Coherent array structure light turbulence compensation method and system based on gradient back propagation

Abstract

The invention provides a coherent array structure light turbulence compensation method and system based on gradient back propagation, and the method comprises the steps: calculating the theoretical light intensity distribution of a defocus plane at a set distance after a focusing lens is focused through constructing a light field model of a coherent laser array under a turbulence-free condition; collecting observation light intensity distribution at a set distance after focusing of the focusing lens after transmission of the atmospheric turbulence; and constructing a loss function by using a mean square error between theoretical light intensity distribution and observation light intensity distribution, taking an array parameter as an optimization variable, calculating a gradient by using a gradient back propagation algorithm, and iteratively updating the array parameter. And the theoretical light intensity distribution and the loss value are calculated again after each update until convergence. And finally, subtracting the optimized array parameter from the initial array parameter to obtain a turbulence compensation parameter, and adjusting an array emission state according to the turbulence compensation parameter, thereby realizing effective compensation of wavefront distortion introduced by atmospheric turbulence. According to the invention, distortion is inverted through out-of-focus light intensity, the structure is simple, and the compensation efficiency is high.

Description

Technical Field

[0001] This invention belongs to the field of atmospheric light transmission and adaptive optics technology, specifically relating to a method for compensating for turbulence distortion in the structured light field of a coherent array emission. Background Technology

[0002] Structured light occupies an important position in modern optical technology due to its unique intensity, phase, or polarization spatial distribution. Among them, Hermite-Gaussian beams, Laguerre-Gaussian beams, and their superposition states, among other structured light types, have shown great application potential in cutting-edge fields such as orbital angular momentum multiplexing communication, high-precision quantum sensing, super-resolution microscopic imaging, and optical tweezers manipulation of microparticles, thanks to their respective optical properties.

[0003] Traditional structured light generation methods mainly rely on external optical modulators or intracavity mode selection in lasers. These methods generally suffer from inherent drawbacks such as low output power and slow dynamic mode switching speed, making it difficult to meet the practical application requirements of high power and fast response. Coherent laser arrays, as a novel structured light emission technology, can achieve high-power structured light generation by coordinating the amplitude and phase of multiple sub-beams. When combined with a high-frequency phase modulator, nanosecond-level mode switching can also be achieved, effectively overcoming the technical bottlenecks of traditional methods.

[0004] However, structured light inevitably suffers from atmospheric turbulence during long-distance transmission through the atmosphere. Atmospheric turbulence causes random distortions in the wavefront of the optical field, leading to problems such as mode scattering, energy diffusion, intensity attenuation, and decreased mode purity, severely damaging its optical properties and significantly reducing the performance of the transmission system. Therefore, achieving efficient compensation for atmospheric turbulence distortion is a key prerequisite for the practical application of coherent array structured light technology.

[0005] Existing turbulence compensation techniques are mainly divided into two categories: one is based on additional wavefront sensing devices (such as Shaker-Hartmann wavefront sensors), which obtain turbulence information by detecting a reference beam and then perform compensation. However, this type of method requires additional configuration of beacon light sources and detection systems, resulting in increased equipment size and complexity. Furthermore, the difference in transmission paths between the reference beam and the signal beam introduces compensation errors. The other type is based on inverting the turbulence phase from the light intensity at the receiving surface and reconstructing the wavefront distortion using Fourier transform relationships. However, this scheme relies on low-frequency, low-power devices such as spatial light modulators as phase modulation compensation carriers and cannot be applied to high-power beam transmission scenarios.

[0006] To address the aforementioned technical shortcomings, there is an urgent need for a turbulence distortion compensation method that is structurally simple, has high compensation accuracy, and is highly adaptable, in order to solve the problem of turbulence interference in coherent array structured light transmission. Summary of the Invention

[0007] To overcome the problems of system complexity, low compensation reliability, and narrow applicability in existing turbulence compensation technologies, this invention provides a coherent array structured light turbulence compensation method and system based on gradient backpropagation, which achieves efficient turbulence distortion compensation without additional reference beams and wavefront sensing devices, and improves the transmission quality of structured light under different turbulence intensities.

