Atmospheric turbulence multilayer phase screen simulation mixing method

Abstract

The invention discloses an atmospheric turbulence multilayer phase screen simulation mixing method, which relates to the technical field of atmospheric optics and optical communication, and comprises the following steps: selecting a target area according to a simulation scene, and deploying a laser radar and a ground sensor network in combination with terrain and meteorological data; the laser radar detects scattered light in real time, the ground sensor measures near ground information, after data of the laser radar and the ground sensor is denoised, data of the laser radar and the ground sensor are integrated with data of satellite remote sensing and a ground meteorological station, and high-precision atmospheric parameters are iterated by constructing a three-dimensional space-time grid, applying algorithms such as dynamic time warping and a fusion physical model; according to atmospheric parameters, the correction model for the near-ground complex area is combined with an FFT algorithm and a window function to generate and correct each layer of phase screen; determining the weight according to a self-adaptive weight algorithm, and carrying out parallel calculation and superposition on the multi-layer phase screen; an optical path distortion degree and a phase covariance function are selected as indexes to build a system, and a deep learning mode is used for recognition and a statistical principle is used for verification. The precision of simulating the influence of the turbulence on light propagation is effectively improved.

Description

Technical Field

[0001] The invention relates to the field of atmospheric optics and optical communication technology, in particular to a hybrid method for simulating atmospheric turbulence multi-layer phase screens. Background Art

[0002] In many current application scenarios involving the propagation of light in the atmosphere, atmospheric turbulence is like a thorny problem that seriously restricts system performance. For example, in astronomical observations, when starlight passes through the atmosphere, the optical path distortion and phase fluctuation caused by atmospheric turbulence make the star image blurred and the resolution greatly reduced, making it difficult for astronomers to capture the fine features of celestial bodies; in the field of free-space optical communications, optical signals encounter turbulence during transmission, light intensity flickers, and phase jumps frequently, resulting in a surge in communication bit error rate and poor communication quality, which greatly limits its transmission distance and reliability.

[0003] Traditional atmospheric turbulence simulation methods have many drawbacks. On the one hand, in the process of obtaining atmospheric parameters, the field measurement means are single and often rely on limited meteorological station data. These data have low spatial resolution and are difficult to accurately reflect the true distribution of atmospheric parameters under local complex terrain and changeable meteorological conditions. For example, in urban high-rise areas, mountain canyons and other places, there are significant spatial differences in key parameters such as wind speed, temperature gradient, and atmospheric refractive index structure constant. Traditional measurements cannot meet the needs of high-precision simulation; although meteorological data has a wide coverage, it has poor timeliness and limited accuracy, and cannot provide real-time and accurate input for simulation. On the other hand, the simulation process is rough, and simple single-layer or a few-layer phase screen simulations are mostly used, completely ignoring the complex changes in turbulence characteristics in the vertical stratified structure of the atmosphere. The phase screen generation algorithm is simple and does not fully consider actual factors such as turbulence intermittentity and dynamic changes in wind speed, resulting in the simulation results being far from the impact of real atmospheric turbulence on light propagation, and unable to effectively guide the optimization design of practical application systems. Summary of the invention

[0004] In order to overcome the above problems existing in the prior art, the present invention proposes a hybrid method for simulating a multi-layer phase screen of atmospheric turbulence.

[0005] The technical solution adopted by the present invention to solve the technical problem is: a hybrid method for simulating a multi-layer phase screen of atmospheric turbulence, comprising: S1, select the target simulation area according to the simulation scenario, and deploy the lidar and ground sensor network based on terrain and meteorological data; S2, obtaining the atmospheric parameters of the simulation area, determining the number of simulation layers and the thickness of each layer according to the atmospheric parameters, and setting the wavelength of the simulation light; S3, based on the turbulence model, generates phase screens for each layer; S4, superimposing multiple layers of phase screens according to the weight of each layer's influence on light propagation; S5, verify and optimize the simulation results with the actual observation results; The atmospheric parameters of the simulated area in S2 are obtained by: using laser beam and scattered light field measurements, denoising the lidar data and ground sensor network data to obtain multi-source heterogeneous meteorological data, and fusing the multi-source heterogeneous meteorological data to obtain the atmospheric parameters of the simulated area.

