A nanosecond laser gating imaging slice recovery method and system

Abstract

The application provides a nanosecond laser gating imaging slice recovery method and system, and relates to the technical field of laser imaging slice recovery. The method comprises the following steps: building a nanosecond laser gating imaging platform, collecting nanosecond laser gating imaging slices; constructing a high-low quality slice dataset based on the collected slices; constructing a slice recovery model based on GatedFormer, including a wavelet hybrid attention module, a multi-scale gated feedforward neural network, a global high-frequency information enhancement module and an information injection module; in the training stage, using the high-low quality slice dataset, taking the low-quality slice as the input and the high-quality slice as the target, and training the model through a space-frequency joint loss function; in the inference stage, inputting the slice to be recovered into the trained model, and outputting the recovered slice. The method is highly adaptable to the data form of the nanosecond laser gating imaging slice, and simultaneously considers the recovery quality and the calculation efficiency, thereby meeting the real-time application requirements.

Description

Technical Field

[0001] This invention relates to the field of laser imaging slice restoration technology, and in particular to a nanosecond-level laser-gated imaging slice restoration method and system. Background Technology

[0002] Laser-gated imaging is an active light imaging technology that precisely synchronizes the camera's gating window with the laser pulse emission time, allowing only the laser signal reflected from the target object to pass through. This effectively suppresses interference from ambient light (such as sunlight and artificial light sources) and scattering media (such as smoke, rain, and snow), significantly improving image contrast and clarity. It can be widely used in scenarios such as high-precision ranging, depth sensing, autonomous driving, security monitoring, and industrial inspection.

[0003] With the increasing demand for imaging in high-speed, high-scattering environments, nanosecond-level laser-gated imaging technology has emerged. By controlling the shutter speed on the nanosecond scale, it significantly improves scattering suppression and imaging temporal resolution, thereby achieving better distance measurement accuracy. In nanosecond-level laser-gated imaging, an "image slice" refers to a single frame of image acquired within an extremely short time window (e.g., 500 ps to 3 ns). Due to the extremely short time window, a single frame typically has a low signal-to-noise ratio. Traditional methods such as image intensifiers can amplify the signal, but they easily introduce strong noise, affecting image quality.

[0004] Existing technologies still have the following drawbacks in nanosecond-level laser-gated imaging slice processing:

[0005] 1) Limitations of nanosecond-level laser-gated imaging systems: To achieve high-precision measurement, the system needs to use a short shutter time of nanoseconds, which leads to insufficient photon collection. It is necessary to rely on an image intensifier to increase the gain, but increasing the gain will inevitably amplify electronic noise, causing the slices to become blurry and lose details, which seriously affects the imaging quality.

[0006] 2) Inefficiency of multi-slice accumulation method: Traditional multi-slice accumulation generates a high-quality image by averaging 16 consecutive noisy slices, but this process requires the acquisition of a large number of slices, which is time-consuming and consumes too much computing resources, and cannot adapt to real-time scenarios (such as dynamic environment perception in autonomous driving).

[0007] 3) Inapplicability of existing image restoration methods: Existing image restoration methods (such as those based on convolutional neural networks or Transformers) are optimized for visible light images and do not consider the special characteristics of nanosecond-level laser-gated slices: lack of color information, sparse high-frequency components, and strong coupling between noise and edges. Texture and edge details are crucial to image quality, but existing methods struggle to specifically enhance these features and effectively separate and restore high-frequency details, resulting in poor restoration performance. Summary of the Invention

[0008] To address the aforementioned problems, the present invention aims to provide a nanosecond-level laser-gated imaging slice restoration method and system. On one hand, a slice restoration dataset is constructed, consisting of image pairs with one-to-one correspondences between single-frame slices and accumulated multi-frame slices. Accumulated slices are selected as the real targets for model training, while only single-frame slices are used as model input. This allows for the restoration of multi-frame slices from a single-frame slice during the inference stage, overcoming the problems of poor quality of single-frame slices and difficulty in acquiring accumulated slices. On the other hand, based on the GatedFormer network architecture, a wavelet hybrid attention module is used to separate and enhance high- and low-frequency information. A global high-frequency information enhancement module and an information injection module improve the global consistency of high-frequency details. Spatial and frequency domain optimization is balanced through joint spatial-frequency loss. This method is highly compatible with the data format of nanosecond-level laser-gated imaging slices, balancing restoration quality and computational efficiency to meet real-time application requirements.

