Motion compensation imaging method based on Radon transformation and SWO algorithm
By combining Radon transform and SWO algorithm, motion error of ISAL is quickly estimated and compensated, solving the accuracy and efficiency problem in ISAL imaging and achieving high-quality and efficient motion error compensation effect.
Patent Information
- Application Number
- CN202610167242.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-05
- Publication Date
- 2026-05-12
AI Technical Summary
Among the existing ISAL motion error compensation methods, the cross-correlation method has low accuracy and the PGA algorithm is easily affected by noise, while the direct optimization method has too long an iteration time and is prone to getting trapped in local extrema, making it difficult to balance imaging quality and efficiency.
A motion error compensation method for inverse synthetic aperture lidar is proposed, which employs Radon transform combined with SWO algorithm. The method uses Radon integral transform to quickly estimate the coarse radial velocity, constructs a local constraint search space, and uses SWO algorithm to perform fine estimation of multiple parameters, constructs a phase compensation function for imaging.
It significantly improves image quality, increases image contrast, reduces entropy, shortens imaging time, overcomes local extremum traps, and achieves high-precision and high-efficiency motion error compensation.
Smart Images
Figure CN122017881A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of lidar signal processing, specifically to a joint compensation algorithm for ISAL maneuvering targets based on Radon transform and SWO (SpiderWasp Optimization) algorithm. Background Technology
[0002] ISAL (Inverse Synthetic Aperture Laser) operates at extremely short wavelengths (typically on the micrometer scale). Compared to traditional microwave radar, it can provide more detailed target features (such as texture and shape details), making it of significant strategic value in defense and aerospace fields. However, the extremely short wavelength makes ISAL imaging highly sensitive to the relative motion between the target and the radar. Even minute non-uniform movements can introduce huge phase errors into the echo signal, leading to defocusing or complete image failure. Therefore, how to quickly and accurately estimate and compensate for motion errors from the echo data is the core bottleneck of ISAL technology.
[0003] Currently, existing ISAL motion error compensation methods are mainly divided into two categories:
[0004] 1) Cross-correlation method combined with PGA algorithm: Cross-correlation method is used for simple envelope alignment, and then PGA (phase gradient autofocus) algorithm is used for phase compensation. This method has low accuracy in ISAL, the error caused by separate compensation will affect the image quality, and the PGA algorithm is easily affected by the distribution of strong scattering points.
[0005] 2) Direct Optimization Parameter Estimation Method: This method uses swarm intelligence algorithms (such as PSO) to directly estimate the target parameters. Although it can obtain high-quality images, it suffers from a large search space, excessively long iteration time, and a tendency to get trapped in local optima.
[0006] Therefore, there is an urgent need for an ISAL motion compensation method that can both ensure imaging accuracy and improve convergence speed while avoiding local extrema. Summary of the Invention
[0007] The technical problem to be solved by this invention is that the existing cross-correlation method plus PGA method has low compensation accuracy, and the direct optimization method has too long iteration time and is prone to getting trapped in local extrema. The invention proposes an inverse synthetic aperture lidar motion error compensation method using Radon transform combined with SWO algorithm.
[0008] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a motion-compensated imaging method based on Radon transform and SWO algorithm, comprising the following steps: Step S1: Perform deskewing reception processing and range pulse compression on the echo signal of ISAL (Inverse Synthetic Aperture LiDAR) to obtain a one-dimensional range image signal containing motion errors; Step S2: Perform Radon integral transform on the signal amplitude diagram output in step S1, search for the maximum energy peak in the Radon domain, and obtain the rough radial velocity of the target. Step S3: Based on the rough radial velocity obtained in step S2, construct a local constraint search space and set the velocity search range and acceleration search range; Step S4: Within the local constraint search space defined in step S3, initialize the SWO algorithm population and iteratively search for the optimal velocity and acceleration estimates; Step S5: Construct a phase compensation function using the optimal velocity and acceleration estimates to compensate the original data, and perform azimuth FFT processing to obtain the ISAL image.
[0009] Preferably, in step S1, the one-dimensional range image signal includes motion error. Represented as:
[0010] In the formula, The signal amplitude after pulse compression. To save time, For absolute time, For slow time, For bandwidth, At the speed of light, The distance to the target. It is the imaginary unit.
[0011] Preferably, in step S2, the formula for calculating the Radon integral transform is:
[0012] in, This is the amplitude diagram of a one-dimensional distance image signal. These are Radon domain coordinates.
[0013] Preferably, step S2 further includes: binarizing the distance-compressed image to suppress background noise and highlight the target trajectory skeleton, followed by Radon transform; in the Radon domain The system traverses the search for the coordinates of the maximum energy peak and calculates the approximate radial velocity using the following formula. :
[0014] in, This is the pulse repetition frequency.