[0008] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows:

[0009] A coherent array structured light turbulence compensation method based on gradient backpropagation includes the following steps: Determine the target structured light field and the initial array parameters of the coherent laser array; Based on Kirchhoff diffraction theory and paraxial approximation, a physical forward model of the coherent laser array from the near-field light field to the far-field light field of the emitting surface under turbulence-free conditions is constructed. Based on the physical forward model, the theoretical light intensity distribution of the defocus plane at a set distance behind the focal point of the focusing lens is calculated and reconstructed. Under atmospheric turbulent transmission conditions, a defocused detection plane is set at a predetermined distance behind the focal point of the focusing lens of the coherent laser array to collect the observed light intensity distribution of the structured light after atmospheric turbulent transmission on the defocused detection plane. The loss function is constructed based on the difference between the theoretical light intensity distribution and the observed light intensity distribution. The partial derivatives of the current loss function with respect to each array parameter are calculated using the gradient backpropagation algorithm. The array parameters are iteratively updated with a preset learning rate. The iteration continues until the loss function converges or the preset maximum number of iterations is reached, and the final updated array parameters are output. The turbulence compensation parameters are obtained by subtracting the final updated array parameters from the corresponding initial array parameters. The array parameters of the coherent laser array under atmospheric turbulent transmission conditions are adjusted according to the turbulence compensation parameters to compensate for atmospheric turbulence distortion.

[0010] On the other hand, a coherent array structured light turbulence compensation system based on gradient backpropagation is provided to implement the aforementioned coherent array structured light turbulence compensation method based on gradient backpropagation, including: A coherent laser array is composed of multiple laser emitting units whose phase and tilt parameters can be independently controlled. A focusing lens is used to converge the beam emitted by a coherent laser array and form a transmission beam; The defocus detection module includes a CCD camera positioned at a set distance behind the focal lens, used to acquire the observed light intensity pattern of the transmitted light beam on the defocus plane after it has been transmitted through atmospheric turbulence. The control module stores the initial array parameters of the coherent laser array. Based on Kirchhoff diffraction theory and paraxial approximation, it constructs a physical forward model of the coherent laser array under turbulence-free conditions, from the near-field to the far-field light field of the emitting surface. Based on this physical forward model, it calculates and reconstructs the theoretical light intensity distribution at a set distance behind the focal point of the focusing lens. It acquires the observed light intensity pattern collected by the defocus detection module, constructs a loss function based on the difference between the theoretical and observed light intensity distributions, calculates the partial derivatives of the current loss function with respect to each array parameter using the gradient backpropagation algorithm, iteratively updates the array parameters at a preset learning rate, and iterates until the loss function converges or reaches a preset maximum number of iterations, outputting the final updated array parameters. The difference between the final updated array parameters and the corresponding initial array parameters is used to obtain turbulence compensation parameters. Based on these turbulence compensation parameters, a control signal is generated to adjust the array parameters of the coherent laser array under actual atmospheric turbulence transmission conditions, thereby compensating for atmospheric turbulence distortion.

[0011] The present invention has the following technical effects: This invention provides a coherent array structured light turbulence compensation method based on gradient backpropagation. It constructs an optical field model of the coherent laser array under turbulence-free conditions and calculates the theoretical light intensity distribution at a predetermined distance behind the focal point of the focusing lens. The observed light intensity distribution at the same distance behind the focal point is acquired after actual atmospheric turbulence propagation. A loss function is constructed using the mean square error between the theoretical and observed light intensity distributions. The piston phase and tilt parameters of each beam element in the array are used as optimization variables. The gradient backpropagation algorithm is used to calculate the gradient and iteratively update these parameters. After each update, the theoretical light intensity distribution and loss value are recalculated until convergence. Finally, the difference between the optimized array parameters and the initial array parameters is used to obtain the turbulence compensation parameters. Based on these parameters, the array emission state is adjusted, thereby achieving effective compensation for wavefront distortion introduced by atmospheric turbulence. This invention, by inverting distortion through defocused light intensity, has the advantages of simple structure and high compensation efficiency. This invention eliminates the need for additional reference beams, beacon light sources, and wavefront sensing devices. Turbulence compensation can be achieved solely through a single coherent laser array combined with backend detection and algorithm processing, significantly reducing system size, weight, and complexity, and enhancing its engineering practicality.