[0006] The specific process of the above-mentioned atmospheric turbulence multi-layer phase screen simulation hybrid method for fusing the multi-source heterogeneous meteorological data is as follows: S21, conduct reliability assessment on satellite remote sensing data, ground weather station data, lidar and ground sensor network data, and exclude abnormal data; S22, based on the geographic coordinates of the simulation area, maps satellite remote sensing data, ground meteorological station data, and lidar and ground sensor network data into a unified grid. In the time dimension, the dynamic time warping algorithm is used in combination with the atmospheric parameter change rate estimation model to flexibly match the time series of different data sources; S23, constructing a fusion physical model, taking the atmospheric circulation vector field in the satellite remote sensing data after quality assessment and spatiotemporal adaptation, the air pressure and precipitation data of the ground meteorological station, the atmospheric refractive index structure parameters detected by the lidar, and the near-surface temperature, humidity, wind speed and direction data of the ground sensor network as input variables, numerically solving the physical model through the finite volume method and the finite difference method, and iteratively calculating the fused atmospheric parameters; S24, using the Monte Carlo simulation method, comprehensively quantify the uncertainty of each data source, model parameter uncertainty and numerical calculation error in the fusion process, characterize the reliability range of the atmospheric parameters after fusion in the form of uncertainty ellipse or confidence interval, substitute the fusion result into the atmospheric turbulence multi-layer phase screen simulation process for preview, compare the deviation of the preview result with the historical measured data or high-confidence theoretical model result, and adjust the physical model parameters, data source weight distribution or data screening strategy in reverse according to the direction and size of the deviation, and execute the fusion process again until the uncertainty of the fusion result is reduced to meet the preset accuracy requirements.

[0007] In the above-mentioned atmospheric turbulence multi-layer phase screen simulation hybrid method, S3 specifically includes: S31, selects turbulence models based on the atmospheric environment characteristics of the simulated area. The Kolmogorov model is used in relatively stable atmospheric areas, and a modified model that considers turbulence intermittency is used in near-ground complex terrain areas. Based on the selected model, combined with the atmospheric parameters of different altitude layers collected by S1, key parameter values ​​are determined, including accurate setting of atmospheric refractive index structure constants, inner scales, and outer scales for each layer; S32, using the fast Fourier transform algorithm, based on the power spectrum formula of the selected turbulence model, combined with the determined equal parameters, a random number matrix that conforms to the turbulence power spectrum distribution is generated in the frequency domain, and the frequency domain random number matrix is ​​converted to the spatial domain to obtain a preliminary phase screen; the Hanning window and Blackman window function are used to perform windowing processing to obtain relatively accurate initial phase screens of each layer; S33, collects wind speed and direction information of each layer, calculates horizontal and vertical wind speed components for each layer of phase screen according to wind speed conditions, translates the phase screen according to wind speed components according to simulation time step, simulates the dragging effect of wind on turbulent phase screen, and obtains the corrected phase screen of each layer.

[0008] In the above-mentioned atmospheric turbulence multi-layer phase screen simulation hybrid method, S4 specifically includes: S41, assigning preliminary weights according to the measured or estimated values ​​at different altitudes; S42, constructing a superposition model based on a mathematical weighted algorithm, taking the generated and corrected phase screens of each layer as input data, and performing weighted summation calculation point by point according to the weights determined in S31.

[0009] In the above-mentioned hybrid method for simulating a multi-layer phase screen of atmospheric turbulence, the larger the measured or estimated value of different altitude layers in S41, the stronger the interference to light propagation is usually, and the higher the initial weight assigned is; the thicker the atmospheric layer is given a bonus in the weight distribution; the relative position relationship between each layer and the light source and the receiving end is analyzed, and the weight ratio of the layer close to the light source or the receiving end is increased.