[0015] Preferably, in step S3, the strategy for constructing the local constraint search space is as follows: the velocity search range is set to... ,in The approximate radial velocity obtained in step S2, The preset velocity fluctuation threshold is used; the acceleration search range is set to... .
[0016] Preferably, the speed fluctuation threshold The acceleration search range is set to 5 m / s. .
[0017] Preferably, in step S4, the SWO algorithm uses image contrast as the fitness function. The calculation formula is as follows:
[0018] in, Represents the grayscale value of the image. and These represent the dimensions of the image.
[0019] Preferably, the specific process of step S4 includes: Randomly generated within the local constraint search space Individuals containing velocity and acceleration information are used as the initial population; Calculate the fitness value for each individual; The individual's position is updated based on its fitness value until a preset termination condition is met; Output the velocity estimate corresponding to the maximum fitness value and acceleration estimates .
[0020] Preferably, in step S5, the phase compensation function The expression is:
[0021] In the formula, For frequency modulation slope, The wavelength of light. It is the imaginary unit.
[0022] In a second aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.
[0023] The beneficial effects of this invention are: 1) This invention solves the problem of image defocusing caused by minute motion errors in ISAL (In-Situ Image Alignment), significantly improving image quality. Existing cross-correlation methods suffer from sub-resolution cell alignment errors, which introduce drastic periodic phase jumps, causing the PGA algorithm to fail and resulting in severe azimuth ghosting and defocusing. This invention uses the SWO algorithm for multi-parameter fine estimation, accurately capturing and compensating for residual minute velocity and acceleration errors. Experimental data shows that the image contrast generated by this method is improved from 1.9794 to 5.9067 compared to traditional methods, and the image entropy value is reduced from 8.5018 to 6.7631. The final imaging result shows a clear aircraft outline, well-focused scattering points, no obvious sidelobe interference, and effective correction of range curvature.
[0024] 2) This invention overcomes the efficiency bottleneck of traditional optimization algorithms, significantly shortening imaging time. When using algorithms like Particle Swarm Optimization (PSO) for blind search, the large search space leads to extremely long iteration times (e.g., 11.43 seconds in the experiment). This invention innovatively proposes a two-layer architecture of "coarse estimation guiding fine search": First, utilizing the integral gain characteristic of the Radon transform, the target trajectory skeleton is quickly extracted from the noise and a coarse radial velocity is locked; based on this, a local constraint search space is constructed, transforming the "blind optimization" in high-dimensional space into a small-scale "fixed-point optimization." Experiments show that this invention shortens the algorithm iteration time to 4.59 seconds, improving efficiency by approximately 60%, meeting the needs of rapid ISAL imaging.
[0025] 3) Overcoming the local optima trap and enhancing the algorithm's robustness: Traditional PGA algorithms heavily rely on strong scattering points and are easily affected by noise; while conventional optimization algorithms are prone to getting trapped in local optima. The SWO algorithm used in this invention has strong global optimization capabilities, and combined with the accurate initial value guidance provided by the Radon transform, it effectively avoids the algorithm getting trapped in local optima. Whether in low signal-to-noise ratio environments or for complex moving targets, this invention can stably converge to the globally optimal parameters, ensuring the reliability of the algorithm in practical applications.
[0026] 4) This invention achieves an organic combination of signal processing and intelligent optimization. It combines the classic Radon transform (for geometric feature extraction) with modern swarm intelligence algorithms (for parameter optimization). Leveraging the advantages of the Radon transform in handling linear distance travel, it solves the initial value sensitivity problem of the SWO algorithm; simultaneously, it utilizes the high-precision optimization capability of the SWO algorithm to compensate for the limited accuracy of the Radon transform. This joint processing mechanism fully utilizes the advantages of both, providing a novel and efficient solution for large-scale motion error compensation in ISAL. Attached Figure Description
[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0028] Figure 1 This is an algorithm flowchart of the method of the present invention.
[0029] Figure 2 This is a comparison chart of the imaging effects of the method of the present invention and the prior art. Detailed Implementation
[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0031] Example 1: Overall Algorithm Flow This embodiment provides a joint compensation imaging method for ISAL maneuvering targets based on Radon transform and SWO algorithm. Its core lies in employing a two-layer optimization strategy of "coarse estimation guiding fine search." For example... Figure 1 As shown, the specific processing flow includes the following steps: Step S1: Echo signal model construction and preprocessing First, the raw echo signal received by ISAL undergoes de-chirp reception processing, followed by range-compression. This step converts the time-domain echo into a range-frequency domain signal, making the range movement of the target scattering point in the two-dimensional plane (range-azimuth plane) appear as an observable tilted trajectory. The final output is a one-dimensional range image signal that includes motion errors. .
