A dynamic encryption imaging method and device based on orbital angular momentum speckle

Abstract

The application provides a dynamic encryption imaging method and device based on orbital angular momentum speckle, and the method comprises the following steps: step 1, generating an initial Gaussian light beam output by a coherent light source, and adjusting the polarization state of the initial Gaussian light beam through a polarization control unit, and adjusting the aperture of the light beam through an expansion optical system; step 2, generating an orbital angular momentum vortex light beam corresponding to a topological charge number; step 3, forming an orbital angular momentum speckle field corresponding to the topological charge number one by one, and taking the orbital angular momentum speckle field as a dynamic encryption key; step 4, continuously illuminating a dynamic target and realizing information multiplexing coding; step 5, obtaining an encrypted intensity image carrying multiple frames of dynamic information; and step 6, outputting multiple frames of dynamic imaging results. The application realizes efficient optical encryption imaging of a dynamic scene under single-exposure conditions, and has the technical effects of large key space, strong controllability, simple system structure, strong anti-interference ability and the like.

Description

Technical Field

[0001] This invention relates to the fields of optical imaging and information security, specifically to a dynamic encrypted imaging method and apparatus based on orbital angular momentum speckle. Background Technology

[0002] Optical encryption has garnered widespread attention in information security, communication confidentiality, and optical imaging due to its high parallelism and high-speed information processing capabilities. For dynamic scene imaging, achieving high-quality, multi-frame rapid acquisition while ensuring information security is a core research challenge.

[0003] Existing optical encryption technologies generally involve a trade-off between key randomness, controllability, and system feasibility: while highly random keys offer stronger security, they are difficult to precisely control and replicate; highly controllable keys facilitate system implementation and debugging, but their encryption strength is limited; and schemes relying on complex, high-dimensional degrees of freedom, while theoretically possessing high security, often suffer from complex system structures and insufficient stability. This inherent contradiction restricts the simultaneous achievement of high security and system reliability within the same imaging system.

[0004] Orbital angular momentum (OAM) beams, due to their high degree of freedom in topological charge number and large theoretical dimensionality, are considered a potential means of high-dimensional optical encryption. However, existing research on OAM-based imaging encryption still has significant shortcomings. On the one hand, related work mostly focuses on the encoding and transmission of one-dimensional numerical information or the reconstruction of static lattice targets, with less research on imaging encryption for two-dimensional dynamic scenes, failing to fully utilize the encoding potential of OAM modes in the spatiotemporal dimension. On the other hand, the inherent phase singularity of OAM beams results in a ring-shaped intensity distribution, with a significant central blind zone in higher-order modes, making it difficult to provide uniform or controllable illumination conditions for the target, thus affecting the complete acquisition of spatial information. In addition, OAM modes are highly sensitive to perturbations such as scattering media, refractive index fluctuations, and optical aberrations, easily inducing modal crosstalk and reducing the stability and reliability of the encryption and decryption process under non-ideal conditions. Furthermore, existing research focuses more on mode multiplexing and transmission performance optimization in the field of communication, lacking a systematic encoding-modulation-decryption framework for two-dimensional dynamic imaging, making it difficult to achieve effective encryption and reconstruction of multi-frame dynamic information under single-exposure conditions.

[0005] Therefore, there is an urgent need to propose a dynamic imaging encryption method and device based on orbital angular momentum modulation, which can achieve controllable modulation and high security of encryption keys while ensuring physical stability, and is applicable to multi-frame dynamic imaging under single exposure conditions, so as to fully explore the application potential of high-dimensional modal freedom of orbital angular momentum in dynamic imaging encryption. Summary of the Invention

[0006] Purpose of the invention: To address the problems of existing optical encryption imaging technologies in dynamic scene imaging, such as difficulty in balancing key controllability and security, limited imaging dimensions, insufficient ability to acquire multi-frame information under single-exposure conditions, and low system stability, this invention provides a dynamic encryption imaging method and device based on orbital angular momentum speckle.

