Anisotropic full-link degradation simulation method and system for high-altitude high-speed optical imaging
By introducing an anisotropic spatial frequency calculation model and a low-frequency tilt compensation plane, the problems of anisotropic ambiguity and imprecise multi-physics coupling in high-altitude and high-speed optical imaging simulation are solved, and high-precision simulation results are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HEFEI INSTITUTE OF PHYSICAL SCIENCE CHINESE ACADEMY OF SCIENCES
- Filing Date
- 2026-05-15
- Publication Date
- 2026-06-12
AI Technical Summary
Existing high-altitude high-speed optical imaging simulation methods fail to accurately reflect the anisotropic ambiguity caused by aerodynamic flow field stretching, lack low-frequency aerodynamic jitter compensation, and have inaccurate multi-physics coupling, resulting in inaccurate simulation results.
An anisotropic spatial frequency calculation model based on flow field angle and flow field stretching ratio is introduced. A low-frequency tilt compensation plane is designed, and the phases of aerodynamic shock wave and shear layer, system defocus phase and anisotropic turbulent target phase are superimposed to achieve multi-physics field coupling simulation.
It improves the accuracy of simulation results, accurately reproduces the directional stretching and blurring of the light wavefront caused by the high-altitude and high-speed flow field environment, enhances the accuracy of the aerodynamic jitter model, and ensures that the end-to-end degradation image conforms to the real optical physics laws.
Smart Images
Figure CN122194474A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of photoelectric imaging, and in particular to an anisotropic end-to-end degradation simulation method and system for high-altitude high-speed optical imaging. Background Technology
[0002] In related technologies, existing optical imaging simulation methods based on atmospheric turbulence typically use standard Kolmogorov or standard von Kármán spectra to construct power spectral density, generate a static random phase screen using fast Fourier transform, then convert the static random phase screen into a point spread function, and directly convolve and superimpose it with the target image to obtain the simulated image.
[0003] This technique has the following problems and drawbacks: First, when a high-altitude, high-speed aircraft is in flight, the surrounding air forms a strong shear flow field, and the turbulent vortices are stretched, exhibiting strong anisotropy. This results in directional stretching and blurring of the image, and the isotropic assumption in related techniques cannot accurately reflect this flow field stretching effect. Second, the traditional Fast Fourier Transform method suffers from a lack of low-frequency energy when generating the phase screen, causing the simulated image to lack the low-frequency aerodynamic jitter phenomenon unique to high-altitude platforms. Third, related techniques often confuse atmospheric phase distortion with aerosol amplitude attenuation without performing rigorous physical decoupling, leading to inaccurate multiphysics composite simulations.
[0004] In summary, existing high-altitude, high-speed vehicle imaging simulations suffer from several problems, including failure to consider anisotropic ambiguity caused by aerodynamic flow field stretching, lack of low-frequency aerodynamic jitter compensation, and imprecise multi-physics coupling, leading to inaccurate simulation results. Summary of the Invention
[0005] Based on this, it is necessary to provide an anisotropic end-to-end degradation simulation method and system for high-altitude high-speed optical imaging to address the aforementioned technical problems. This method introduces an anisotropic spatial frequency calculation model based on the flow field angle and the flow field stretching ratio, which can accurately reproduce the directional stretching ambiguity caused by the high-altitude high-speed flow field environment on the light wavefront. This breaks the limitation of related technologies that can only simulate isotropic ambiguity. Then, a low-frequency tilt compensation plane is designed. After superimposing this compensation plane, the system's point spread function exhibits a random centroid shift consistent with the actual target range observation height, greatly improving the accuracy of the aerodynamic jitter model. Finally, the aerodynamic shock wave is physically superimposed with the shear layer phase, the system defocus phase, and the anisotropic turbulent target phase to achieve multi-physics coupling, thereby improving the accuracy of the simulation results.
