Automatic aberration correction system and method for remote sensing light field imaging

Abstract

The invention relates to the technical field of optical imaging, and discloses an automatic aberration correction system and method for remote sensing light field imaging, and the method comprises the steps: constructing an aberration sensitive feature set through dynamic projection coding and multispectral synchronous collection, and improving the aberration perception capability of a complex scene in combination with polarized light information; establishing a three-dimensional aberration field model based on a light field gradient decomposition and multi-focal length coupling model; calculating an optimal focusing parameter and a digital phase compensation mask in combination with physical driving optimization and data driving compensation; and dynamic adaptive correction is realized based on real-time aberration monitoring and light field entropy minimization optimization. According to the method, aberration such as spherical aberration, coma aberration and curvature of field can be effectively inhibited, and the resolution and depth estimation precision of remote sensing light field imaging are improved.

Description

Technical Field

[0001] The present invention relates to the field of optical imaging technology, and in particular to an automated aberration correction system and method for remote sensing light field imaging. Background Art

[0002] Remote sensing light field imaging technology is widely used in fields such as Earth observation, environmental monitoring, agricultural remote sensing, and target identification. Traditional remote sensing imaging systems typically rely on large-aperture optical lenses and high-resolution sensors. However, due to atmospheric disturbances, lens aberrations, and sensor non-uniformity, imaging quality is often limited, making it difficult to obtain high-precision three-dimensional light field information.

[0003] As a novel light field acquisition technology, multifocal-length microlens arrays can improve the system's depth of field and spatial resolution through multi-focal-depth imaging. However, due to the focal length differences between microlens units and inherent aberrations in optical systems, such as spherical aberration, coma, and field curvature, directly acquired light field data often suffer from geometric distortion, focus shift, and intensity non-uniformity, affecting subsequent image restoration and depth estimation accuracy. Therefore, an efficient, automated aberration correction method is needed to improve the quality and precision of light field imaging and achieve more accurate remote sensing measurements. Summary of the Invention

[0004] In view of this, the present invention proposes an automated aberration correction system and method for remote sensing light field imaging, aiming to effectively solve the above problems and improve the quality and accuracy of remote sensing light field imaging.

[0005] The present invention proposes an automated aberration correction method for remote sensing light field imaging. The method is based on adaptive correction technology using dynamic projection and multi-focal length light field feedback, and includes the following steps:

[0006] Data acquisition: Through dynamic projection coding and multi-spectral synchronous acquisition, we build an aberration-sensitive feature set and combine it with polarization information to improve the aberration perception ability of complex scenes;

[0007] Aberration diagnosis: Based on the light field gradient decomposition and multi-focal length coupling model, a three-dimensional aberration field model is established;

[0008] Correction calculation: Combine physical drive optimization with data-driven compensation to calculate the optimal focusing parameters and digital phase compensation mask;

[0009] Closed-loop feedback: Dynamic adaptive correction is achieved based on real-time aberration monitoring and light field entropy minimization optimization.

[0010] Preferably, the data collection step includes:

[0011] The near-infrared spot projection device is controlled to generate a sparse spot with random temporal and spatial encoding. The spot position P(x, t) and shape S(x, y, t) change in a pseudo-random sequence in consecutive frames:

[0012]

[0013] Among them, P0 represents the initial spot position; A n represents the amplitude; ω n Indicates frequency; φ n Indicates phase shift; N indicates harmonic order;

[0014] An image detector is used to synchronously capture light field images in the visible and near-infrared bands, and the deformation of the light spot under the multi-focal length microlens array is calculated. A polarization detection module is introduced to analyze the changes in the light spot under different polarization angles to improve the sensitivity to complex medium aberrations:

[0015]

[0016] Among them, Δr represents the deformation of the light spot; n1 and n2 represent the refractive indices of different bands; r0 and r1 represent the radii of the light spot in the visible light and near-infrared bands.

[0017] Preferably, in the aberration diagnosis step, a light field gradient decomposition model is established based on a multi-focal length microlens array to decompose the aberration into a lateral aberration gradient GL and an axial aberration gradient GA:

[0018]

[0019] Where x and z represent the coordinates of the light spot on the image plane; f represents the focal length of the microlens;

[0020] The aberration propagation relationship is constructed by using the imaging differences of the same light spot by microlenses with different focal lengths:

[0021] D = WX;

[0022] Where D represents the aberration distribution matrix; X represents the aberration characteristic matrix; and W represents the coupling weight matrix between microlenses.