[0007] The method includes the following steps:

[0008] Step 1: Generate an initial Gaussian beam output from a coherent light source, and control the polarization state of the initial Gaussian beam through a polarization control unit (the polarization control unit is existing technology, consisting of a half-wave plate and a thin-film linear polarizer). At the same time, adjust the beam aperture through a beam expanding optical system (the beam expanding optical system is existing technology, and the one used here is a variable magnification beam expander from Daheng Optoelectronics) to meet the spatial sampling conditions for subsequent phase modulation and speckle modulation.

[0009] Step 2: Using a spatial light modulator, a spiral phase distribution with different topological charges is loaded onto a Gaussian beam to generate an orbital angular momentum vortex beam with the corresponding topological charge.

[0010] Step 3: The orbital angular momentum vortex beam is modulated by a speckle modulator (the speckle modulator is existing technology, and here it is a piece of frosted glass) to form an orbital angular momentum speckle field that corresponds one-to-one with the topological charge number, and the orbital angular momentum speckle field is used as a dynamic encryption key.

[0011] Step 4: Switch the orbital angular momentum speckle key corresponding to different topological charge numbers or different combinations of topological charge numbers according to the preset timing or encoding rules, continuously illuminate the dynamic target and realize information multiplexing encoding;

[0012] Step 5: Under single-exposure conditions, acquire the target intensity information after illumination by the orbital angular momentum speckle key to obtain an encrypted intensity image carrying multiple frames of dynamic information;

[0013] Step 6: Input the encrypted strength image into the decryption and reconstruction network, combine it with the corresponding orbital angular momentum speckle key information, demultiplex and reconstruct the multi-frame dynamic information, and output the multi-frame dynamic imaging results.

[0014] In step 2, the topological charge number is an integer or fractional value, and the orbital angular momentum speckle fields corresponding to different topological charge numbers are statistically distinguishable from each other.

[0015] In step 3, the speckle modulation device includes a scattering medium, a random phase plate, a rough surface, or a combination thereof.

[0016] In step 4, the preset timing or encoding rules include switching different orbital angular momentum modes in chronological order, or combining two or more orbital angular momentum modes within the same time period.

[0017] At any moment The speckle key is generated from the following parameter vector:

[0018] ,

[0019] in It is the key space and parameter vector related to the orbital angular momentum (OAM) degrees of freedom. Represented as:

[0020] ,

[0021] in Represents the moment Topological charge combination of orbital angular momentum (OAM); The x and y coordinates represent the spatial location of the k-th orbital angular momentum OAM mode, respectively. For the first In this method, the weight coefficients at different positions at each time point are all taken as 1. This represents the number of combinations of orbital angular momentum (OAM) modes at time t.

[0022] After propagation through frosted glass scattering, the multiplexed orbital angular momentum (OAM) beam is converted into a speckle illumination field. The speckle pattern generated at time t is used as the physical encryption key, expressed as:

[0023] ,

[0024] in This represents the speckle pattern generated at time t. It is a scattering medium;

[0025] The generation of each orbital angular momentum (OAM) speckle is controlled by the number of OAM modes and their combination; using time-varying speckle fields to illuminate moving objects, the modulated intensity distribution is obtained as follows:

[0026] ,

[0027] in Let be the intensity distribution of the moving object at time t after passing through a time-varying speckle field illumination scheme. This represents the background noise during the measurement process. Indicates at time The object function.

[0028] In step 5, during a single exposure, the dynamic target undergoes continuous modulation by two or more different orbital angular momentum speckle keys within the exposure time, thereby mapping the time series information onto the encryption strength image.

[0029] By utilizing time-multiplexed speckle illumination in a single exposure, a compressed and encrypted measurement image was acquired, represented as follows:

[0030] ,

[0031] in This represents the compressed and encrypted measurement value of a single frame captured by the camera, where T represents the number of frames in the encrypted video; the temporal information of the dynamic scene is compressed onto a two-dimensional measurement map, using a key space. A high-dimensional, time-independent speckle key is obtained, thereby enabling encryption.