[0006] Firstly, this application provides a method for simulating the anisotropic end-to-end degradation of high-altitude, high-speed optical imaging, including: Acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field image; The anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase are calculated. The aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase are physically superimposed to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. Generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
[0007] In one embodiment, the calculation process for the initial phase of anisotropic turbulence includes: Obtain atmospheric coherence length, flow field stretching ratio, and flow field angle; A two-dimensional basic spatial frequency grid is constructed, and the coordinates of the two-dimensional basic spatial frequency grid are rotated according to the flow field angle to generate a rotated coordinate system. The square of the anisotropic spatial frequency is calculated based on the rotated coordinate system and the flow field stretching ratio. Anisotropic power spectral density is calculated based on the von Kármán spectral model using the atmospheric coherence length and the square of the anisotropic spatial frequency. Random noise is added to the anisotropic power spectral density, and then an inverse Fourier transform is performed to generate the initial phase of the anisotropic turbulence.
[0008] In one embodiment, the process of constructing the low-frequency tilt compensation plane includes: Generate random tilt coefficients; A first two-dimensional spatial coordinate grid is constructed, and the phase value of each coordinate point in the first two-dimensional spatial coordinate grid is adjusted based on a random tilt coefficient to generate a low-frequency tilt compensation plane.
[0009] In one embodiment, the calculation process for the phase of the aerodynamic shock wave and the shear layer includes: Obtain the flow field angle; Construct a second two-dimensional spatial coordinate network, and rotate the second two-dimensional spatial coordinate network according to the flow field angle; Based on the mathematical model of bow shock wave, the phase distribution of the main shock wave is generated according to the rotated second two-dimensional spatial coordinate network; Shear layer noise is added to the phase distribution of the main shock wave and then multiplied by the shock wave intensity parameter to generate the phase of the aerodynamic shock wave and the shear layer.
[0010] In one embodiment, generating a full-link degradation image based on a geometric field-of-view image and a composite phase plane includes: Generate a global point spread function based on the composite phase plane; A blurred image is generated by convolving the geometric field of view image with a global point spread function. The blurred image is subjected to transmittance attenuation processing, and environmental noise is superimposed to generate a full-link degradation image.
[0011] In one embodiment, during the perspective projection transformation of the original clear image, if the pitch angle is less than a preset vertical threshold, the distortion margin is calculated based on the sine value of the pitch angle, and a target perspective quadrilateral vertex system is constructed. The original clear image is then subjected to perspective projection transformation based on the target perspective quadrilateral vertex system.
[0012] Secondly, this application also provides an anisotropic end-to-end degradation simulation system for high-altitude high-speed optical imaging, comprising: The perspective projection transformation module is used to acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field of view image. The phase superposition module is used to calculate the anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase. It physically superimposes the aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. The degradation image generation module is used to generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
[0013] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps: Acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field image; The anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase are calculated. The aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase are physically superimposed to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. Generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
[0014] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps: Acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field image; The anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase are calculated. The aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase are physically superimposed to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. Generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
[0015] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps: Acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field image; The anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase are calculated. The aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase are physically superimposed to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. Generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
[0016] This application employs the aforementioned anisotropic end-to-end degradation simulation method and system for high-altitude, high-speed optical imaging, which has the following beneficial effects: 1. An anisotropic spatial frequency calculation model based on flow field angle and flow field stretching ratio is introduced, which can accurately reproduce the directional stretching ambiguity caused by the high-altitude high-speed flow field environment on the light wavefront, breaking the limitation of related technologies that can only simulate isotropic ambiguity. Then, a low-frequency tilt compensation plane is designed. After superimposing this compensation plane, the system point spread function will show a random centroid shift consistent with the actual target observation height, which greatly improves the accuracy of the aerodynamic jitter model. Finally, the aerodynamic shock wave is physically superimposed with the shear layer phase, the system defocus phase, and the anisotropic turbulent target phase to achieve multi-physics coupling, thereby improving the accuracy of the simulation results.