[0023] Preferably, the correction calculation includes:

[0024] Reverse calculation of the optimal focusing displacement Δf of the imaging objective:

[0025]

[0026] Wherein, k represents the aberration sensitivity coefficient of the optical system;

[0027] The aberration compensation model is trained using a generative adversarial network, which inputs the original distorted light field image and outputs the corrected image:

[0028] I′=Ⅰ-G(I,θ);

[0029] Among them, I represents the original light field image; I′ represents the corrected image; G(I,θ) represents the aberration compensation map generated by the GAN model.

[0030] Preferably, the closed-loop feedback includes:

[0031] Aberration monitoring is performed every fixed frame rate T, and the imaging objective lens and digital phase mask parameters are dynamically adjusted:

[0032]

[0033] Among them, θ t System parameters at time step t; θ t+1 represents the system parameters at time step t+1; H represents the light field entropy; α represents the adjustment step size.

[0034] Preferably, the light field entropy evaluation formula is as follows:

[0035]

[0036] Among them, p i Represents the normalized probability of different gray levels of the light field image; N represents the number of gray levels.

[0037] Preferably, the edge response consistency of the multi-focal length microlens is used as the aberration suppression constraint:

[0038]

[0039] Among them, R i represents the edge response of microlens i; represents the average edge response; ∈ represents the error threshold.

[0040] Preferably, the aberration correction calculation further includes zoom optimization:

[0041]

[0042] Where Z represents the zoom optimization objective function; f i represents the focal length of microlens i; λ represents the wavelength.

[0043] Preferably, a regularization constraint is used to improve the stability of aberration correction:

[0044]

[0045] Among them, J represents the optimization objective function; λ represents the regularization weight; θ j represents the system parameters of the jth frame; θj-1 Indicates the system parameters of the j-1th frame.

[0046] The present invention also proposes an automated aberration correction system for remote sensing light field imaging, which is used to implement the above-mentioned automated aberration correction method for remote sensing light field imaging. The automated aberration correction system for remote sensing light field imaging comprises:

[0047] Light field data acquisition module, used to perform dynamic projection encoding and multi-spectral synchronous acquisition, including a near-infrared spot projection device, a multi-focal length microlens array and an image detector, used to obtain aberration-sensitive feature sets;

[0048] An aberration diagnosis module is used to establish a three-dimensional aberration field model based on light field gradient decomposition and a multi-focal length coupling model, including a data processing unit and a feature analysis unit;

[0049] A correction calculation module, which is used to combine physical-driven optimization and data-driven compensation to calculate the optimal focusing parameters and digital phase compensation mask, including an imaging objective lens, a digital phase compensation unit, and an optimization calculation unit;

[0050] The closed-loop feedback control module is used to achieve dynamic adaptive correction based on real-time aberration monitoring and light field entropy minimization optimization, including an aberration monitoring unit, a feedback control unit and a focusing drive unit.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] Aberration diagnosis based on dynamic optical markers: Utilizing spatiotemporally encoded near-infrared spot projection, the aberration response of the optical system is actively stimulated, breaking through the limitations of traditional static calibration methods.

[0053] Multi-focal length coupling correction: Redundant observation is constructed through the focal length diversity of the microlens array to achieve high-dimensional analysis of the aberration field and improve correction accuracy.

[0054] Hybrid optimization compensation: Combining physics-driven optimization (parameter adjustment based on the optical aberration model) and data-driven compensation (deep learning end-to-end correction of high-order aberrations), it takes into account both computational efficiency and the ability to compensate for complex aberrations.

[0055] Real-time closed-loop correction: Adopting the light field entropy minimization strategy, the imaging objective lens group and digital phase mask are continuously monitored and dynamically adjusted to achieve adaptive aberration suppression and local detail enhancement.