[0032] In step 6, the decryption and reconstruction network is a learning network based on physical model constraints, a data-driven learning network, or a hybrid network structure combining a learning network based on physical model constraints and a data-driven learning network.

[0033] In step 6, the decryption and reconstruction network includes a U-net backbone, a coarse-to-fine feature fusion module C2FM, a feature enhancement module, an encoder, and a decoder;

[0034] At the encoder-decoder connection and in the decoder section, a coarse-to-fine feature fusion module (C2FM) is inserted.

[0035] The coarse-to-fine feature fusion module C2FM includes two parallel branches: one for coarse feature extraction and the other for fine-grained feature extraction. The features from both branches are then fused and output. First, the input to the coarse-to-fine feature fusion module C2FM is divided according to the number of frames in the reconstructed sequence. Group, pair The mean of the group features is calculated and used as the coarse feature extraction branch, represented as follows: The grouped features are then compared with the first group of features in each of the n groups to calculate their positional offset vectors. Feature alignment is then performed using pixel interpolation, as shown below:

[0036] ,

[0037] in In reference space Below, the feature tensor at time t after feature alignment;

[0038] in Represents the t-th feature. This represents the spatial feature bias mapping of the t-th set of features to the reference space. This represents the bilinear interpolation kernel. For the source feature space domain, Represents continuous sampling coordinates in the reference space. Represents discrete source pixels;

[0039] The aligned features are fused with the original grouped features to generate fine-grained branch features;

[0040] The coarse feature extraction branch and the fine feature extraction branch are concatenated, and then residual fusion is performed with the initial features after feature extraction, which are then combined with the features of the grouped initial features.

[0041] The feature enhancement module comprises four branches. The first three branches are the difference between multi-scale convolution kernel features and max pooling features. The multi-scale convolution kernels are, in order, convolution kernels with a size of 3. of The kernel size is 5. of The kernel size is of The fourth branch is the Directional Feature Extraction (DFE) module. The DFE module uses convolutional kernels with different directions to extract features in the horizontal and vertical directions, then concatenates and reduces the dimensionality. Finally, it performs residual linking with the input features. The process is represented as follows:

[0042] ,

[0043] in The feature map representing the input direction feature extraction module (DFE). The kernel size is The convolution operator extracts horizontal features. The kernel size is The convolution operator extracts vertical features, and [;] represents the concatenation of channel features. Adjust the channel operator. The output features of the Directional Feature Extraction (DFE) module represent the direction of the feature extraction.

[0044] The training method for the decryption and reconstruction network includes: training loss Defined as:

[0045] ,

[0046] in It is a preset weight. Indicates the reconstructed image. SSIM represents the structural similarity between two images, where the image is labeled with the ground truth (GT).

[0047] The present invention also provides a dynamic encryption imaging device based on orbital angular momentum speckle implemented by the method described above, comprising a coherent light source module, a beam control module, a spatial light modulation module, a speckle modulation module, an imaging acquisition module, and a decryption and reconstruction module;

[0048] The coherent light source module is used to generate an initial Gaussian beam;

[0049] The beam control module is used to perform polarization control and beam expansion on the initial Gaussian beam;

[0050] The spatial light modulation module is used to load helical phase distributions with different topological charges to generate orbital angular momentum vortex beams.

[0051] The speckle modulation module is used to modulate the orbital angular momentum vortex beam into an orbital angular momentum speckle field that corresponds one-to-one with the topological charge number, and the orbital angular momentum speckle field serves as a dynamic encryption key.

[0052] The imaging acquisition module is used to acquire encrypted images carrying multiple frames of dynamic information under single exposure conditions.

[0053] The decryption and reconstruction module is used to demultiplex and reconstruct the encrypted strength image by combining the orbital angular momentum speckle key information.

[0054] The spatial light modulation module is configured to load spiral phase distributions corresponding to different topological charges or combinations thereof according to a preset time sequence or encoding rules within a single exposure time, so as to achieve time switching of orbital angular momentum modes.