[0017] 2. This application strictly divides the degradation process into coherent phase distortion and incoherent amplitude attenuation. First, all phase distortions are superimposed in the complex domain to generate a global point spread function, and then transmittance attenuation is performed in the spatial domain. This realizes the physical closed loop of multi-physics field coupled simulation and ensures that the generated end-to-end degradation image conforms to the real optical physics laws. Attached Figure Description
[0018] Figure 1 This is a flowchart of an anisotropic end-to-end degradation simulation method for high-altitude high-speed optical imaging in one embodiment; Figure 2This is a comparative heatmap of the isotropic and anisotropic point diffusion functions in one embodiment; wherein, Figure 2 In the graph (a), the point spread function is the pure turbulent degradation image. Figure 2 In the image, (b) represents the point spread function of the superimposed shock wave structure image; Figure 3 This is a phase comparison experimental data image of a pure turbulence degradation image and a composite degradation image of a superimposed shock wave structure in one embodiment; wherein, Figure 3 In the image, (a) represents the phase of the pure turbulence degradation image. Figure 3 (b) in the image represents the phase of the superimposed shock wave structure image; Figure 4 This is a visual comparison experimental data image of a pure turbulence degradation image and a composite degradation image of a superimposed shock wave structure, as shown in another embodiment; wherein, Figure 4 Image (a) in the image is a purely turbulent degradation image. Figure 4 Image (b) in the image is a superimposed shock wave structure; Figure 5 This is a comparative experimental data graph showing the end-to-end degradation effect of the proposed solution compared to Comparative Example 1 and Comparative Example 2 in one embodiment; wherein, Figure 5 Figure (a) in the figure shows the simulation results of the proposed solution. Figure 5 Figure (b) in the diagram shows the simulation results of the scheme in Comparative Example 1. Figure 5 Figure (c) in the figure shows the simulation results of the comparative example 2 scheme. Detailed Implementation
[0019] 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.
[0020] This application provides a definition of the relevant terms: Aerodynamic shock wave: refers to the surface of abrupt change in density and refractive index formed when the air at the front of an aircraft is strongly compressed during high-speed flight. In the physical optics model of this application, it is specifically characterized as a strong phase retardation layer that causes the light beam to have directional stretching characteristics.
[0021] Shear layer: This refers to the turbulent mixing layer generated by the interaction between high-speed airflow and the surface of the aircraft's optical window. Due to the high-frequency random fluctuations in its internal gas density and refractive index, random high-frequency phase distortion components are generated on the wavefront of the beam propagation when passing through this layer.
[0022] System defocus phase: This refers to the quadratic wavefront phase error caused by factors such as thermal difference, focusing deviation, or inherent physical structure of the lens in the optical imaging system. In mathematical models, it is usually represented as a quadratic surface phase that is proportional to the square of the pupil coordinates.
[0023] Random tilt coefficient: This represents a random variable used to simulate the low-frequency beam jitter phenomenon caused by aerodynamic flow or platform vibration. In the wavefront phase decomposition model, it corresponds to the tilt (Tip / Tilt) term coefficient in the Zernike polynomial decomposition, which can be used to construct a low-frequency tilt compensation plane.
[0024] Atmospheric coherence length: Represents the sub-aperture diameter with a root mean square value of 1 radian for the wavefront phase undulation of light waves propagating through the atmosphere at a given wavelength. It characterizes the maximum transverse scale at which the atmosphere can maintain the coherence of light waves.
[0025] Flow field stretching ratio: indicates the relative stretching of the flow field in a certain direction.
[0026] Flow field angle: represents the angle between the fluid velocity vector at a certain point and the reference coordinate axis.
[0027] like Figure 1 As shown, this application provides an anisotropic end-to-end degradation simulation method for high-altitude high-speed optical imaging, including: S100: Acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field image.
[0028] In one embodiment, during the perspective projection transformation of the original clear image, if the pitch angle is less than a preset vertical threshold, the distortion margin is calculated based on the sine value of the pitch angle, and a target perspective quadrilateral vertex system is constructed. The original clear image is then subjected to perspective projection transformation based on the target perspective quadrilateral vertex system.