[0056] Enhanced remote sensing measurement capabilities: Significantly improve the spatial resolution, depth accuracy, and multispectral imaging consistency of light field imaging systems, making them suitable for high-precision remote sensing mapping and remote sensing observation tasks in complex environments BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0058] Figure 1 A flow chart of the automated aberration correction method for photosensitive field imaging of the present invention;

[0059] Figure 2 This is a functional block diagram of the automated aberration correction method system for photosensitive field imaging of the present invention. DETAILED DESCRIPTION

[0060] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0061] See Figure 1 This embodiment provides an automated aberration correction method for remote sensing light field imaging. The method is based on adaptive correction technology using dynamic projection and multi-focal length light field feedback, and includes the following steps:

[0062] Data acquisition: Through dynamic projection coding and multi-spectral synchronous acquisition, we build an aberration-sensitive feature set and combine it with polarization information to improve the aberration perception ability of complex scenes;

[0063] Aberration diagnosis: Based on the light field gradient decomposition and multi-focal length coupling model, a three-dimensional aberration field model is established;

[0064] Correction calculation: Combine physical drive optimization with data-driven compensation to calculate the optimal focusing parameters and digital phase compensation mask;

[0065] Closed-loop feedback: Dynamic adaptive correction is achieved based on real-time aberration monitoring and light field entropy minimization optimization.

[0066] It is understood that this embodiment proposes an automated aberration correction method for remote sensing light field imaging, which utilizes dynamic projection technology and adaptive correction technology of multi-focal length light field feedback. The specific steps are as follows:

[0067] First, data acquisition is performed: dynamic projection coding and multi-spectral synchronous acquisition are used to construct a set of features sensitive to aberrations. In addition, by incorporating polarization information, the ability to perceive aberrations in complex scenes is further enhanced.

[0068] Then, aberration diagnosis is performed: a three-dimensional aberration field model is constructed based on the light field gradient decomposition technology and the multi-focal length coupling model;

[0069] Then, correction calculations are performed: by combining physical-driven optimization methods and data-driven compensation technology, the optimal focusing parameters and digital phase compensation mask are calculated;

[0070] Finally, a closed-loop feedback loop is implemented: based on real-time aberration monitoring technology and light field entropy minimization optimization strategy, dynamic adaptive correction of aberration is achieved;

[0071] Through the above steps, this embodiment provides an efficient and accurate automatic aberration correction method, which can significantly improve the quality and accuracy of remote sensing light field imaging.

[0072] In some embodiments of the present application, the data collection step includes:

[0073] The near-infrared spot projection device is controlled to generate a sparse spot with random temporal and spatial encoding. The spot position P(x, t) and shape S(x, y, t) change in a pseudo-random sequence in consecutive frames:

[0074]

[0075] Among them, P0 represents the initial spot position; A n represents the amplitude; ω n Indicates frequency; φ n Indicates phase shift; N indicates harmonic order;

[0076] An image detector is used to synchronously capture light field images in the visible and near-infrared bands, and the deformation of the light spot under the multi-focal length microlens array is calculated. A polarization detection module is introduced to analyze the changes in the light spot under different polarization angles to improve the sensitivity to complex medium aberrations:

[0077]

[0078] Among them, Δr represents the deformation of the light spot; n1 and n2 represent the refractive indices of different bands; r0 and r1 represent the radii of the light spot in the visible light and near-infrared bands.

[0079] It is understandable that this embodiment further refines the specific process of data acquisition. In this process, the near-infrared spot projection device not only generates dynamically coded spots, but also has the characteristics of spatiotemporal random coding, that is, the position P(x, t) and shape S(x, y, t) of the spot change in a pseudo-random sequence in consecutive frames. This change can be described by a series of parameters, such as the initial spot position P0, amplitude A, and the position of the spot. n , frequency ω n , phase shift φ n And harmonic order N, etc.

[0080] Furthermore, during data acquisition, image detectors are used to synchronously capture light field images in the visible and near-infrared bands. Analysis of these images allows calculation of the deformation of the light spot under the multifocal microlens array. The magnitude of this deformation can be determined using parameters such as the spot radius r0 and r1 in the visible and near-infrared bands, as well as the refractive indices n1 and n2 in different bands.

