executed and completed at the same moment, but may be executed at different moments. The execution order of these sub-steps or stages is not necessarily sequential, but may be executed in turn or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0114] In one embodiment, as in Figure 7As shown, a terahertz band large-size rotating target imaging system 100 is provided, including an echo acquisition module 11, a migration cancellation module 13, a spatially varied compensation module 15, a windowing module 17, and an imaging module 19. The echo acquisition module 11 acquires inverse synthetic aperture radar echo data of a uniformly rotating large-size target and performs line-frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large-size target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle. The migration cancellation module 13 performs KT transform on the baseband echo signal to eliminate first-order range-crossing migration and uses fast Gaussian grid non-uniform fast Fourier transform to eliminate second-order range-crossing migration. The spatially varied compensation module 15 constructs a spatially varied phase compensation function and performs phase error compensation on the baseband echo signal after eliminating range-crossing migration, then performs an azimuth-directed fast Fourier transform to obtain a two-dimensional imaging result. The windowing module 17 is used to construct a U-shaped two-dimensional window, apply a window to the two-dimensional imaging result, and calculate the image contrast of the local image of the two-dimensional imaging result. The imaging module 19 is used to iteratively estimate the two parameters to be estimated using a particle swarm optimization algorithm under the criterion of maximizing the local image contrast. When the iteration terminates, the corresponding local image contrast reaches its maximum, and the final estimated two parameters to be estimated and the well-focused imaging result of a uniformly rotating large target are obtained. The two parameters to be estimated include the distance-to-center offset and rotation speed of the uniformly rotating large target.

[0115] The aforementioned terahertz band large-size rotating target imaging system 100, through a design concept based on local image maximum contrast and particle swarm optimization for range center offset and rotational speed estimation, as well as focused imaging, fully leverages the high efficiency of the particle swarm optimization algorithm. Combining the defocusing characteristics caused by target rotation, it proposes a rotation parameter estimation method based on local image maximum contrast. After performing KT correction on the baseband echo signal for first-order cross-range cell migration, second-order cross-range cell migration correction is performed using the estimated parameters and a fast Gaussian grid non-uniform fast Fourier transform. Then, second-order spatially variable phase compensation is performed, followed by azimuth-direction fast Fourier transform. After two-dimensional windowing, the particle swarm optimization algorithm is used to iteratively estimate the two parameters. When the iteration terminates, the corresponding local image contrast reaches its maximum, ultimately yielding a well-focused range Doppler imaging result.

[0116] Compared to traditional methods, the proposed technical solution accurately corrects the migration of first-order and second-order cross-range units. After performing second-order spatially variable phase compensation and azimuth fast Fourier transform, the image can be focused and imaged. Finally, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated, which greatly reduces the amount of computation in the search. It has the characteristics of simplicity and efficiency, and ultimately achieves the goal of effectively improving the performance of large-size target imaging in the terahertz band.

[0117] In one embodiment, during the process of KT transforming the baseband echo signal to eliminate first-order cross-range cell migration, the migration cancellation module 13 can specifically be used to construct a new slow-time variable and then approximate it with the baseband echo signal to obtain a baseband echo signal with KT transform-eliminated first-order cross-range cell migration; the slow-time variable τ m for:

[0118]

[0119] Among them, f c For radar carrier frequency, t m For azimuth slow time, γ is the frequency modulation. fr represents the fast time interval, and fr represents the fast time frequency.

[0120] In one embodiment, the migration elimination module 13 defines a resampling factor k during the process of eliminating second-order cross-range cell migration using a fast Gaussian grid non-uniform fast Fourier transform. resample for:

[0121]

[0122] Where ω is the target's rotational speed, τ m To construct new slow-time variables.

[0123] For specific limitations regarding the terahertz band large-size rotating target imaging system 100, please refer to the corresponding limitations of the terahertz band large-size rotating target imaging method above, which will not be repeated here. Each module in the aforementioned terahertz band large-size rotating target imaging system 100 can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of a device with data processing capabilities, or stored in software in the memory of the aforementioned device, so that the processor can call and execute the operations corresponding to each module. The aforementioned device can be, but is not limited to, various types of radar data calculation and processing devices already existing in the art.