[0010] In the above-mentioned atmospheric turbulence multi-layer phase screen simulation hybrid method, S5 specifically includes: S51, selecting the degree of optical path distortion as a key verification indicator, and quantifying it into an optical path distortion indicator value by measuring the deviation between the simulated light propagation path and the actual observed light propagation path; S52, build a verification system with high-speed data acquisition and synchronization functions, including the data generated by each layer of phase screen, the superimposed light field data, and the corresponding actual observation field data; S53, matching and comparing the preprocessed simulated data features with the actual observed data features; S54, calculate the deviation range between the simulation data and the actual observation data on the selected verification index, and determine the reliability of the simulation result through the set confidence interval. If the deviation exceeds the confidence interval, it is determined that optimization is required.

[0011] In the above-mentioned atmospheric turbulence multi-layer phase screen simulation hybrid method, the S51 uses the phase covariance function as an auxiliary verification indicator to compare the change trend of the phase covariance function between the simulation and the actual observation.

[0012] In the above-mentioned atmospheric turbulence multi-layer phase screen simulation hybrid method, the S52 performs standardization and denoising on the collected simulation data and actual observation data respectively through the data preprocessing module in the verification system.

[0013] The beneficial effects of the present invention are as follows: (1) the present invention uses multi-source heterogeneous meteorological data fusion technology, adopts the collaboration of laser radar and ground sensor network, and combines satellite remote sensing and ground meteorological station data for deep fusion, which greatly improves the accuracy and spatiotemporal resolution of atmospheric parameter acquisition. Whether it is a mountainous area with complex terrain or a coastal area with changeable weather, it can accurately capture the subtle changes in atmospheric parameters, laying a solid foundation for subsequent high-precision simulation, so that the input at the simulation start end is more in line with the actual atmospheric state.

[0014] (2) Based on precise atmospheric parameters, the number of simulation layers and the thickness of each layer are determined scientifically and reasonably, and appropriate turbulence models are selected according to the characteristics of different regions. For example, the conventional Kolmogorov model is used in the stratosphere, and a modified model that takes into account the intermittent nature of turbulence is used in complex areas near the ground. Combined with advanced phase screen generation algorithms and wind speed correction mechanisms, the generated phase screens of each layer can accurately reflect the true evolution of atmospheric turbulence from near the ground to the high altitude, and the simulated light propagation affected by turbulence is more realistic.

[0015] (3) A unique multi-layer phase screen superposition weight determination strategy comprehensively considers the atmospheric parameters, thickness and position relationship of each layer with the light source receiving end, ensuring that the superimposed phase screen combination accurately restores the overall impact of atmospheric turbulence. It can more accurately predict signal fading and bit errors in optical communications, and effectively improve the accuracy of star image restoration in astronomical observations, thus helping to optimize systems and achieve performance breakthroughs in related fields in all aspects.

[0016] (4) A rigorous simulation result verification and optimization process, from multi-index verification, intelligent data acquisition and processing system construction to iterative optimization based on deviation feedback, ensures that the simulation results continue to approach the actual situation, continuously improves the simulation reliability, reduces the disconnection between simulation and actual application, and makes R&D design more forward-looking and effective. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0019] This embodiment discloses a hybrid method for simulating a multi-layer phase screen of atmospheric turbulence, such as Figure 1As shown, it mainly includes: 1. Clarify the simulation target and area: According to the application scenario of the simulation, such as the light propagation simulation of an astronomical observatory in a mountain area, or the simulation of low-altitude optical communication links in coastal cities, clarify the geographical scope and key areas of the simulation target area. Combined with the regional topographic map and historical meteorological data, a preliminary estimate of the possible complexity of atmospheric turbulence is made to provide a reference for subsequent equipment deployment and parameter setting.