[0081] This refined data acquisition process not only improves the accuracy and reliability of data acquisition, but also provides richer and more accurate information for subsequent aberration diagnosis and correction calculations, thereby further improving the performance and accuracy of the entire automated aberration correction system for remote sensing light field imaging.

[0082] In some embodiments of the present application, in the aberration diagnosis step, a light field gradient decomposition model is established based on a multi-focal length microlens array to decompose the aberration into a lateral aberration gradient GL and an axial aberration gradient GA:

[0083]

[0084] Where x and z represent the coordinates of the light spot on the image plane; f represents the focal length of the microlens;

[0085] The aberration propagation relationship is constructed by using the imaging differences of the same light spot by microlenses with different focal lengths:

[0086] D = WX;

[0087] Where D represents the aberration distribution matrix; X represents the aberration characteristic matrix; and W represents the coupling weight matrix between microlenses.

[0088] It is understandable that the present embodiment adopts a more sophisticated model to process aberrations in the aberration diagnosis step. Specifically, based on the multi-focal length microlens array, a light field gradient decomposition model is established. This model can decompose aberrations into lateral aberration gradient GL and axial aberration gradient GA, thereby more accurately understanding the nature of aberrations. Through this model, the imaging differences of the same light spot by microlenses of different focal lengths can be utilized to construct the propagation relationship of aberrations. This relationship can be described by the aberration distribution matrix D, the aberration characteristic matrix X and the coupling weight matrix W between microlenses, providing a powerful mathematical tool for subsequent correction calculations. The refinement of this step not only improves the accuracy of aberration diagnosis, but also provides a more solid theoretical basis for the entire correction process, further enhancing the stability and reliability of the automated aberration correction system for remote sensing light field imaging.

[0089] In some embodiments of the present application, the correction calculation includes:

[0090] Reverse calculation of the optimal focusing displacement Δf of the imaging objective:

[0091]

[0092] Wherein, k represents the aberration sensitivity coefficient of the optical system;

[0093] The aberration compensation model is trained using a generative adversarial network, which inputs the original distorted light field image and outputs the corrected image:

[0094] I′=Ⅰ-G(I,θ);

[0095] Among them, I represents the original light field image; I′ represents the corrected image; G(I,θ) represents the aberration compensation map generated by the GAN model.

[0096] It can be understood that in the correction calculation step, this embodiment adopts a more accurate and efficient method to determine the optimal focusing displacement Δf of the imaging objective lens. This calculation process takes into account the aberration sensitivity coefficient k of the optical system, so that the focal length of the imaging system can be more accurately predicted and adjusted to achieve the best imaging effect. In addition, a generative adversarial network (GAN) is used to train the aberration compensation model. This model can receive the original distorted light field image as input and output a corrected image. Through the powerful learning ability of the GAN model, effective compensation for aberrations can be achieved, further improving the imaging quality. Specifically, the corrected image I′ is obtained by processing the original light field image I with the GAN model G(I,θ), where θ represents the parameters of the GAN model. The introduction of this step not only improves the efficiency and accuracy of the correction calculation, but also provides a more powerful image processing capability for the entire automated aberration correction system for remote sensing light field imaging, enabling it to better adapt to complex and changing imaging environments.

[0097] In some embodiments of the present application, the closed-loop feedback includes:

[0098] Aberration monitoring is performed every fixed frame rate T, and the imaging objective lens and digital phase mask parameters are dynamically adjusted:

[0099]

[0100] Among them, θ t System parameters at time step t; θ t+1 represents the system parameters at time step t+1; H represents the light field entropy; α represents the adjustment step size.

[0101] It is understandable that in the closed-loop feedback step, this embodiment adopts a real-time monitoring and dynamic adjustment strategy to ensure the continued effectiveness of the correction process. Specifically, the system will perform aberration monitoring once every fixed frame rate T. The setting of this frequency is intended to balance the real-time nature of the correction and the computational burden of the system. In each monitoring, the system will evaluate the imaging quality based on the current light field entropy H, and dynamically adjust the imaging objective and digital phase mask parameters accordingly. This adjustment is performed iteratively, and each step of the adjustment will take into account the system parameters θ of the previous step. t , and the adjustment step α calculated by the light field entropy H, so as to obtain the next system parameter θ t+1 This process ensures that the correction system can continuously track and adapt to changes in the imaging environment, maintaining high stability and precision in imaging quality. Through this closed-loop feedback mechanism, the automated aberration correction system for remote sensing light field imaging of this embodiment can not only achieve accurate aberration correction in the initial stage, but also maintain excellent imaging performance throughout the entire imaging process.