[0012] The present invention has high compensation accuracy and reliability. By using the light intensity distribution off-focus plane as the compensation basis, it utilizes the additional spatial frequency information contained therein to alleviate the ill-conditioned problem of phase recovery and ensure the uniqueness and physical validity of the compensation solution. Experimental verification shows that under weak and medium turbulence conditions, the purity of the structured light mode after compensation exceeds 95%, and the mode purity can still exceed 90% under strong turbulence conditions.

[0013] The present invention has a fast convergence speed: by reasonably setting the learning rate of the gradient backpropagation algorithm, it can achieve a model purity of more than 95% in only 10 iterations under weak turbulence conditions, 20 iterations in moderate turbulence, and 95% in strong turbulence after 65 iterations. The average iteration time is 220ms, and it has real-time compensation capability. After FPGA hardware acceleration, it can be further improved to nanosecond level response.

[0014] This invention is highly versatile: it is applicable to various structured lights such as Hermite-Gaussian beams, Laguerre-Gaussian beams and their arbitrary superposition states, and has good adaptability to coherent arrays of different sizes. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a coherent array structured optical turbulence compensation method based on gradient backpropagation in one embodiment; Figure 2 This is a schematic diagram of a coherent array structured optical turbulence compensation system based on gradient backpropagation in one embodiment. Figure 3 This is a comparison of structured light before and after compensation under different turbulence intensities in one embodiment. Detailed Implementation

[0017] 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 a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] Reference Figure 1 A coherent array structured light turbulence compensation method based on gradient backpropagation, provided in one embodiment, includes: Determine the target structured light field and the initial array parameters of the coherent laser array; Based on Kirchhoff diffraction theory and paraxial approximation, a physical forward model of the coherent laser array from the near-field light field to the far-field light field of the emitting surface under turbulence-free conditions is constructed. Based on the physical forward model, the theoretical light intensity distribution of the defocus plane at a set distance behind the focal point of the focusing lens is calculated and reconstructed. Under atmospheric turbulent transmission conditions, a defocused detection plane is set at a predetermined distance behind the focal point of the focusing lens of the coherent laser array to collect the observed light intensity distribution of the structured light after atmospheric turbulent transmission on the defocused detection plane. The loss function is constructed based on the difference between the theoretical light intensity distribution and the observed light intensity distribution. The partial derivatives of the current loss function with respect to each array parameter are calculated using the gradient backpropagation algorithm. The array parameters are iteratively updated with a preset learning rate. The iteration continues until the loss function converges or the preset maximum number of iterations is reached, and the final updated array parameters are output. The turbulence compensation parameters are obtained by subtracting the final updated array parameters from the corresponding initial array parameters. The array parameters of the coherent laser array under atmospheric turbulent transmission conditions are adjusted according to the turbulence compensation parameters to compensate for atmospheric turbulence distortion.

[0019] In the above embodiments, a single coherent array accomplishes the dual functions of structured light generation and turbulence compensation. The array parameters of the coherent laser array include the piston phase and tilt parameters of each laser emitting unit. The initial array parameters of the coherent laser array are set according to the target structured light field.

[0020] In one embodiment, the constructed coherent laser array is an N×N square coherent laser array, where each sub-beam is a linearly polarized fundamental mode Gaussian beam. j The light field of the sub-beam at the emitting surface, i.e., z=0 for:

[0021] in j Number the sub-beams. j =1,2,......,n, where n is the total number of sub-beams in the coherent laser array; x , y () represents any coordinate position on the launch surface, x j , y j ) is the first j The coordinates of the sub-beam on the emitting surface. A j For the first j The amplitude of the individual beams, ω 0 represents the beam waist width. d For sub-beam aperture, circ Functions within the circular domain; The near-field optical field of the coherent laser array at the emitting surface, i.e., z=0 for:

[0022] in j For the first j Phase of individual beams k For wave vector, k =2π / λ, where λ is the laser wavelength. θ jx and θ jy For the first j The tilt angle of the sub-beam in the x and y directions. i The imaginary unit is exp, which represents the exponentiation of e. The physical forward model calculates the optical field from the near field of the emitting surface to the far field behind the focusing lens using Fourier transform. Based on the physical forward model, it calculates and reconstructs the optical field at a set distance behind the focal point of the focusing lens. L The theoretical light intensity distribution at the defocus plane is as follows:

[0023] in( u , v )for z = L The far-field coordinates of the point, where λ is the laser wavelength. f For the focal length of the focusing lens, L For the propagation distance, F(·) represents the Fourier transform.