[0055] The speckle modulation module is used to generate orbital angular momentum speckle fields that correspond to different topological charge numbers and are statistically distinguishable from each other, so as to form a dynamic encryption key.

[0056] The speckle modulation module is a replaceable or reconfigurable structure to enable dynamic updating of the orbital angular momentum speckle key.

[0057] The orbital angular momentum speckle key is defined by at least one of the following: the topological charge number value, the topological charge number combination method, and the topological charge number combination time switching order.

[0058] The decryption and reconstruction module includes a pre-trained decryption and reconstruction network, which uses prior information of the orbital angular momentum speckle key to demultiplex multiple frames of dynamic information in the encrypted strength image.

[0059] Without the corresponding orbital angular momentum speckle key information, the encrypted strength image cannot reconstruct a dynamic imaging result with temporal consistency.

[0060] This invention introduces a dynamic encryption mechanism combining orbital angular momentum mode modulation and speckle coding, achieving secure multiplexing and stable decryption and reconstruction of multi-frame dynamic information under single-exposure conditions. Compared with existing technologies, this invention has the following advantages:

[0061] (1) This invention utilizes the combination of orbital angular momentum mode and speckle modulation to construct a dynamic encryption key, which significantly expands the encryption key space and improves information security in the dynamic imaging process while maintaining the stability and realizability of the physical system.

[0062] (2) By switching different orbital angular momentum speckle keys according to preset time series or encoding rules, the multiplexing encoding of dynamic information of multiple frames under single exposure conditions is realized, breaking through the acquisition limitation of traditional imaging systems in the time dimension and improving dynamic imaging efficiency.

[0063] (3) The present invention uses a single camera device to complete the encrypted information acquisition, without the need for a multi-channel or multi-camera structure. The system composition is simple and easy to implement and integrate with engineering.

[0064] (4) The decryption and reconstruction network combining the prior information of the orbital angular momentum speckle key effectively improves the stability and robustness of the multi-frame dynamic information demultiplexing process, enabling the present invention to obtain reliable dynamic imaging results in complex imaging environments. Attached Figure Description

[0065] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0066] Figure 1 This is a flowchart of the method of the present invention.

[0067] Figure 2 This is a schematic diagram of the specific structure of the decryption network in an embodiment of the present invention.

[0068] Figure 3 This is a schematic diagram of the specific structure of the experimental apparatus in an embodiment of the present invention.

[0069] Figure 4 This is a schematic diagram of the simulation results in an embodiment of the present invention.

[0070] Figure 5 This is a schematic diagram of the actual experimental results in the embodiments of the present invention. Detailed Implementation

[0071] like Figure 1 As shown, this embodiment provides a dynamic densification imaging method based on orbital angular momentum speckle, including the following steps:

[0072] Step 1: Generate an initial Gaussian beam output from a coherent light source, and control the polarization state of the initial Gaussian beam through a polarization control unit. At the same time, adjust the beam aperture through a beam expanding optical system to meet the spatial sampling conditions for subsequent phase modulation and speckle modulation.

[0073] Step 2: Using a spatial light modulator, a spiral phase distribution with different topological charges is loaded onto the Gaussian beam to generate an orbital angular momentum vortex beam with corresponding topological charges.

[0074] Step 3: Modulate the orbital angular momentum vortex beam using a speckle modulator to form an orbital angular momentum speckle field that corresponds one-to-one with the topological charge number, and use the orbital angular momentum speckle field as a dynamic encryption key.

[0075] Step 4: Switch the orbital angular momentum speckle key corresponding to different topological charge numbers or their combinations according to the preset timing or encoding rules, continuously illuminate the dynamic target and realize information multiplexing encoding;

[0076] Step 5: Under single-exposure conditions, acquire the target intensity information after illumination by the orbital angular momentum speckle key to obtain an encrypted intensity image carrying multiple frames of dynamic information;

[0077] Step 6: Input the encrypted strength image into the decryption and reconstruction network, combine it with the corresponding orbital angular momentum speckle key information, demultiplex and reconstruct the multi-frame dynamic information, and output the multi-frame dynamic imaging results.