[0102] In some embodiments of the present application, the light field entropy evaluation formula is as follows:

[0103]

[0104] Among them, p i Represents the normalized probability of different gray levels of the light field image; N represents the number of gray levels.

[0105] It is understandable that this embodiment also proposes a light field entropy evaluation formula for quantifying the imaging quality of light field images. This formula is based on the normalized probability p of different gray levels in the light field image. iThe entropy of the light field is calculated using the number of grayscale levels N, accurately reflecting the detail richness and uniformity of the light field image. Using this evaluation formula to monitor image quality in real time during the closed-loop feedback step ensures that the correction system can accurately capture changes in the imaging environment and make timely adjustments, thereby maintaining high stability and precision in image quality. This innovative light field entropy evaluation formula not only enhances the intelligence level of automated aberration correction systems for remote sensing light field imaging, but also lays a solid foundation for their widespread adoption in practical applications.

[0106] In some embodiments of the present application, the edge response consistency of the multi-focal length microlens is used as the aberration suppression constraint:

[0107]

[0108] Among them, R i represents the edge response of microlens i; represents the average edge response; ∈ represents the error threshold.

[0109] It is understandable that this embodiment also proposes an innovative aberration suppression strategy to enhance the correction effect. Specifically, the edge response consistency of multi-focal length microlenses is used as a constraint condition for aberration suppression. In actual operation, the edge response Ri of each microlens i is calculated and compared with the average edge response. When the deviation between the edge response of a certain microlens and the average edge response exceeds a preset error threshold ∈, it is considered that the microlens has aberrations, and is corrected accordingly. The implementation of this strategy can not only effectively suppress the generation of aberrations, but also further improve the stability and accuracy of the imaging system. In this way, the automated aberration correction system for remote sensing light field imaging of this embodiment can not only maintain excellent performance in complex imaging environments, but also inject new vitality into the development of the field of remote sensing imaging.

[0110] In some embodiments of the present application, the aberration correction calculation further includes zoom optimization:

[0111]

[0112] Where Z represents the zoom optimization objective function; f i represents the focal length of microlens i; λ represents the wavelength.

[0113] It is understandable that the present embodiment further considers the application of zoom optimization in aberration correction calculations. Zoom optimization aims to achieve a wider imaging range and higher imaging quality by adjusting the focal length of the microlens. In this process, a zoom optimization objective function Z is defined, which takes into account parameters such as the focal length fi of the microlens i and the wavelength λ. By optimizing this objective function, the best focal length combination can be found, thereby expanding the field of view of the imaging system while maintaining high resolution. The introduction of this step not only enhances the flexibility of the correction system, but also provides a broader space for its application in various imaging tasks. In this way, the automated aberration correction system for remote sensing light field imaging of this embodiment can not only achieve accurate aberration correction in a static environment, but also maintain stable imaging performance in dynamically changing scenes, making an important contribution to the technological progress and application expansion in the field of remote sensing imaging.

[0114] In some embodiments of the present application, regularization constraints are used to improve the stability of aberration correction:

[0115]

[0116] Among them, J represents the optimization objective function; λ represents the regularization weight; θ j represents the system parameters of the jth frame; θ j-1 Indicates the system parameters of the j-1th frame.

[0117] It is understandable that this embodiment also proposes a method for improving the stability of aberration correction by using regularization constraints. In actual operation, the correction process is often affected by various noises and interferences, which may lead to instability of the correction results. In order to overcome this problem, this embodiment introduces a regularization constraint and applies it to the optimization objective function J. The regularization weight λ is used to balance the relationship between correction accuracy and stability, ensuring that the stability of the system is maintained while pursuing high-precision correction. In addition, the continuity of system parameters between adjacent frames is also considered, that is, the relationship between the system parameter θj of the j-th frame and the system parameter θj-1 of the j-1-th frame. Through this constraint, the fluctuations in the correction process can be further reduced and the robustness of the entire system can be improved. The introduction of this innovative regularization constraint method not only enhances the stability of the automated aberration correction system for remote sensing light field imaging, but also provides a strong guarantee for its long-term stable operation in complex and changeable imaging environments.