[0029] This application first acquires a raw, clear image of the target area. This raw, clear image can be a remotely sensed image acquired under ideal conditions, a simulated image, or a pre-processed high signal-to-noise ratio image. Further, a perspective projection transformation is performed on the raw, clear image to generate a geometric field-of-view image. During this process, a perspective projection relationship can be established based on the flight altitude, attitude parameters, and intrinsic and extrinsic parameters of the imaging platform. When the pitch angle is less than a preset vertical threshold, it is preferable to calculate the distortion margin based on the sine of the pitch angle, and construct a target perspective quadrilateral vertex system based on this distortion margin, completing the image mapping through homography transformation.
[0030] It should be noted that the perspective projection model can be implemented using a pinhole imaging model or other equivalent projection models.
[0031] In one embodiment, this application loads a raw, sharp image with a resolution of 512x512. The observation altitude is set to 50,000 meters, and the pitch angle is 0 degrees. The trapezoidal distortion caused by strabismus is calculated based on the sine of the pitch angle, a perspective transformation matrix is constructed, and applied to the raw, sharp image to obtain a geometrically deformed field-of-view image.
[0032] S200, calculate the anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase. Physically superimpose the aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane.
[0033] After completing the geometric field-of-view image modeling, the phase distortion introduced during atmospheric propagation is constructed. Specifically, the phase distortion introduced in this application includes anisotropic turbulence phase, aerodynamic shock wave and shear layer phase, and system defocus phase. Specifically: the anisotropic turbulence phase is preferably obtained by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. The anisotropic turbulence initial phase characterizes the strong anisotropy exhibited by the stretching of turbulent vortices, while the low-frequency tilt compensation plane is used to characterize the low-frequency aerodynamic jitter phenomenon unique to high-altitude platforms. The aerodynamic shock wave and shear layer phase is used to characterize the high-frequency disturbances of the shear layer. The system defocus phase characterizes the defocus aberrations caused by different defocus amounts when the positions of the microscope objectives in the system's object light path and reference light path are not perfectly matched.
[0034] In one embodiment, the calculation process for the initial phase of anisotropic turbulence includes: obtaining the atmospheric coherence length, flow field stretching ratio, and flow field angle; constructing a two-dimensional basic spatial frequency grid, rotating the two-dimensional basic spatial frequency grid according to the flow field angle to generate a rotated coordinate system; calculating the square of the anisotropic spatial frequency based on the rotated coordinate system and the flow field stretching ratio; calculating the anisotropic power spectral density based on the von Kármán spectral model according to the atmospheric coherence length and the square of the anisotropic spatial frequency; adding random noise to the anisotropic power spectral density and performing an inverse Fourier transform to generate the initial phase of anisotropic turbulence.
[0035] First, a two-dimensional basic spatial frequency grid is generated. The grid is then rotated according to the flow field angle to obtain a rotated coordinate system. The longitudinal components of the rotated coordinate system are multiplied by the flow field stretching ratio to calculate the square of the anisotropic spatial frequency with directional stretching characteristics. Using the square of the anisotropic spatial frequency and the von Kármán spectral model formula, the anisotropic power spectral density is calculated. Finally, a complex random noise matrix is generated. The complex random noise matrix is multiplied by the square root of the anisotropic power spectral density, and an inverse Fourier transform is performed. The real part is taken as the initial phase of the atmospheric turbulence.
[0036] In one embodiment, the atmospheric coherence length is set to 20 cm, the flow field angle to 0 degrees, and the flow field stretching ratio to 2.0. A grid is constructed in the two-dimensional spatial frequency domain. After rotating the coordinate axes by 0 degrees, the frequency components on the vertical axis are multiplied by a coefficient of 2.0. The adjusted frequency components are substituted into the von Kármán spectral model formula to generate anisotropic power spectral density. An inverse fast Fourier transform is performed using Gaussian white noise to extract the real part and obtain the initial phase of the anisotropic turbulence.