[0020] 2. Equipment deployment and parameter setting: Select a suitable model and parameter configuration of LiDAR, and install it in a location with high terrain, wide field of view and avoid obvious obstructions according to the regional terrain, to ensure that the laser beam can cover the key simulation area to the greatest extent and realize all-round scanning and detection of the atmosphere. At the same time, according to the complexity and accuracy requirements of the near-ground, rationally plan the layout of the ground sensor network, and encrypt the sensor layout in areas with large terrain fluctuations and sensitive airflow changes, such as valleys and gaps between high-rise buildings, to ensure comprehensive collection of fine near-ground atmospheric parameters.

[0021] 3. Data collection, processing and integration: a. After the lidar is started, it collects scattered light information after the laser interacts with atmospheric molecules and aerosols in real time, and uses the built-in high-precision algorithm to quickly calculate parameters such as atmospheric refractive index structure constant and wind speed profile. The data collection frequency is set according to the dynamic needs of the simulation, such as several times per second in areas with variable weather. The ground sensor network operates synchronously, and thermometers, hygrometers, and anemometers accurately measure near-ground temperature and humidity, wind speed and direction, and transmit data to the data processing center at a high frequency (such as several times per minute) to complement the lidar data.

[0022] b. After receiving the two types of data, the data processing center immediately executes the weighted average algorithm and Kalman filter algorithm to remove noise interference. Among them, the weighted average algorithm dynamically allocates weights based on sensor accuracy and stability to enhance the influence of high-quality data; the Kalman filter algorithm predicts the credibility of current data based on historical data trends and accurately removes abnormal fluctuations.

[0023] c. Integrate the processed data with satellite remote sensing data and ground meteorological station data.

[0024] d. For satellite remote sensing data, professional image processing software is used to select high-credibility data subsets based on indicators such as cloud coverage, data quality identification, and consistency comparison with historical data of the same period, and information on areas with severe cloud cover and data anomalies is eliminated; for ground meteorological station data, abnormalities are checked by remotely retrieving equipment operation logs and calibration records, combined with recent changes in the surrounding environment (such as new buildings and vegetation changes) to ensure data reliability; for lidar and ground sensor network data, high-quality data is selected based on signal strength stability and the rationality of the measurement value fluctuation range.

[0025] e. A three-dimensional space-time grid model is constructed with the origin of the simulated regional geographic coordinates, and the spatial interpolation algorithm is used to map different source data to a unified grid to ensure data spatial consistency. In the time dimension, a dynamic time warping algorithm is used, combined with an atmospheric parameter change rate estimation model, and an elastic time matching window is set according to different altitude layers and regional characteristics to achieve accurate alignment of multi-source data time series. Finally, a fusion physical model is constructed based on the Navier-Stokes equation, the energy conservation equation, and the water vapor phase change equilibrium equation. The atmospheric circulation vector field in the satellite remote sensing data that has been quality evaluated and processed for space-time adaptation, the air pressure and precipitation data of the ground meteorological station, the atmospheric refractive index structure parameters detected by the lidar, and the near-ground temperature, humidity, wind speed and direction data of the ground sensor network are used as input variables. The physical model is numerically solved by the finite volume method and the finite difference method, and the fused atmospheric parameters are iteratively calculated.

[0026] The Monte Carlo simulation method is used to comprehensively quantify the uncertainties of various data sources, model parameter uncertainties and numerical calculation errors in the fusion process. The reliability range of the atmospheric parameters after fusion is characterized in the form of uncertainty ellipses or confidence intervals. The fusion results are substituted into the atmospheric turbulence multi-layer phase screen simulation process for preview. The deviations of the preview results and the historical measured data or high-confidence theoretical model results are compared. According to the direction and size of the deviation, the physical model parameters, data source weight distribution or data screening strategy are adjusted in reverse, and the fusion process is executed again until the uncertainty of the fusion result is reduced to meet the preset accuracy requirements.