[0078] The imaging and encryption framework employed in this invention is based on time-multiplexed speckle illumination and is controlled by a dynamically designed key based on orbital angular momentum (OAM). In this framework, the encryption key is physically embodied in the speckle pattern itself, while a time-varying set of parameters governs the specific generation of each speckle pattern. At time... The speckle key is generated from the following parameter vector: ,in It is the key space related to the orbital angular momentum (OAM) degrees of freedom, and the relevant parameter vector is represented as follows: ,in Represents the moment The topological charge combination of orbital angular momentum (OAM). The x and y coordinates represent the spatial location of the k-th orbital angular momentum OAM mode, respectively. For the first In this method, the weight coefficients at different positions at each time point are all set to 1. Represents the first The number of combinations of orbital angular momentum (OAM) modes at each time step. After propagation through frosted glass scattering, the multiplexed OAM beam is converted into a speckle illumination field. At time step... The generated speckle pattern is used as the physical encryption key, and its expression is written as: The generation of each orbital angular momentum (OAM) speckle is controlled by the number and combination of OAM modes. Using time-varying speckle fields to illuminate a moving object, the modulated intensity distribution is obtained as follows:

[0079] ,

[0080] in Indicates at time The object function. By utilizing time-multiplexed speckle illumination in a single exposure, the system acquired a compressed and encrypted measurement, represented as:

[0081] ,

[0082] in The background noise during the measurement process is represented by T, and the number of frames in the encrypted video is represented by T. The temporal information of the dynamic scene is compressed onto a two-dimensional measurement map, using the orbital angular momentum (OAM) parameter space. A high-dimensional, time-independent speckle key is obtained for encryption. The encrypted measurements and the calibrated speckle code are fed into the reconstruction network, which outputs the reconstructed decrypted video sequence.

[0083] The decryption and reconstruction network is data-driven, and its structure includes a U-Net backbone, a coarse-to-fine fusion module (C2FM), and a feature enhancement module. Figure 2 The detailed structure of the encrypted compressed video reconstruction network is shown. The overall network structure consists of an encoder and a decoder, with coarse-to-fine feature fusion modules (C2FM) inserted at the connection points between the two and in the decoder section. C2FM includes two parallel branches: one for coarse feature extraction and the other for fine-grained feature extraction. The features from both branches are then fused and the output is presented. First, the input to the coarse-to-fine feature fusion module C2FM is divided according to the number of frames in the reconstructed sequence. Group, for this The mean of the group features is calculated and used as the coarse feature extraction branch, represented as follows: The grouped features are then compared with the first group of features to calculate positional offset vectors, and feature alignment is performed using pixel interpolation. The structure is as follows:

[0084] ,

[0085] in In reference space Below, the feature tensor at time t after feature alignment; This represents the t-th feature group. This represents the spatial feature bias mapping of the t-th set of features to the reference space. This represents the bilinear interpolation kernel. For the source feature space domain, It represents the continuous sampling coordinates of the reference space. Representing discrete source pixels. The aligned features are fused with the original grouped features to generate fine-grained branch features. The coarse feature branches and fine-grained feature branches are concatenated and then residually fused with the initial grouped features after feature extraction.

[0086] The Feature Enhanced Module (FEM) comprises four main branches. The first three branches are the difference between multi-scale convolutional kernel features and max-pooling features. The multi-scale convolutional kernels are, in order: , , The fourth branch is the Directional Feature Extraction (DFE) module. The DFE module uses convolutional kernels with different orientations to extract features in the horizontal and vertical directions, then concatenates and reduces the dimensionality before performing residual linking with the input features. The process is represented as follows:

[0087] ,

[0088] in Feature map representing the input DFE module. The kernel size is The convolution operator extracts horizontal features. The kernel size is The convolution operator extracts vertical features; [;] represents channel feature concatenation. The channel adjustment operator facilitates residual connection with the original features. This represents the output characteristics of the DFE module.