[0037] The formula for the von Kármán spectral model is as follows:
[0038] in, The phase wavefront power spectral density; The length of atmospheric coherence; For spatial high-frequency cutoff parameters related to the internal scale of turbulence; Spatial low-frequency cutoff parameters related to the external scale of turbulence; These are spatial frequency components.
[0039] In one embodiment, the specific computation mechanism of the anisotropic spatial frequency grid is as follows: The generated size is Fundamental frequency matrix and The flow field angle is set to... ,calculate The cosine and sine values.
[0040] Then, calculate the transverse frequency components after rotation. for:
[0041] Calculate the longitudinal frequency components after rotation for:
[0042] Finally, the flow field stretching ratio parameter is introduced. Calculate the squared values of the anisotropic spatial frequencies:
[0043] when When this occurs, it indicates that the flow field is stretched in the longitudinal direction, which manifests as a lengthening of the turbulent structure in the corresponding direction in the spatial domain.
[0044] This application introduces an anisotropic spatial frequency calculation model based on the flow field angle and the flow field stretching ratio, which can accurately reproduce the directional stretching ambiguity caused by the high-altitude high-speed flow field environment on the light wavefront, breaking the limitation of related technologies that can only simulate isotropic ambiguity.
[0045] In one embodiment, the process of constructing the low-frequency tilt compensation plane includes: generating random tilt coefficients; constructing a first two-dimensional spatial coordinate grid; adjusting the phase value of each coordinate point in the first two-dimensional spatial coordinate grid based on the random tilt coefficients; and generating the low-frequency tilt compensation plane.
[0046] The random tilt coefficients generated in this application include horizontal random Gaussian variables and vertical random Gaussian variables that follow a normal distribution. By constructing a two-dimensional spatial coordinate grid, the horizontal random Gaussian variables are multiplied by the horizontal coordinate of the two-dimensional spatial coordinate grid, and the vertical random Gaussian variables are multiplied by the vertical coordinate of the two-dimensional spatial coordinate grid. The multiplication results are then summed to obtain a low-frequency tilt compensation plane for simulating aerodynamic vibration.
[0047] In one embodiment, when generating the phase screen using the Fourier domain filtering method employed in related technologies, the center point of the spatial frequency grid is typically forced to be [value missing]. This results in the generated image failing to represent the overall macroscopic jitter characteristics of the image. This embodiment establishes a system from... arrive A two-dimensional linear coordinate system mesh, where the x-coordinate of the two-dimensional linear coordinate system mesh is... The vertical axis is Using a random number generator, a variance of is generated. lateral random tilt coefficient and longitudinal random tilt coefficient .
[0048] This application calculates low-frequency tilt compensation. The formula is:
[0049] The calculated low-frequency tilt compensation It is directly added to the high-frequency phase plane obtained by inverse fast Fourier transform. Here, the high-frequency phase plane is the initial phase of the anisotropic turbulence.
[0050] This application addresses the persistent problem of low-frequency missing parameters in traditional Fourier transform phase screens by designing a low-frequency tilt compensation plane. Mathematical experimental data demonstrates that after superimposing this compensation plane, the system's point spread function exhibits a random centroid shift consistent with actual target range observations, significantly improving the accuracy of the aerodynamic jitter model.
[0051] In one embodiment, the calculation process of the aerodynamic shock wave and shear layer phase includes: obtaining the flow field angle; constructing a second two-dimensional spatial coordinate network and rotating the second two-dimensional spatial coordinate network according to the flow field angle; generating the main shock wave phase distribution based on the bow shock wave mathematical model and the rotated second two-dimensional spatial coordinate network; adding shear layer noise to the main shock wave phase distribution and multiplying it by the shock wave intensity parameter to generate the aerodynamic shock wave and shear layer phase.