[0027] 4. Simulate atmospheric environmental characteristics: The data processing center transmits the fused high-precision atmospheric parameters to the simulation module, and the simulation module simulates the regional atmospheric environmental characteristics based on parameter analysis. For relatively stable high-altitude areas such as the stratosphere, the Kolmogorov turbulence model is selected based on the atmospheric parameters corresponding to the altitude. Combined with the temperature, pressure and other information of the layer, key parameters such as the atmospheric refractive index structure constant, inner scale and outer scale are accurately set to ensure that the model reflects the true turbulent characteristics. In near-ground complex terrain areas, such as busy urban neighborhoods and the bottom of mountain canyons, a modified model that takes into account the intermittent nature of turbulence is used, and the model parameters are optimized and accurately assigned using the turbulence outbreak frequency and intensity change data measured in the field.

[0028] 5. Generate the initial phase screen of each layer: Use the fast Fourier transform (FFT) algorithm to generate the phase screen. According to the power spectrum formula of the selected turbulence model and the determined equal parameters, use the pseudo-random number generator in the frequency domain to generate a random number matrix that conforms to the turbulence power spectrum distribution. Convert the matrix to the spatial domain through FFT to obtain a preliminary phase screen. Considering the boundary effect of the FFT algorithm, use window functions such as the Hanning window and the Blackman window to perform windowing on the phase screen. According to the phase screen size and simulation accuracy requirements, reasonably select the window function parameters to suppress boundary "winding" and other problems, and obtain the initial phase screen of each layer.

[0029] 6. Correction of phase screens at each layer: Deploy high-precision wind speed measurement equipment, such as ultrasonic anemometers, at each altitude in the simulation area to collect wind speed and direction information in real time. The data transmission frequency matches the simulation time step to ensure timeliness. After each layer of phase screen is generated, the horizontal and vertical translations are accurately calculated using the vector translation algorithm based on the wind speed component and the simulation time step, and the phase screen is translated to simulate the dragging effect of the wind, and the corrected phase screens of each layer are generated to truly reflect the dynamic changes of atmospheric turbulence under the action of wind.

[0030] 7. Determine the weight of each layer of phase screen: The simulation system uses an adaptive weight allocation algorithm to preliminarily allocate weights based on the measured or estimated atmospheric refractive index structure constants at different altitudes. The larger the value, the stronger the interference to light propagation is usually, and the higher the initial weight allocated. The weight allocation is dynamically adjusted with real-time changes. Consider the thickness of each layer, and according to the theoretical model of light propagation in atmospheric layers of different thicknesses, appropriate weights are added to thick atmospheric layers, and the addition ratio is optimized based on the simulation verification results. Analyze the relative positions of each layer with the light source and the receiving end, and use the principles of geometric optics to increase the weight ratio of layers close to key parts, and comprehensively determine the precise weight of each layer.

[0031] 8. Superposition of each layer of phase screens: Construct a superposition model based on an efficient mathematical weighted algorithm, take the generated and corrected phase screens of each layer as input, and use parallel computing technology to perform weighted summation calculation point by point to improve the calculation efficiency. For example, suppose there are N layers of phase screens, and the coordinates on the i-th layer of phase screen are The phase value at , and its corresponding weight is , then the phase value of the total phase screen at this point after superposition is: , quickly obtain a multi-layer phase screen combination that simulates the overall impact of atmospheric turbulence.

[0032] 9. Verification and optimization of the model: a) Select the degree of optical path distortion as the key verification indicator. Deploy high-precision optical path measurement equipment, such as an optical path measurement instrument based on the principle of interference, at key nodes of the simulated light propagation path, such as near the light source and in front of the receiving end. Accurately measure the deviation between the simulated light propagation path and the actual observed light propagation path. Use a high-speed data acquisition card to collect data in real time and quantify it into an optical path distortion index value. At the same time, use the phase covariance function as an auxiliary verification indicator. With the help of professional optical analysis software, such as the optical toolbox based on MATLAB, compare the change trend of the phase covariance function between the simulation and the actual observation, and evaluate the accuracy of the simulation results from multiple angles.