[0024] Under actual atmospheric turbulent transport conditions, a distance is set behind the focal point of the focusing lens of the coherent laser array. L A defocused detection plane is set up, and a CCD camera is used to collect the observed light intensity distribution on the defocused detection plane after the structured light has been transmitted through atmospheric turbulence. In one embodiment, L=20mm, and the CCD camera detection resolution is set to 224×224 pixels to obtain complete spatial frequency characteristics including turbulence distortion information.

[0025] The loss function is constructed using the mean square error (MSE) between the theoretical and observed light intensity distributions. The Adam optimizer is used, and the piston phase of the loss function for each laser emitting unit is calculated via gradient backpropagation. j and tilt parameters θ jx , θ jy The partial derivatives are used to iteratively update the array parameters with a preset learning rate. For example, in one embodiment, the piston phase learning rate is set to 0.1, and the x and y tilt parameter learning rate is set to 10. -6 Each time the array parameters are updated, the theoretical light intensity distribution of the defocus plane is updated through the physical forward model, and the loss function value is recalculated.

[0026] The convergence condition of the loss function is: the loss function is less than a set threshold, or the change of the loss function in multiple consecutive iterations is less than the set threshold.

[0027] In one embodiment, the convergence condition is set as follows: when the loss function value is less than a preset threshold or the number of iterations reaches 100, the array parameter update is stopped. At this time, the array parameters can reconstruct the distortion effect of turbulence on the beam. The turbulence compensation parameters are obtained by subtracting the current array parameters from the initial array parameters. The coherent array emits according to these parameters based on the original parameters, thereby achieving effective compensation for turbulence distortion.

[0028] In another embodiment, the coherent array structured light turbulence compensation system based on gradient backpropagation includes: A coherent laser array, composed of multiple laser emitting units whose phase and tilt parameters can be independently controlled, is used to generate a target structured light field under the control of a control module; A focusing lens is used to converge the beam emitted by a coherent laser array and form a transmission beam; The defocus detection module includes a CCD camera positioned at a focal distance L behind the focusing lens, used to acquire the observed light intensity pattern of the transmitted light beam on the defocus plane after it has been transmitted through atmospheric turbulence. The control module stores the initial array parameters of the coherent laser array. Based on Kirchhoff diffraction theory and paraxial approximation, it constructs a physical forward model of the coherent laser array under turbulence-free conditions, from the near-field to the far-field light field of the emitting surface. Based on this physical forward model, it calculates and reconstructs the theoretical light intensity distribution at the defocused plane at the focal distance z=L of the focusing lens. It acquires the observed light intensity pattern collected by the defocus detection module, constructs a loss function based on the difference between the theoretical and observed light intensity distributions, calculates the partial derivatives of the current loss function with respect to each array parameter using the gradient backpropagation algorithm, iteratively updates the array parameters at a preset learning rate, and iterates until the loss function converges or reaches the preset maximum number of iterations, outputting the final updated array parameters. The difference between the final updated array parameters and the corresponding initial array parameters is used to obtain turbulence compensation parameters. Based on these turbulence compensation parameters, a control signal is generated to adjust the array parameters of the coherent laser array under actual atmospheric turbulence transmission conditions, thereby compensating for atmospheric turbulence distortion.

[0029] Reference Figure 2This is a schematic diagram of a coherent array structured optical turbulence compensation system based on gradient backpropagation in one embodiment. It includes a seed laser 1, a preamplifier 2, an optical fiber beam splitter 3, a phase modulator 4, an optical fiber amplifier 5, a collimator array 6, a focusing lens 7, an atmospheric turbulence region 8, a beam splitter 9, a first camera 10, a second camera 11, and a control module 12. The first camera 10 is used to observe the far-field beam shape. The second camera 11 is used for defocus plane detection, acting as a defocus detection module. The second camera 11 is positioned 20mm behind the focal point of the focusing lens 7 at the defocus detection plane, with a detection resolution of 224×224 pixels, and is used to achieve feedback correction of turbulence distortion. The control module 12 uses an industrial computer equipped with an NVIDIA RTX 3090 graphics card to run the gradient backpropagation compensation algorithm and to implement real-time updates of array parameters.