[0052] After rotating the second spatial coordinate system according to the flow field angle, the main shock wave phase distribution is generated based on the rotated second two-dimensional spatial coordinate network using the bow shock wave mathematical model, and a random white noise matrix is generated. The random white noise matrix is convolved using a smooth convolution kernel to obtain the shear layer noise. The main shock wave phase distribution and the shear layer noise are added proportionally and multiplied by the shock wave intensity parameter to obtain the aerodynamic shock wave and shear layer phase.
[0053] This application sets the shock wave Mach intensity parameter to 1.5. An exponential bow-shaped shock wave phase is constructed using the bow-shaped equation, while shear layer noise is obtained by two-dimensional convolution of spatial random noise using a smoothing matrix. The bow-shaped shock wave phase, shear layer noise, and anisotropic turbulence phase are directly matrix-superimposed to form a composite phase plane with a size of 512x512. This composite phase plane is used to uniformly describe the wavefront distortion characteristics of light waves during propagation and imaging.
[0054] In one embodiment, the mathematical construction process of the phase between the aerodynamic shock wave and the shear layer is given, as follows: Set the flow field rotation angle to This generates a rotated coordinate grid. The x-coordinate of the rotated coordinate grid is... Mach intensity parameters Based on the local atmospheric density and aerodynamic thermodynamic properties, the shock wave intensity amplitude coefficient is calculated. Simultaneously set the curvature coefficient of the bow-shaped shock front. Shock wave detachment distance Shock layer thickness parameters .
[0055] Using a bow-shaped shock wave mathematical model, the main shock wave phase distribution is generated based on the above parameters, resulting in a white noise matrix with the same size as the image. A mean convolution kernel of a certain size is used to smooth the white noise matrix, and the smoothed result is extracted as the high-frequency perturbation of the shear layer. The bow-shaped shock wave and the high-frequency perturbation of the shear layer are added together according to weights to form the final composite phase plane.
[0056] In this application, the bow shock wave is calculated. The mathematical equation is:
[0057] S300 generates a full-link degradation image based on the geometric field of view image and the composite phase plane.
[0058] In one embodiment, generating a full-link degradation image based on a geometric field-of-view image and a composite phase plane includes: generating a global point spread function based on the composite phase plane; performing a convolution operation on the geometric field-of-view image using the global point spread function to generate a blurred image; performing transmittance attenuation processing on the blurred image and superimposing environmental noise to generate a full-link degradation image.
[0059] In this application, the steps of transmittance attenuation processing include: calculating the exponent of the product of the constant base and the negative aerosol optical thickness as atmospheric transmittance; extracting the high quantile pixel values of the blurred image as ambient light estimation values; multiplying the blurred image by the atmospheric transmittance, and adding the product of the ambient light estimation value and a factor minus the difference in atmospheric transmittance to obtain an image containing the fogging effect.
[0060] In one embodiment, the transmittance attenuation processing is performed after the blurred image has undergone spatial frequency domain convolution, using an exponential attenuation formula to simulate the optical absorption of aerosols and ambient light scattering.
[0061] This application strictly divides the degradation process into coherent phase distortion and incoherent amplitude attenuation. First, all phase distortions are superimposed in the complex domain to generate a global point spread function, and then transmittance attenuation is performed in the spatial domain. This realizes the physical closed loop of multi-physics field coupled simulation and ensures that the generated end-to-end degradation image conforms to the real optical physics laws.
[0062] Figure 2 This demonstrates a comparison of the morphology of the point spread function in the spatial frequency domain. For example... Figure 2 As shown in (a), the energy distribution of the point spread function in the purely turbulent degradation image exhibits a perfectly centrosymmetric circular diffuse spot, displaying isotropic fuzzy characteristics; while... Figure 2 As shown in (b), when the bow-shaped shock wave is coupled with the shear layer phase, the shape of the point spread function undergoes significant distortion. The high-energy region is violently stretched along the shock wave curvature, exhibiting a narrow, asymmetric elliptical or arc-shaped tail. The comparison of the point spread function shape in the spatial frequency domain intuitively proves that the anisotropic spatial frequency model of this application successfully reproduces the directional stretching and splitting effect of high-altitude, high-speed flow fields on light.