[0089] The specific training method is as follows: The training loss combines MSE loss and SSIM loss, and is defined as:

[0090] ,

[0091] in Set to 0.3, Indicates the reconstructed image. The image represents the ground truth (GT) image. The network's learning rate was set to 0.0004, and it employed an SGD optimizer and cosine annealing measurements. Training was performed using the PyTorch framework, with computations executed on an NVIDIA A40 GPU.

[0092] like Figure 3 As shown, this embodiment of the invention also provides a dynamic encryption imaging device based on orbital angular momentum speckle for implementing the above method. The device includes a coherent light source module, a beam control module, a spatial light modulation module, a speckle modulation module, an imaging acquisition module, and a decryption and reconstruction module.

[0093] The system comprises the following modules: a coherent light source module to generate an initial Gaussian beam with spatial coherence; a beam manipulation module located on the output path of the coherent light source module to perform polarization state manipulation and beam shaping on the initial Gaussian beam; a spatial light modulation module to load helical phase distributions with different topological charges onto the beam to generate orbital angular momentum vortex beams with corresponding topological charges, and to switch between different orbital angular momentum modes according to a preset time sequence or encoding rules; and a speckle modulation module to perform speckle modulation on the orbital angular momentum vortex beams to form orbital angular momentum speckle fields corresponding one-to-one with different topological charges, which serve as encryption keys in the dynamic encrypted imaging process.

[0094] The beam output from the coherent light source module sequentially passes through the beam control module, spatial light modulation module, and speckle modulation module before illuminating the dynamic target under test. The light signal modulated by the target is acquired by the imaging acquisition module under single-exposure conditions to obtain an encrypted strength image carrying multiple frames of dynamic information. The decryption and reconstruction module receives the encrypted strength image obtained by the imaging acquisition module and, in conjunction with the corresponding orbital angular momentum speckle key information, demultiplexes and reconstructs the multiple frames of dynamic information in the encrypted strength image.

[0095] Figure 4 In the middle (a), the phase amplitude diagram of the OAM key at different times is shown, including the position of different OAM codes and the corresponding topology charge value at different positions; Figure 4 In the middle (b), (d), and (f), the labels of the true values ​​in the simulation dataset are shown, and from top to bottom they represent single-frame images in the time series. Figure 4 (c), (e), and (g) are time series reconstructed through the decryption network, such as Figure 4 As shown, the reconstruction results have a high peak signal-to-noise ratio and structural consistency with the ground truth labels; numerical simulation was used to verify the ability of the method of the present invention to reuse and reconstruct dynamic information of multiple frames under single exposure conditions. Figure 5 (a) shows the phase diagram of the OAM key at different times in the real experiment; Figure 5(b) represents the OAM speckle modulated by the speckle modulator in the actual experiment; Figure 5 In the experiment, (g)-(j) are the single-frame encrypted compression measurement values ​​of the football symbol, car symbol, handwritten number 7 and handwritten letter V. Figure 5 Figures (c) to (f) show the time series reconstructed through the decryption network. Real optical experiments further verified the encryption imaging and decryption reconstruction performance of this invention in dynamic scenes. Experimental results show that this invention, while maintaining a simple system structure, can achieve efficient encryption and stable decryption of dynamic information, demonstrating good practicality and promotional value.