[0030] A coherent laser array is composed of a seed laser 1, a preamplifier 2, an fiber beam splitter 3, a phase modulator 4, a fiber amplifier 5, and a collimator array 6. The seed laser 1 is a 1064nm linearly polarized seed laser. The seed laser emitted by the seed laser 1 is pre-amplified by the preamplifier 2 and then split into multiple unit beams by the fiber beam splitter 3. Each unit beam has a phase modulator 4, a fiber amplifier 5, and an adaptive collimator sequentially along its transmission path. Each unit beam undergoes phase modulation by the corresponding phase modulator 4, power amplification by the fiber amplifier 5, and collimation by the adaptive collimator. All adaptive collimators are arranged in the desired array configuration to form the collimator array 6. The beam emitted by the coherent laser array, after passing through the focusing lens 7, undergoes wavefront distortion as it passes through the atmospheric turbulence region 8. The wavefront-distorted light beam is split by beam splitter 9, with one part transmitted to the first camera 10 and the other part to the second camera 10. The second camera, positioned at a focal distance L behind the focusing lens 7, captures the intensity pattern of the transmitted beam on the defocus plane after passing through atmospheric turbulence. The second camera 11 captures the intensity distribution pattern and transmits it to the control module 12.

[0031] The atmospheric turbulence region 8 can be an atmospheric turbulence region in a real application environment, or it can be an atmospheric turbulence region 8 simulated based on the Kolmogorov model in the experiment. The atmospheric turbulence region 8 simulated based on the Kolmogorov model is generated by generating a turbulence phase screen using the power spectrum method, which can simulate three scenarios: weak turbulence (r0=3mm), moderate turbulence (r0=2mm), and strong turbulence (r0=1mm).

[0032] In any of the above embodiments, the atmospheric turbulence region 8 can be simulated using the Kolmogorov model, and a turbulence phase screen can be generated by the power spectrum method. The coherence length r0 = 3 mm in the weak turbulence scenario, r0 = 2 mm in the moderate turbulence scenario, and r0 = 1 mm in the strong turbulence scenario.

[0033] The control module can use an industrial computer equipped with an NVIDIA RTX 3090 graphics card to run the gradient backpropagation compensation algorithm and realize the real-time update of array parameters.

[0034] One embodiment is used to generate a Laguerre-Gaussian beam LG. 02 Taking a topological load of l=2 as an example, the specific implementation steps of the turbulence compensation method provided by this invention are as follows: System initialization: Seed laser 1 emits 1064nm linearly polarized laser light, which passes through preamplifier 2, fiber beam splitter 3, phase modulator 4, fiber amplifier 5, and collimator array 6. The control module controls a 5×5 square coherent laser array to generate the target structured light field, i.e., target LG, according to the target structured light field settings. 02 The beam, at this time the piston phase φ of the coherent laser array is initially determined according to LG. 02 Beam phase distribution settings: initial values ​​of tilt parameters θx and θy are set to 0.

[0035] Turbulent transport and light intensity detection: In atmospheric turbulence region 8, the Kolmogorov model can be used to simulate weak, moderate, and strong turbulence scenarios. A turbulent phase screen is generated using the power spectral density method. The coherence length r0 = 3 mm in the weak turbulence scenario, 2 mm in the moderate turbulence scenario, and 1 mm in the strong turbulence scenario. The LG emitted by the coherent laser array (e.g., a 5×5 square coherent laser array)... 02 After passing through the focusing lens 7, the light beam undergoes wavefront distortion as it passes through the atmospheric turbulence region 8. The distorted light beam forms a light intensity distribution, i.e., the observed light intensity distribution, at the defocused detection plane 20mm behind the focal point of the focusing lens 7. The second camera 11 acquires the observed light intensity distribution pattern and transmits it to the control module 12.