[0063] Figure 3 This demonstrates a comparison of two-dimensional wavefront phase distributions in the spatial domain. For example... Figure 3 As shown in (a), the phase distribution of pure turbulence exhibits random, irregular, cloud-like fluctuations, reflecting low-frequency and high-frequency phase random perturbations under Kolmogorov theory; Figure 3As shown in (b), after the introduction of the shock structure, a bright band with extremely high contrast and obvious bow-shaped geometric features (i.e., the shock front) appears in the originally chaotic random phase background. The comparison of the two-dimensional wavefront phase distribution in the spatial domain shows that the system has successfully carried out a rigorous physical superposition of random coherent distortion (turbulence) and structural coherent distortion (aerodynamic shock).
[0064] Figure 4 The comparison of the final degraded images output from the end-to-end simulation is shown. For example... Figure 4 As shown in (a), the image under pure turbulent conditions only exhibits a uniform decrease in overall sharpness, accompanied by slight jitter blurring without directionality; while... Figure 4 As shown in (b), under the combined degradation of shock waves and turbulence, the overall contrast of the image not only decreases due to aerosol attenuation, but more importantly, the edges of the target in the image exhibit severe asymmetric stretching, ghosting, and directional tearing of high-frequency details in a specific direction. The final degraded image output from the end-to-end simulation is in high agreement with the observations from real high-altitude high-speed wind tunnel optical experiments, proving that the simulation model in this application has extremely high physical fidelity.
[0065] like Figure 5 As shown, based on the above scheme, this application also provides relevant comparative examples: Comparative Example 1 employs imaging simulation methods from related technologies. In this comparative example, the flow field stretching scaling parameter is forcibly set to 1.0, and the construction step of the low-frequency tilt compensation plane is skipped. The remaining geometric projection and convolution steps are consistent with those of this application.
[0066] The images generated in Comparative Example 1 are compared and analyzed with those generated in this application. The comparison data shows that the images generated in Comparative Example 1 exhibit perfect circular diffuse spots, which is seriously inconsistent with the elliptical stretching and blurring phenomenon observed in high-altitude high-speed wind tunnel experiments. Furthermore, the centroid of the image sequence generated in Comparative Example 1 remains stable, losing the strong aerodynamic jitter characteristics of a real flight environment.
[0067] Comparative Example 2 illustrates the deficiency in related technologies due to the lack of physical mechanism decoupling. In this comparative example, the complex plane composite phase superposition step is skipped, and the atmospheric coherence length is directly used to estimate the scalar contrast loss of the entire system, thereby reducing the overall image brightness. This application, however, first performs coherent interferometry calculations in the complex domain to generate a global point spread function, and then independently performs aerosol incoherent energy attenuation.
[0068] Experimental comparisons show that the image generated by Comparative Example 2 fails to accurately reflect the dual physical degradation phenomenon of local edges being distorted by high-frequency phases while the overall background is submerged by aerosol grayscale. The theoretical goodness of fit of the proposed solution in image restoration tests is far superior to that of Comparative Example 2.
[0069] Secondly, this application also provides an anisotropic end-to-end degradation simulation system for high-altitude high-speed optical imaging, comprising: The perspective projection transformation module is used to acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field of view image. The phase superposition module is used to calculate the anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase. It physically superimposes the aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. The degradation image generation module is used to generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
[0070] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.
[0071] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.
[0072] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.
[0073] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.
[0074] Those skilled in the art will understand that all or part of the processes in 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 described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.
[0075] 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.