[0096] This invention provides a dynamic encryption imaging method and apparatus based on orbital angular momentum speckle. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment of the invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. A dynamic encryption imaging method based on orbital angular momentum speckle, characterized in that, Includes the following steps: Step 1: Generate an initial Gaussian beam output from a coherent light source, and control the polarization state of the initial Gaussian beam through a polarization control unit. At the same time, adjust the beam aperture through a beam expanding optical system to meet the spatial sampling conditions for subsequent phase modulation and speckle modulation. Step 2: Using a spatial light modulator, a spiral phase distribution with different topological charges is loaded onto a Gaussian beam to generate an orbital angular momentum vortex beam with the corresponding topological charge. Step 3: Modulate the orbital angular momentum vortex beam using a speckle modulator to form an orbital angular momentum speckle field that corresponds one-to-one with the topological charge number, and use the orbital angular momentum speckle field as a dynamic encryption key. Step 4: Switch the orbital angular momentum speckle key corresponding to different topological charge numbers or different combinations of topological charge numbers according to the preset timing or encoding rules, continuously illuminate the dynamic target and realize information multiplexing encoding; Step 5: Under single-exposure conditions, acquire the target intensity information after illumination by the orbital angular momentum speckle key to obtain an encrypted intensity image carrying multiple frames of dynamic information; Step 6: Input the encrypted strength image into the decryption and reconstruction network, combine it with the corresponding orbital angular momentum speckle key information, demultiplex and reconstruct the multi-frame dynamic information, and output the multi-frame dynamic imaging results.

2. The method according to claim 1, characterized in that, In step 2, the topological charge number is an integer or fractional value, and the orbital angular momentum speckle fields corresponding to different topological charge numbers are statistically distinguishable from each other.

3. The method according to claim 2, characterized in that, In step 3, the speckle modulation device includes a scattering medium, a random phase plate, a rough surface, or a combination thereof.

4. The method according to claim 3, characterized in that, In step 4, the preset timing or encoding rules include switching different orbital angular momentum modes in chronological order, or combining two or more orbital angular momentum modes within the same time period. At any moment The speckle key is generated from the following parameter vector: ， in It is the key space and parameter vector related to the orbital angular momentum (OAM) degrees of freedom. Represented as: ， in Represents the moment Topological charge combination of orbital angular momentum (OAM); The x and y coordinates represent the spatial location of the k-th orbital angular momentum OAM mode, respectively. For the first Weighting coefficients at different positions at each time point; This represents the number of combinations of orbital angular momentum (OAM) modes at time t. After propagation through frosted glass scattering, the multiplexed orbital angular momentum (OAM) beam is converted into a speckle illumination field. The speckle pattern generated at time t is used as the physical encryption key, expressed as: ， in This represents the speckle pattern generated at time t. It is a scattering medium; The generation of each orbital angular momentum (OAM) speckle is controlled by the number of OAM modes and their combination; using time-varying speckle fields to illuminate moving objects, the modulated intensity distribution is obtained as follows: ， in Let be the intensity distribution of the moving object at time t after passing through a time-varying speckle field illumination scheme. This represents the background noise during the measurement process. Indicates at time The object function.

5. The method according to claim 4, characterized in that, In step 5, during a single exposure, the dynamic target undergoes continuous modulation by two or more different orbital angular momentum speckle keys within the exposure time, thereby mapping the time series information onto the encryption strength image. By utilizing time-multiplexed speckle illumination in a single exposure, a compressed and encrypted measurement image was acquired, represented as follows: ， in This represents the compressed and encrypted measurement value of a single frame captured by the camera, where T represents the number of frames in the encrypted video; the temporal information of the dynamic scene is compressed onto a two-dimensional measurement map, using a key space. A high-dimensional, time-independent speckle key is obtained, thereby enabling encryption.

6. The method according to claim 5, characterized in that, In step 6, the decryption and reconstruction network is a learning network based on physical model constraints, a data-driven learning network, or a hybrid network structure combining a learning network based on physical model constraints and a data-driven learning network.