[0036] Based on Kirchhoff diffraction theory and paraxial approximation, control module 12 constructs a physical forward model of the coherent laser array from the near-field to the far-field light field of the emitting surface under turbulence-free conditions. Based on this physical forward model, it calculates and reconstructs the theoretical light intensity distribution at the defocused plane at the focal distance z=L of the focusing lens. The difference between the theoretical and observed light intensity distributions is used to construct the MSE loss function. The Adam optimizer is used to initialize the learning rate, with a piston phase learning rate of 0.1 and a tilt parameter learning rate of 10. -6The gradient of the loss function with respect to the piston phase φ, the x-direction tilt parameter θx, and the y-direction tilt parameter θy is calculated using the gradient backpropagation algorithm. The array parameters are updated based on the gradient calculation results, and this process is iterated until the loss function converges or the preset maximum number of iterations is reached. The final updated array parameters are then output. The difference between the final updated array parameters and the corresponding initial array parameters is used to obtain the turbulence compensation parameters. Control signals are generated based on these turbulence compensation parameters to adjust the array parameters of the coherent laser array under actual atmospheric turbulence propagation conditions, thereby compensating for atmospheric turbulence distortion.

[0037] Compensation effect verification: The purity of the structured light mode was calculated by capturing the observed light intensity distribution pattern formed by the defocused detection plane after compensation using the second camera. (Refer to...) Figure 3 Comparison of structured light before and after compensation under different turbulence intensities. Figure 2 The first column and first row represent the vortex beam with a standard topological charge of 2. The second column and second row show the far-field beam pattern generated by the laser array under turbulence-free conditions. The second to fourth columns show the far-field beam pattern of the laser array under weak, medium, and strong turbulence, respectively, as well as the beam pattern after compensation using the gradient backpropagation-based coherent array structured light turbulence compensation method proposed in this invention. Experimental results show that the mode purity after weak / medium / strong turbulence compensation is ≥90%, meeting the requirements of practical applications.

[0038] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A coherent array structured light turbulence compensation method based on gradient backpropagation, characterized in that, include: Determine the target structured light field and the initial array parameters of the coherent laser array; Based on Kirchhoff diffraction theory and paraxial approximation, a physical forward model of the coherent laser array from the near-field light field to the far-field light field of the emitting surface under turbulence-free conditions is constructed. Based on the physical forward model, the theoretical light intensity distribution of the defocus plane at a set distance behind the focal point of the focusing lens is calculated and reconstructed. Under atmospheric turbulent transmission conditions, a defocused detection plane is set at a predetermined distance behind the focal point of the focusing lens of the coherent laser array to collect the observed light intensity distribution of the structured light after atmospheric turbulent transmission on the defocused detection plane. The loss function is constructed based on the difference between the theoretical light intensity distribution and the observed light intensity distribution. The partial derivatives of the current loss function with respect to each array parameter are calculated using the gradient backpropagation algorithm. The array parameters are iteratively updated with a preset learning rate. The iteration continues until the loss function converges or the preset maximum number of iterations is reached, and the final updated array parameters are output. The turbulence compensation parameters are obtained by subtracting the final updated array parameters from the corresponding initial array parameters. The array parameters of the coherent laser array under atmospheric turbulent transmission conditions are adjusted according to the turbulence compensation parameters to compensate for atmospheric turbulence distortion.

2. The coherent array structured light turbulence compensation method based on gradient backpropagation according to claim 1, characterized in that, The array parameters of the coherent laser array include the piston phase and tilt parameters of each laser emitting unit, and the initial array parameters of the coherent laser array are set according to the target structured light field.

3. The coherent array structured light turbulence compensation method based on gradient backpropagation according to claim 2, characterized in that, The coherent laser array is an N×N square coherent laser array, and each sub-beam in the array is a linearly polarized fundamental Gaussian beam.

4. The coherent array structured light turbulence compensation method based on gradient backpropagation according to claim 3, characterized in that, The first in a coherent laser array j The light field of the sub-beam at the emitting surface, i.e., z=0 for: in j Number the sub-beams. j =1,2,......,n, where n is the total number of sub-beams in the coherent laser array; ( x , y () represents any coordinate position on the launch surface, x j , y j ) is the first j The coordinates of the sub-beam on the emitting surface. A j For the first j The amplitude of the individual beam ω 0 represents the beam waist width. d For sub-beam aperture, circ Functions within the circular domain; The near-field optical field of the coherent laser array at the emitting surface, i.e., z=0 for: in j For the first j Phase of individual beams k For wave vector, k =2π / λ, where λ is the laser wavelength. θ jx and θ jy For the first j The tilt angle of the sub-beam in the x and y directions. i The imaginary unit is exp, which represents the exponentiation of e. The physical forward model calculates the optical field from the near field of the emitting surface to the far field behind the focusing lens using Fourier transform. Based on the physical forward model, it calculates and reconstructs the optical field at a set distance behind the focal point of the focusing lens. L The theoretical light intensity distribution at the defocus plane is as follows: in( u , v )for z = L The far-field coordinates of the point, where λ is the laser wavelength. f For the focal length of the focusing lens, L For the propagation distance, F(·) represents the Fourier transform.