[0118] See Figure 2 This embodiment further proposes an automated aberration correction system for remote sensing light field imaging, which is used to implement the automated aberration correction method for remote sensing light field imaging. The automated aberration correction system for remote sensing light field imaging includes:

[0119] Light field data acquisition module, used to perform dynamic projection encoding and multi-spectral synchronous acquisition, including a near-infrared spot projection device, a multi-focal length microlens array and an image detector, used to obtain aberration-sensitive feature sets;

[0120] An aberration diagnosis module is used to establish a three-dimensional aberration field model based on light field gradient decomposition and a multi-focal length coupling model, including a data processing unit and a feature analysis unit;

[0121] A correction calculation module, which is used to combine physical-driven optimization and data-driven compensation to calculate the optimal focusing parameters and digital phase compensation mask, including an imaging objective lens, a digital phase compensation unit, and an optimization calculation unit;

[0122] The closed-loop feedback control module is used to achieve dynamic adaptive correction based on real-time aberration monitoring and light field entropy minimization optimization, including an aberration monitoring unit, a feedback control unit and a focusing drive unit.

[0123] It will be appreciated that this embodiment further describes in detail the various modules and functions of the automated aberration correction system for remote sensing light field imaging. The light field data acquisition module, as the front end of the entire system, undertakes the important task of dynamic projection coding and multi-spectral synchronous acquisition. This module uses a near-infrared spot projection device to generate a sparse light spot with spatiotemporal random coding characteristics. It utilizes a multi-focal microlens array and image detector to synchronously capture light field images in the visible and near-infrared bands, thereby acquiring a rich set of aberration-sensitive features. These feature sets provide a solid foundation for subsequent aberration diagnosis and correction.

[0124] The aberration diagnosis module builds a three-dimensional aberration field model based on the acquired light field data, utilizing light field gradient decomposition and a multi-focal length coupling model. This model accurately describes the spatial distribution of aberrations, providing a powerful mathematical tool for subsequent correction calculations. The data processing unit and feature analysis unit play a key role in this module, processing and analyzing the light field data to extract key aberration features.

[0125] The correction calculation module is the core of the system. It combines physical-driven optimization and data-driven compensation to calculate optimal focusing parameters and digital phase compensation masks. The imaging objective, digital phase compensation unit, and optimization calculation unit work together to achieve precise correction of the imaging system. By reverse-calculating the optimal focusing displacement of the imaging objective and using a generative adversarial network to train an aberration compensation model, the correction calculation module can output a corrected, high-quality image.

[0126] The closed-loop feedback control module monitors image quality in real time and dynamically adjusts imaging parameters based on a light field entropy minimization optimization strategy to achieve adaptive correction. The aberration monitoring unit monitors changes in aberrations, and the feedback control unit adjusts the imaging objective and digital phase mask parameters based on the monitoring results. The focus drive unit is responsible for executing the specific adjustments. Through this closed-loop feedback mechanism, the system can continuously track and adapt to changes in the imaging environment, maintaining high stability and precision in image quality.

[0127] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or a combination of software and hardware embodiments. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. An automated aberration correction method for remote sensing light field imaging, characterized in that: include: Data acquisition: Through dynamic projection coding and multi-spectral synchronous acquisition, we build an aberration-sensitive feature set and combine it with polarization information to improve the aberration perception ability of complex scenes; Aberration diagnosis: Based on the light field gradient decomposition and multi-focal length coupling model, a three-dimensional aberration field model is established; Correction calculation: Combine physical drive optimization with data-driven compensation to calculate the optimal focusing parameters and digital phase compensation mask; Closed-loop feedback: Dynamic adaptive correction is achieved based on real-time aberration monitoring and light field entropy minimization optimization.