[0032] Step S2: Coarse velocity estimation based on Radon transform To address the linear range travel characteristics of the target caused by translational motion in the ISAL echo, a Radon integral transform is performed on the signal amplitude map output in step S1. Specifically, the range-compressed image is first binarized to effectively suppress background noise and highlight the target's trajectory skeleton. Then, the image is projected into the Radon domain (parameter space), and the peak point with the highest energy is searched throughout the Radon domain using the integral gain characteristic of the Radon transform. The coordinates of this peak point directly reflect the trajectory's tilt angle, and the target's approximate radial velocity can be derived from geometric relationships. .
[0033] Step S3: Construct the local constraint search space To address the issues of slow convergence and susceptibility to local optima caused by blind searching over a wide range in global optimization algorithms, this invention utilizes the coarse radial velocity obtained in step S2. Apply physical constraints to the search space. Set a small velocity fluctuation range. The search range for speed is locked within Within this framework, the search range for acceleration is set based on the target's potential maneuvers. This constraint transforms the originally high-dimensional, global blind optimization into a small-scale "fixed-point optimization," thus laying the foundation for subsequent refined estimation.
[0034] Step S4: Fine-grained multi-parameter estimation based on the SWO algorithm Within the local constraint space defined in step S3, initialize the SWO algorithm.
[0035] (1) Population initialization: randomly generated Each of the following individuals contains a set of velocities to be estimated. and acceleration parameter.
[0036] (2) Fitness calculation: Image contrast (in variance form) is used as the fitness function to evaluate focus quality. The current parameters are substituted into the phase compensation formula to correct the image, and the contrast of the corrected image is calculated. The higher the contrast, the clearer the image focus.
[0037] (3) Iterative Update: Individual spider wasps update their positions by mimicking the optimization behavior of spider wasps in nature, based on their fitness values. After multiple iterations, when a preset termination condition is met (such as reaching the maximum number of iterations or a fitness threshold), the optimal velocity estimate corresponding to the maximum fitness value is output. and acceleration estimates The SWO algorithm has a powerful global optimization capability, and can accurately estimate the residual small velocity and acceleration errors that the Radon transform fails to eliminate.
[0038] Step S5: Phase Compensation and Imaging Using the optimal parameters output in step S4 and A full-aperture phase compensation function, including a fast-time envelope term and a slow-time phase term, is constructed. Precise phase compensation is then performed on the original one-dimensional range image data to eliminate range curvature and azimuth defocus. Finally, an azimuth FFT (Fast Fourier Transform) is performed to obtain a high-resolution, highly focused ISAL image.
[0039] Example 2: Key Mathematical Model and Algorithm Details This embodiment elaborates on the key mathematical models in the above process to ensure the feasibility of the technical solution.
[0040] 1. Signal Model In step S1, after deskewing reception and range pulse compression, the one-dimensional range image signal containing motion error is obtained. Represented as:
[0041] In the formula: The signal amplitude after pulse compression; Indicates fast time; Indicates slow time; For signal bandwidth; The speed of light; For the goal in slow time The distance and journey of time; It is the imaginary unit.
[0042] 2. Radon Transform and Velocity Calculation In step S2, the amplitude diagram of the signal is... The Radon integral transform is performed as follows:
[0043] in, Let be the impulse function. Polar radius, It is the polar angle. In the Radon domain After finding the maximum energy peak point, its angle is extracted. The rough radial velocity is calculated using the following formula. :
[0044] In the formula, This is the pulse repetition frequency.
[0045] 3. Fitness function (image contrast) In step S4, the optimization objective of the SWO algorithm is to maximize image contrast. This invention uses contrast, defined in variance form, as the fitness function. The calculation formula is:
[0046] In the formula: This indicates the image at the pixel level after compensation. The grayscale value at that location; and These represent the dimensions of the image in the range and azimuth directions, respectively. This formula calculates the variance of image intensity. The larger the variance, the more obvious the difference between light and dark in the image, and the better the focusing effect.
[0047] 4. Full aperture phase compensation function In step S5, the phase compensation function is constructed. It includes compensation for both linear (velocity) and quadratic (acceleration) terms:
[0048] In the formula: This is the frequency modulation slope; The wavelength is the laser wavelength. The first half of this function is used to correct for distance curvature (fast-time envelope term), and the second half is used to correct for phase error (slow-time phase term).
[0049] Example 3: Simulation Experiment and Effect Verification To verify the effectiveness and superiority of the method of the present invention, a simulation experiment was set up and compared with the prior art.
[0050] 1. Simulation Condition Settings A scatter plot model of the aircraft was established for simulation. The main parameters are shown in the table below: Light wavelength: 1550nm Laser modulation bandwidth: 4GHz Pulse repetition frequency: 50kHz Single pulse duration: 0.2 μs Synthetic aperture pulse count: 512 Target distance: 3km Target motion parameters: translational velocity 35.5 m / s, acceleration 5 m / s² 2. Algorithm parameter configuration Radon coarse estimation stage: Based on the detected peak values, a rough radial velocity is calculated. .