[0033] b) Build a verification system with high-speed data acquisition and synchronization functions, and use a distributed data acquisition architecture to capture the output data of each link of the simulation in real time, including the generation of each layer of phase screen, the light field data after superposition, and the corresponding actual observation site data to ensure data integrity. The system has a built-in intelligent data preprocessing module, which uses different standardization algorithms and denoising filters to standardize and denoise the two types of data according to the data type and source, ensuring the uniformity of data quality and laying the foundation for accurate comparative analysis.

[0034] c) Use pattern recognition algorithms based on deep learning, such as convolutional neural network models, to accurately match and compare the features of pre-processed simulated data with the features of actual observed data, focus on the data differences of key nodes of light propagation such as light sources and receivers, and deeply mine the implicit features of the data; based on statistical principles, calculate the deviation range of simulated data and actual observed data on selected verification indicators, set the confidence interval of indicators such as the degree of optical path distortion to [-3σ, 3σ] (σ is the standard deviation), use hypothesis testing methods to determine the reliability of simulation results, and start the optimization process if the deviation exceeds the interval.

[0035] d) During optimization, first review the atmospheric parameters, accurately re-measure key parameters using lidar and other methods, calibrate the original data, compare the measurement results with previous data, and optimize the parameters using data fusion technology; deeply analyze the phase screen generation and superposition algorithms, adjust the window function parameters, increase the number of FFT points, optimize the weight allocation logic, and introduce adaptive mechanisms, such as dynamic adjustment of weights based on feedback control; re-substitute the optimized parameters and algorithms into the simulation, repeat the verification steps, and iterate the optimization until the simulation results meet the high-precision requirements, ensuring that the final simulation accurately reproduces the actual impact of atmospheric turbulence on light propagation.

[0036] The above embodiments are only exemplary embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art may make various modifications or equivalent substitutions to the present invention within the essence and protection scope of the present invention, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present invention.

Claims

1. A hybrid method for simulating atmospheric turbulence multi-layer phase screens, characterized in that: include: S1, select the target simulation area according to the simulation scenario, and deploy the lidar and ground sensor network based on terrain and meteorological data; S2, obtaining the atmospheric parameters of the simulation area, determining the number of simulation layers and the thickness of each layer according to the atmospheric parameters, and setting the wavelength of the simulation light; S3, based on the turbulence model, generates phase screens for each layer; S4, superimposing multiple layers of phase screens according to the weight of each layer's influence on light propagation; S5, verify and optimize the simulation results with the actual observation results; The atmospheric parameters of the simulated area in S2 are obtained by: using laser beam and scattered light field measurements, denoising the lidar data and ground sensor network data to obtain multi-source heterogeneous meteorological data, and fusing the multi-source heterogeneous meteorological data to obtain the atmospheric parameters of the simulated area.

2. The atmospheric turbulence multi-layer phase screen simulation hybrid method according to claim 1 is characterized in that: The specific process of data fusion of the multi-source heterogeneous meteorological data is as follows: S21, conduct reliability assessment on satellite remote sensing data, ground weather station data, lidar and ground sensor network data, and exclude abnormal data; S22, based on the geographic coordinates of the simulation area, maps satellite remote sensing data, ground meteorological station data, and lidar and ground sensor network data into a unified grid. In the time dimension, the dynamic time warping algorithm is used in combination with the atmospheric parameter change rate estimation model to flexibly match the time series of different data sources; S23, constructing a fusion physical model, taking the atmospheric circulation vector field in the satellite remote sensing data after quality assessment and spatiotemporal adaptation, the air pressure and precipitation data of the ground meteorological station, the atmospheric refractive index structure parameters detected by the lidar, and the near-surface temperature, humidity, wind speed and direction data of the ground sensor network as input variables, numerically solving the physical model through the finite volume method and the finite difference method, and iteratively calculating the fused atmospheric parameters; S24, using the Monte Carlo simulation method, comprehensively quantify the uncertainty of each data source, model parameter uncertainty and numerical calculation error in the fusion process, characterize the reliability range of the atmospheric parameters after fusion in the form of uncertainty ellipse or confidence interval, substitute the fusion result into the atmospheric turbulence multi-layer phase screen simulation process for preview, compare the deviation of the preview result with the historical measured data or high-confidence theoretical model result, and adjust the physical model parameters, data source weight distribution or data screening strategy in reverse according to the direction and size of the deviation, and execute the fusion process again until the uncertainty of the fusion result is reduced to meet the preset accuracy requirements.