[0124] In one embodiment, a target imaging device is also provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: acquiring inverse synthetic aperture radar echo data of a uniformly rotating large-size target and performing line frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large-size target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle; performing KT transform on the baseband echo signal to eliminate first-order range cell migration, and using fast Gaussian grid non-uniform fast Fourier transform to eliminate second-order range cell migration; constructing a spatially variable phase compensation function and applying it to the elimination... After phase error compensation, the baseband echo signal after the range cell migration is subjected to azimuth fast Fourier transform to obtain the two-dimensional imaging result. After constructing a "U"-shaped two-dimensional window, the two-dimensional imaging result is windowed and the image contrast of the local image of the two-dimensional imaging result is calculated. Under the criterion of maximizing the local image contrast, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated. When the iteration terminates, the corresponding local image contrast reaches the maximum, and the final estimated two parameters to be estimated and the well-focused imaging result of the uniformly rotating large target are obtained. The two parameters to be estimated include the distance center offset and rotation speed of the uniformly rotating large target.

[0125] It is understood that, in addition to the memory and processor mentioned above, the target imaging device also includes other hardware and software components not listed in this specification. The specific components can be determined according to the model of the specific filter device in different application scenarios, and will not be listed and described in detail in this specification.

[0126] In one embodiment, when the processor executes the computer program, it can also implement the steps or sub-steps added in the various embodiments of the above-described terahertz band large-size rotating target imaging method.

[0127] In one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored. When executed by a processor, the computer program performs the following steps: acquiring inverse synthetic aperture radar echo data of a uniformly rotating large-size target and performing line frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large-size target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle; performing KT transform on the baseband echo signal to eliminate first-order range-crossing cell migration, and using fast Gaussian grid non-uniform fast Fourier transform to eliminate second-order range-crossing cell migration; constructing a spatially variable phase compensation function and applying it to eliminate range-crossing cell migration. After phase error compensation, the baseband echo signal after cell migration is subjected to azimuth fast Fourier transform to obtain a two-dimensional imaging result. After constructing a "U"-shaped two-dimensional window, the two-dimensional imaging result is windowed and the image contrast of the local image of the two-dimensional imaging result is calculated. Under the criterion of maximizing the local image contrast, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated. When the iteration terminates, the corresponding local image contrast reaches the maximum, and the final estimated two parameters to be estimated and the well-focused imaging result of the uniformly rotating large target are obtained. The two parameters to be estimated include the distance-center offset and rotation speed of the uniformly rotating large target.

[0128] In one embodiment, when the computer program is executed by the processor, it can also implement the steps or sub-steps added in the various embodiments of the above-described terahertz band large-size rotating target imaging method.

[0129] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), memory bus DRAM (RDRAM), and interface DRAM (DRDRAM), etc.

[0130] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0131] The above embodiments merely illustrate several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, all of which fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for imaging large-size rotating targets in the terahertz band, characterized in that, Including the following steps: The inverse synthetic aperture radar echo data of a uniformly rotating large target is acquired and subjected to line-frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle. The baseband echo signal is subjected to KT transform to eliminate first-order cross-range cell migration, and fast Gaussian grid non-uniform fast Fourier transform is used to eliminate second-order cross-range cell migration. After constructing a spatially variable phase compensation function and compensating for the phase error of the baseband echo signal after eliminating the migration of the distance cell, a fast Fourier transform in the azimuth direction is performed to obtain the two-dimensional imaging result. After constructing a "回"-shaped two-dimensional window, windowing the two-dimensional imaging result and calculating the image contrast of the local image of the two-dimensional imaging result; the "回"-shaped two-dimensional window is , where,( m , n [[ID=6 represent the row and column indices of the two-dimensional image matrix respectively, the starting point of the number of columns of the zero-set part of the "回"-shaped two-dimensional window is ns , the end point is ne , the starting point of the number of rows is ms and the end point is me , M and N represent the total number of rows and columns of the two-dimensional image matrix respectively; Under the criterion of maximizing local image contrast, the particle swarm optimization algorithm is used to iteratively estimate the two parameters to be estimated. When the iteration terminates, the local image contrast reaches its maximum, and the final estimated two parameters and the well-focused imaging result of the uniformly rotating large target are obtained. The two parameters to be estimated include the distance-to-center offset and rotation speed of the uniformly rotating large target.