2. The automated aberration correction method for remote sensing light field imaging according to claim 1, wherein: The data collection steps include: The near-infrared spot projection device is controlled to generate a sparse spot with random temporal and spatial encoding. The spot position P(x, t) and shape S(x, y, t) change in a pseudo-random sequence in consecutive frames: Among them, P0 represents the initial spot position; A n represents the amplitude; ω n Indicates frequency; φ n Indicates phase shift; N indicates harmonic order; An image detector is used to synchronously capture light field images in the visible and near-infrared bands, and the deformation of the light spot under the multi-focal length microlens array is calculated. A polarization detection module is introduced to analyze the changes in the light spot under different polarization angles to improve the sensitivity to complex medium aberrations: Among them, Δr represents the deformation of the light spot; n1 and n2 represent the refractive indices of different bands; r0 and r1 represent the radii of the light spot in the visible light and near-infrared bands.

3. The automated aberration correction method for remote sensing light field imaging according to claim 2, wherein: In the aberration diagnosis step, a light field gradient decomposition model is established based on the multi-focal length microlens array to decompose the aberration into the lateral aberration gradient GL and the axial aberration gradient GA: Where x and z represent the coordinates of the light spot on the image plane; f represents the focal length of the microlens; The aberration propagation relationship is constructed by using the imaging differences of the same light spot by microlenses with different focal lengths: D = WX; Where D represents the aberration distribution matrix; X represents the aberration characteristic matrix; and W represents the coupling weight matrix between microlenses.

4. The automated aberration correction method for remote sensing light field imaging according to claim 3, wherein: The correction calculation includes: Reverse calculation of the optimal focusing displacement Δf of the imaging objective: Wherein, k represents the aberration sensitivity coefficient of the optical system; The aberration compensation model is trained using a generative adversarial network, which inputs the original distorted light field image and outputs the corrected image: I′=IG(I,θ); Among them, I represents the original light field image; I′ represents the corrected image; G(I,θ) represents the aberration compensation map generated by the GAN model.

5. The automated aberration correction method for remote sensing light field imaging according to claim 4, wherein: The closed-loop feedback includes: Aberration monitoring is performed every fixed frame rate T, and the imaging objective lens and digital phase mask parameters are dynamically adjusted: Among them, θ t System parameters at time step t; θ t+1 represents the system parameters at time step t+1; H represents the light field entropy; α represents the adjustment step size.

6. The automated aberration correction method for remote sensing light field imaging according to claim 5, wherein: The light field entropy evaluation formula is as follows: Among them, p i Represents the normalized probability of different gray levels of the light field image; N represents the number of gray levels.

7. The automated aberration correction method for remote sensing light field imaging according to claim 6, wherein: The edge response consistency of multi-focal length microlenses is used as the aberration suppression constraint: Among them, R i represents the edge response of microlens i; represents the average edge response; ∈ represents the error threshold.

8. The automated aberration correction method for remote sensing light field imaging according to claim 7, wherein: The aberration correction calculation also includes zoom optimization: Where Z represents the zoom optimization objective function; f i represents the focal length of microlens i; λ represents the wavelength.

9. The automated aberration correction method for remote sensing light field imaging according to claim 8, wherein: Improving the stability of aberration correction using regularization constraints: Among them, J represents the optimization objective function; λ represents the regularization weight; θ j represents the system parameters of the jth frame; θ j-1 Indicates the system parameters of the j-1th frame.

10. An automated aberration correction system for remote sensing light field imaging, used to implement the automated aberration correction method for remote sensing light field imaging according to any one of claims 1 to 9, characterized in that: The automated aberration correction system for remote sensing light field imaging comprises: Light field data acquisition module, used to perform dynamic projection encoding and multi-spectral synchronous acquisition, including a near-infrared spot projection device, a multi-focal length microlens array and an image detector, used to obtain aberration-sensitive feature sets; An aberration diagnosis module is used to establish a three-dimensional aberration field model based on light field gradient decomposition and a multi-focal length coupling model, including a data processing unit and a feature analysis unit; A correction calculation module, which is used to combine physical-driven optimization and data-driven compensation to calculate the optimal focusing parameters and digital phase compensation mask, including an imaging objective lens, a digital phase compensation unit, and an optimization calculation unit; The closed-loop feedback control module is used to achieve dynamic adaptive correction based on real-time aberration monitoring and light field entropy minimization optimization, including an aberration monitoring unit, a feedback control unit and a focusing drive unit.

Images