3. The atmospheric turbulence multi-layer phase screen simulation hybrid method according to claim 1 is characterized in that: The S3 specifically includes: S31, selects turbulence models based on the atmospheric environment characteristics of the simulated area. The Kolmogorov model is used in relatively stable atmospheric areas, and a modified model that considers turbulence intermittency is used in near-ground complex terrain areas. Based on the selected model, combined with the atmospheric parameters of different altitude layers collected by S1, key parameter values ​​are determined, including accurate setting of atmospheric refractive index structure constants, inner scales, and outer scales for each layer; S32, using the fast Fourier transform algorithm, based on the power spectrum formula of the selected turbulence model, combined with the determined equal parameters, a random number matrix that conforms to the turbulence power spectrum distribution is generated in the frequency domain, and the frequency domain random number matrix is ​​converted to the spatial domain to obtain a preliminary phase screen; the Hanning window and Blackman window function are used to perform windowing processing to obtain relatively accurate initial phase screens of each layer; S33, collects wind speed and direction information of each layer, calculates horizontal and vertical wind speed components for each layer of phase screen according to wind speed conditions, translates the phase screen according to wind speed components according to simulation time step, simulates the dragging effect of wind on turbulent phase screen, and obtains the corrected phase screen of each layer.

4. The atmospheric turbulence multi-layer phase screen simulation hybrid method according to claim 1 is characterized in that: The S4 specifically includes: S41, assigning preliminary weights according to the measured or estimated values ​​at different altitudes; S42, constructing a superposition model based on a mathematical weighted algorithm, taking the generated and corrected phase screens of each layer as input data, and performing weighted summation calculation point by point according to the weights determined in S41.

5. The atmospheric turbulence multi-layer phase screen simulation hybrid method according to claim 4 is characterized in that: The larger the measured or estimated value of different altitude layers in S41, the stronger the interference to light propagation is usually, and the higher the initial weight assigned to it; the thicker the atmospheric layer is given a bonus in the weight distribution; the relative position relationship between each layer and the light source and the receiving end is analyzed, and the weight ratio of the layer close to the light source or the receiving end is increased.

6. The atmospheric turbulence multi-layer phase screen simulation hybrid method according to claim 1 is characterized in that: The S5 specifically includes: S51, selecting the degree of optical path distortion as a key verification indicator, and quantifying it into an optical path distortion indicator value by measuring the deviation between the simulated light propagation path and the actual observed light propagation path; S52, build a verification system with high-speed data acquisition and synchronization functions, including the data generated by each layer of phase screen, the superimposed light field data, and the corresponding actual observation field data; S53, matching and comparing the preprocessed simulated data features with the actual observed data features; S54, calculate the deviation range between the simulation data and the actual observation data on the selected verification index, and determine the reliability of the simulation result through the set confidence interval. If the deviation exceeds the confidence interval, it is determined that optimization is required.

7. The atmospheric turbulence multi-layer phase screen simulation hybrid method according to claim 6 is characterized in that: The S51 uses the phase covariance function as an auxiliary verification indicator to compare the change trends of the phase covariance functions of the simulation and the actual observation.

8. The atmospheric turbulence multi-layer phase screen simulation hybrid method according to claim 6 is characterized in that: The S52 performs standardization and denoising processing on the collected simulation data and actual observation data respectively through the data preprocessing module in the verification system.

Images