2. The terahertz band large-size rotating target imaging method according to claim 1, characterized in that, The process of eliminating first-order cross-range cell migration by performing KT transform on the baseband echo signal includes: A new slow-time variable is constructed and then input into the baseband echo signal for approximation processing to obtain the baseband echo signal with KT transform eliminating first-order cross-range cell migration; the slow-time variable for: in, For radar carrier frequency, For location, slow time, To adjust the frequency, For the sake of speed, For fast time frequency.

3. The terahertz band large-size rotating target imaging method according to claim 1 or 2, characterized in that, During the process of eliminating second-order cross-range cell migration using a fast Gaussian grid non-uniform fast Fourier transform, a resampling factor is defined. for: in, The rotational speed of the target To construct new slow-time variables.

4. A large-size rotating target imaging system in the terahertz band, characterized in that, include: The echo acquisition module is used to acquire the inverse synthetic aperture radar echo data of a uniformly rotating large target and perform line frequency modulation processing to obtain the baseband echo signal of the uniformly rotating large target; the baseband echo signal includes a second-order Taylor expansion of the sine and cosine of the accumulation angle. The migration cancellation module is used to perform KT transform on the baseband echo signal to eliminate first-order cross-range cell migration, and to use fast Gaussian grid non-uniform fast Fourier transform to eliminate second-order cross-range cell migration. The spatial variation compensation module is used to construct the spatial variation phase compensation function and perform phase error compensation on the baseband echo signal after eliminating the migration of the distance cell, and then perform azimuth fast Fourier transform to obtain the two-dimensional imaging result. A windowing module, configured to construct a "hui"-shaped two-dimensional window, apply windowing to the two-dimensional imaging result, and calculate the image contrast of a local image of the two-dimensional imaging result; the "hui"-shaped two-dimensional window is , where,( m , n [[ID=6)]] respectively represent the row and column indices of the two-dimensional image matrix, the starting point of the number of columns of the zero-set part of the "hui"-shaped two-dimensional window is ns , the end point is ne , the starting point of the number of rows is ms and the end point is me , M and N respectively represent the total number of rows and columns of the two-dimensional image matrix; The imaging module is used to iteratively estimate two parameters to be estimated using a particle swarm optimization algorithm under the criterion of maximizing local image contrast. When the iteration terminates, the local image contrast reaches its maximum, and the final estimated parameters and the well-focused imaging result of the uniformly rotating large target are obtained. The two parameters to be estimated include the distance-to-center offset and rotation speed of the uniformly rotating large target.

5. The terahertz band large-size rotating target imaging system according to claim 4, characterized in that, The process of eliminating first-order cross-range cell migration by performing KT transform on the baseband echo signal includes: A new slow-time variable is constructed and then input into the baseband echo signal for approximation processing to obtain the baseband echo signal with KT transform eliminating first-order cross-range cell migration; the slow-time variable for: in, For radar carrier frequency, For location, slow time, To adjust the frequency, For the sake of speed, For fast time frequency.

6. The terahertz band large-size rotating target imaging system according to claim 4 or 5, characterized in that, During the process of eliminating second-order cross-range cell migration using a fast Gaussian grid non-uniform fast Fourier transform, a resampling factor is defined. for: in, The rotational speed of the target To construct new slow-time variables.

7. A target imaging device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the terahertz band large-size rotating target imaging method according to any one of claims 1 to 3.

Images