7. The method according to claim 6, characterized in that, In step 6, the decryption and reconstruction network includes a U-net backbone, a coarse-to-fine feature fusion module C2FM, a feature enhancement module, an encoder, and a decoder; At the encoder-decoder connection and in the decoder section, a coarse-to-fine feature fusion module (C2FM) is inserted. The coarse-to-fine feature fusion module C2FM includes two parallel branches: one for coarse feature extraction and the other for fine-grained feature extraction. The features from both branches are then fused and output. First, the input to the coarse-to-fine feature fusion module C2FM is divided according to the number of frames in the reconstructed sequence. Group, pair The mean of the group features is calculated and used as the coarse feature extraction branch, represented as follows: The grouped features are then compared with the first group of features in each of the n groups to calculate their positional offset vectors. Feature alignment is then performed using pixel interpolation, as shown below: ， in In reference space Below, the feature tensor at time t after feature alignment; in Represents the t-th feature. This represents the spatial feature bias mapping of the t-th set of features to the reference space. This represents the bilinear interpolation kernel. For the source feature space domain, Represents continuous sampling coordinates in the reference space. Represents discrete source pixels; The aligned features are fused with the original grouped features to generate fine-grained branch features; The coarse feature extraction branch and the fine feature extraction branch are concatenated and then fused with the initial features after feature extraction, which are then combined with the grouped features. The feature enhancement module comprises four branches. The first three branches are the difference between multi-scale convolution kernel features and max pooling features. The multi-scale convolution kernels are, in order, convolution kernels with a size of 3. of The kernel size is 5. of The kernel size is of The fourth branch is the Directional Feature Extraction (DFE) module. The DFE module uses convolutional kernels with different directions to extract features in the horizontal and vertical directions, then concatenates and reduces the dimensionality. Finally, it performs residual linking with the input features. The process is represented as follows: ， in The feature map representing the input direction feature extraction module (DFE). The kernel size is The convolution operator extracts horizontal features. The kernel size is The convolution operator extracts vertical features, and [;] represents the concatenation of channel features. Adjust the channel operator. The output features of the Directional Feature Extraction (DFE) module represent the direction of the feature extraction. The training method for the decryption and reconstruction network includes: training loss Defined as: ， in It is a preset weight. Indicates the reconstructed image. SSIM represents the structural similarity between two images, where the image is labeled with the ground truth (GT).

8. A dynamic encryption imaging device based on orbital angular momentum speckle, implemented using the method described in any one of claims 1 to 7, characterized in that, It includes a coherent light source module, a beam control module, a spatial light modulation module, a speckle modulation module, an imaging acquisition module, and a decryption and reconstruction module; The coherent light source module is used to generate an initial Gaussian beam; The beam control module is used to perform polarization control and beam expansion on the initial Gaussian beam; The spatial light modulation module is used to load helical phase distributions with different topological charges to generate orbital angular momentum vortex beams. The speckle modulation module is used to modulate the orbital angular momentum vortex beam into an orbital angular momentum speckle field that corresponds one-to-one with the topological charge number, and the orbital angular momentum speckle field serves as a dynamic encryption key. The imaging acquisition module is used to acquire encrypted images carrying multiple frames of dynamic information under single exposure conditions. The decryption and reconstruction module is used to demultiplex and reconstruct the encrypted strength image by combining the orbital angular momentum speckle key information.

9. The apparatus according to claim 8, characterized in that, The spatial light modulation module is configured to load spiral phase distributions corresponding to different topological charges or combinations thereof according to a preset time sequence or encoding rules within a single exposure time, so as to achieve time switching of orbital angular momentum modes.

10. The apparatus according to claim 9, characterized in that, The speckle modulation module is used to generate orbital angular momentum speckle fields that correspond to different topological charge numbers and are statistically distinguishable from each other, so as to form a dynamic encryption key. The speckle modulation module is a replaceable or reconfigurable structure to enable dynamic updating of the orbital angular momentum speckle key. The orbital angular momentum speckle key is defined by at least one of the following: the topological charge number value, the topological charge number combination method, and the topological charge number combination time switching order. The decryption and reconstruction module includes a pre-trained decryption and reconstruction network, which uses prior information of the orbital angular momentum speckle key to demultiplex multiple frames of dynamic information in the encrypted strength image. In the absence of corresponding orbital angular momentum speckle key information, the encrypted strength image cannot reconstruct a dynamic imaging result with temporal consistency.

Images