[0076] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for simulating anisotropic end-to-end degradation in high-altitude, high-speed optical imaging, characterized in that, include: Acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field image; The anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase are calculated. These are then physically superimposed to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. The aerodynamic shock wave represents the density and refractive index abrupt change surface formed by the intense compression of the air at the aircraft's front end during high-speed flight. The shear layer represents the turbulent mixing layer generated by the interaction between the high-speed airflow and the surface of the aircraft's optical window. The system defocus phase represents the quadratic wavefront phase error caused by the optical imaging system's own factors. Generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
2. The method according to claim 1, characterized in that, The calculation process for the initial phase of anisotropic turbulence includes: The atmospheric coherence length, flow field stretching ratio, and flow field angle are obtained. The atmospheric coherence length represents the sub-aperture diameter with a root mean square value of 1 radian for the wavefront phase undulation of a light wave as it propagates through the atmosphere at a given wavelength. The flow field stretching ratio represents the relative stretching of the flow field in a certain direction. The flow field angle represents the angle between the fluid velocity vector at a certain point and the reference coordinate axis. A two-dimensional basic spatial frequency grid is constructed, and the coordinates of the two-dimensional basic spatial frequency grid are rotated according to the flow field angle to generate a rotated coordinate system. The square of the anisotropic spatial frequency is calculated based on the rotated coordinate system and the flow field stretching ratio. Anisotropic power spectral density is calculated based on the von Kármán spectral model using the atmospheric coherence length and the square of the anisotropic spatial frequency. Random noise is added to the anisotropic power spectral density, and then an inverse Fourier transform is performed to generate the initial phase of the anisotropic turbulence.
3. The method according to claim 1, characterized in that, The process of constructing the low-frequency tilt compensation plane includes: Generate random tilt coefficients; the random tilt coefficients represent random variables used to simulate the overall jitter of low-frequency beams caused by aerodynamic flow fields or platform vibrations; A first two-dimensional spatial coordinate grid is constructed, and the phase value of each coordinate point in the first two-dimensional spatial coordinate grid is adjusted based on a random tilt coefficient to generate a low-frequency tilt compensation plane.
4. The method according to claim 1, characterized in that, The calculation process for the phase of the aerodynamic shock wave and the shear layer includes: Obtain the flow field angle; Construct a second two-dimensional spatial coordinate network, and rotate the second two-dimensional spatial coordinate network according to the flow field angle; Based on the mathematical model of bow shock wave, the phase distribution of the main shock wave is generated according to the rotated second two-dimensional spatial coordinate network; Shear layer noise is added to the phase distribution of the main shock wave and then multiplied by the shock wave intensity parameter to generate the phase of the aerodynamic shock wave and the shear layer.
5. The method according to any one of claims 1 to 4, characterized in that, Generate a full-link degradation image based on the geometric field-of-view image and the composite phase plane, including: Generate a global point spread function based on the composite phase plane; A blurred image is generated by convolving the geometric field of view image with a global point spread function. The blurred image is subjected to transmittance attenuation processing, and environmental noise is superimposed to generate a full-link degradation image.
6. The method according to claim 1, characterized in that, During the perspective projection transformation of the original clear image, if the pitch angle is less than the preset vertical threshold, the distortion margin is calculated based on the sine value of the pitch angle, and a target perspective quadrilateral vertex system is constructed. The original clear image is then transformed by perspective projection based on the target perspective quadrilateral vertex system.
7. A simulation system for anisotropic end-to-end degradation in high-altitude, high-speed optical imaging, characterized in that, include: The perspective projection transformation module is used to acquire the original clear image of the target area, perform perspective projection transformation on the original clear image, and generate a geometric field of view image. The phase superposition module is used to calculate the anisotropic turbulence phase, the aerodynamic shock wave and shear layer phase, and the system defocus phase. It physically superimposes the aerodynamic shock wave and shear layer phase, the system defocus phase, and the anisotropic turbulence target phase to generate a composite phase plane. The anisotropic turbulence phase is formed by superimposing the anisotropic turbulence initial phase and the low-frequency tilt compensation plane. The degradation image generation module is used to generate a full-link degradation image based on the geometric field of view image and the composite phase plane.
8. A computer 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 method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.