[0051] Search space constraint: Set a velocity fluctuation threshold That is, the speed search range is Set the acceleration search range to .
[0052] SWO algorithm settings: population size set to 20, maximum number of iterations set to 50.
[0053] 3. Experimental Results Analysis: The method of this invention is compared with "traditional cross-correlation method + PGA" and "direct PSO algorithm without Radon guidance". The results are as follows: (1) Subjective evaluation of image quality: Traditional cross-correlation + PGA: Due to the extremely short wavelength of ISAL, the alignment error of the cross-correlation method causes the PGA algorithm to fail, resulting in severe azimuth ghosting and defocus in the imaging results, making the aircraft outline blurry and difficult to distinguish.
[0054] Direct PSO algorithm: It can achieve focusing and produces relatively good image quality.
[0055] The method of this invention: After coarse estimation by Radon transform, the distance curvature is effectively corrected; after fine compensation by SWO algorithm, the aircraft outline is clear, the scattering points are well focused, there is no obvious sidelobe interference, and the imaging quality reaches the optimal level.
[0056] (2) Comparison of objective indicators:
[0057] Contrast and Entropy: The contrast (5.9067) and entropy (6.7631) of the method of this invention are basically on par with the direct PSO algorithm and far superior to the traditional method, which proves the superiority of the invention in imaging accuracy.
[0058] Operational efficiency: The iteration time of the method in this invention is only 4.59 seconds, compared with 11.43 seconds for the direct PSO algorithm, which improves the computational efficiency by about 60%. This is due to the Radon transform quickly determining the "neighborhood" of the optimal solution, avoiding the search in the invalid space.
[0059] In summary, this invention, through the organic combination of Radon transform and SWO algorithm, significantly reduces computation time while ensuring high-precision imaging, thus solving the problem of balancing accuracy and efficiency in ISAL maneuvering target imaging.
[0060] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0061] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A motion-compensated imaging method based on Radon transform and SWO algorithm, characterized in that, Includes the following steps: Step S1: Perform deskewing reception processing and range pulse compression on the ISAL echo signal to obtain a one-dimensional range image signal containing motion errors; Step S2: Perform Radon integral transform on the signal amplitude diagram output in step S1, search for the maximum energy peak in the Radon domain, and obtain the rough radial velocity of the target. Step S3: Based on the rough radial velocity obtained in step S2, construct a local constraint search space and set the velocity search range and acceleration search range; Step S4: Within the local constraint search space defined in step S3, initialize the SWO algorithm population and iteratively search for the optimal velocity and acceleration estimates; Step S5: Construct a phase compensation function using the optimal velocity and acceleration estimates to compensate the original data, and perform azimuth FFT processing to obtain the ISAL image.
2. The method according to claim 1, characterized in that, In step S1, the one-dimensional range image signal containing motion error Represented as: In the formula, The signal amplitude after pulse compression. To save time, For absolute time, For slow time, For bandwidth, At the speed of light, The distance to the target. It is the imaginary unit.
3. The method according to claim 1, characterized in that, In step S2, the formula for calculating the Radon integral transform is: in, This is the amplitude diagram of a one-dimensional distance image signal. Let be the impulse function. Polar radius, Polar angle, These are Radon domain coordinates.
4. The method as described in claim 3, characterized in that, Step S2 further includes: binarizing the distance-compressed image to suppress background noise and highlight the target trajectory skeleton, followed by Radon transform; in the Radon domain The system traverses the search for the coordinates of the maximum energy peak and calculates the approximate radial velocity using the following formula. : in, is the pulse repetition frequency.
5. The method according to claim 1, characterized in that, In step S3, the strategy for constructing the local constraint search space is as follows: the velocity search range is set to... ,in The approximate radial velocity obtained in step S2, The preset speed fluctuation threshold is used; the acceleration search range is set to... .
6. The method according to claim 5, characterized in that, The speed fluctuation threshold The acceleration search range is set to 5 m / s. .
7. The method according to claim 1, characterized in that, In step S4, the SWO algorithm uses image contrast as the fitness function. The calculation formula is as follows: in, Represents the grayscale value of the image. and These represent the dimensions of the image.
8. The method according to claim 7, characterized in that, The specific process of step S4 includes: Randomly generated within the local constraint search space Individuals containing velocity and acceleration information are used as the initial population; Calculate the fitness value for each individual; The individual's position is updated based on its fitness value until a preset termination condition is met; Output the velocity estimate corresponding to the maximum fitness value and acceleration estimates .
9. The method according to claim 1, characterized in that, In step S5, the phase compensation function The expression is: In the formula, For frequency modulation slope, The wavelength of light. It is the imaginary unit.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 9.