5. The coherent array structured light turbulence compensation method based on gradient backpropagation according to any one of claims 1 to 4, characterized in that, The convergence condition of the loss function is: the loss function is less than a set threshold, or the change of the loss function in multiple consecutive iterations is less than the set threshold.

6. A coherent array structured light turbulence compensation system based on gradient backpropagation, characterized in that, The method for implementing coherent array structured light turbulence compensation based on gradient backpropagation as described in any one of claims 1 to 4 includes: A coherent laser array, consisting of multiple laser emitting units whose phase and tilt parameters are independently controlled, is used to generate a target structured light field under the control of a control module. A focusing lens is used to converge the beam emitted by a coherent laser array and form a transmission beam; The defocus detection module includes a CCD camera positioned at a focal distance L behind the focusing lens, used to acquire the observed light intensity pattern of the transmitted light beam on the defocus plane after it has been transmitted through atmospheric turbulence. The control module stores the initial array parameters of the coherent laser array. Based on Kirchhoff diffraction theory and paraxial approximation, it constructs a physical forward model of the coherent laser array under turbulence-free conditions, from the near-field to the far-field light field of the emitting surface. Based on this physical forward model, it calculates and reconstructs the theoretical light intensity distribution at the defocused plane at the focal distance z=L of the focusing lens. It acquires the observed light intensity pattern collected by the defocus detection module, constructs a loss function based on the difference between the theoretical and observed light intensity distributions, calculates the partial derivatives of the current loss function with respect to each array parameter using the gradient backpropagation algorithm, iteratively updates the array parameters at a preset learning rate, and iterates until the loss function converges or reaches the preset maximum number of iterations, outputting the final updated array parameters. The difference between the final updated array parameters and the corresponding initial array parameters is used to obtain turbulence compensation parameters. Based on these turbulence compensation parameters, a control signal is generated to adjust the array parameters of the coherent laser array under actual atmospheric turbulence transmission conditions, thereby compensating for atmospheric turbulence distortion.

7. The coherent array structured light turbulence compensation system based on gradient backpropagation according to claim 6, characterized in that, The coherent laser array includes a seed laser, a preamplifier, a fiber beam splitter, a phase modulator, a fiber amplifier, and a collimator array; The seed laser emitted by the seed laser is pre-amplified by a preamplifier and then split into multiple unit beams by an optical fiber beam splitter. Each unit beam has a phase modulator, an optical fiber amplifier, and an adaptive collimator installed sequentially along its transmission path. Each unit beam is phase-modulated by its corresponding phase modulator, power-amplified by its optical fiber amplifier, and collimated by its adaptive collimator. All the adaptive collimators are arranged in the required array to form a collimator array. The beam emitted by the coherent laser array undergoes wavefront distortion after passing through the focusing lens and traversing the atmospheric turbulence region. The wavefront-distorted beam is then split by a beam splitter, with one part transmitted to the first camera and the other part to the second camera. The second camera, located at a focal distance L behind the focusing lens, captures the observed light intensity pattern of the transmitted beam after passing through the atmospheric turbulence on the defocus plane. The second camera captures the observed light intensity distribution pattern and transmits it to the control module.

8. The coherent array structured light turbulence compensation system based on gradient backpropagation according to claim 7, characterized in that, The seed laser is a 1064nm linearly polarized seed laser.

9. The coherent array structured optical turbulence compensation system based on gradient backpropagation according to claim 7, characterized in that, The coherent laser array is a 5×5 square coherent laser array.

10. The coherent array structured optical turbulence compensation system based on gradient backpropagation according to claim 7, 8, or 9, characterized in that, The atmospheric turbulent region was simulated using the Kolmogorov model. The turbulent phase screen was generated using the power spectrum method. The coherence length r0 = 3 mm in the weak turbulence scenario, r0 = 2 mm in the moderate turbulence scenario, and r0 = 1 mm in the strong turbulence scenario.

Images