[0009] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0010] On the one hand, a nanosecond-level laser-gated imaging slice recovery method is provided, the method comprising the following steps:

[0011] S1. Build a nanosecond-level laser-gated imaging platform and collect nanosecond-level laser-gated imaging slices;

[0012] S2. Construct high- and low-quality slice datasets based on the collected slices;

[0013] S3. Construct a slice restoration model based on GatedFormer, including: a wavelet hybrid attention module to enhance high-frequency and low-frequency features of slices to achieve texture restoration; a multi-scale gated feedforward neural network for multi-scale feature extraction and information flow control; a global high-frequency information enhancement module to further enhance high-frequency features; and an information injection module for multi-scale feature fusion.

[0014] S4. During the training phase, the high- and low-quality slice dataset is used, with low-quality slices as input and high-quality slices as targets, and the slice recovery model is trained using the space-frequency joint loss function.

[0015] S5. In the inference phase, the slice to be restored is input into the trained slice restoration model, and the restored slice is output.

[0016] Optionally, the nanosecond-level laser-gated imaging platform described in step S1 includes: a nanosecond-level laser-gated camera, a multi-channel synchronous timing controller, a high-power laser emitter, and supporting optical components.

[0017] Optionally, the high- and low-quality slice dataset in step S2 includes data pairs consisting of low-quality slices and high-quality slices, wherein the low-quality slices are single-frame slice data and the high-quality slices are 16 frames of cumulative average slice data corresponding to the low-quality slice scene, and the two achieve pixel-level correspondence; the high- and low-quality slice dataset is divided into training set, validation set and test set according to 7:1.5:1.5.

[0018] Optionally, the wavelet hybrid attention module in step S3 specifically includes:

[0019] Wavelet transform is used to decompose the slice into low-frequency features and high-frequency features. The high-frequency features are extracted to extract horizontal / vertical texture through 1×3 and 3×1 depth-separable convolutions, while the low-frequency features are modeled to model global dependencies through decoupled linear attention. Then, the enhanced high-frequency and low-frequency features are fused by inverse wavelet transform to achieve the collaborative restoration of high-frequency details and low-frequency structures.

[0020] Optionally, the multi-scale gated feedforward neural network in step S3 specifically includes:

[0021] A dual-branch structure is adopted, which extracts multi-scale local features through 1×3+3×1 and 1×5+5×1 depthwise separable convolutions respectively. After GELU activation, the information flow is dynamically adjusted through element-wise multiplication gating mechanism to achieve more efficient feature representation and information fusion.

[0022] Optionally, the global high-frequency information enhancement module in step S3 specifically includes:

[0023] Fourier transform is performed on local features at different scales of the encoder to extract high-frequency features, which are then enhanced by convolutional blocks, nonlinear transformations and sigmoid activation. Finally, inverse Fourier transform is performed to obtain enhanced global high-frequency features, thus compensating for the problem of insufficient cross-layer information interaction.

[0024] Optionally, the information injection module in step S3 specifically includes:

[0025] The global high-frequency features output by the global high-frequency information enhancement module are adjusted to a size that matches the features of the decoder, and then injected into the decoder through element-wise multiplication to enhance the global consistency of high-frequency details.

[0026] Optionally, the slice recovery model based on GatedFormer in step S3 further includes a gated block, which combines the output of the wavelet hybrid attention module and the multi-scale gated feedforward neural network through feature splitting to improve information utilization.

[0027] Optionally, the space-frequency joint loss function described in step S4 combines spatial domain constraints and frequency domain constraints.

[0028] On the other hand, a nanosecond-level laser-gated imaging slice restoration system is provided for implementing the method described in any of the above embodiments, the system comprising:

[0029] The data acquisition module is used to build a nanosecond-level laser-gated imaging platform and acquire nanosecond-level laser-gated imaging slices.

[0030] The dataset building module is used to construct high- and low-quality slice datasets based on the collected slices;

[0031] The model building module is used to build a slice restoration model based on GatedFormer, including: a wavelet hybrid attention module to enhance high-frequency and low-frequency features of slices to achieve texture restoration; a multi-scale gated feedforward neural network for multi-scale feature extraction and information flow control; a global high-frequency information enhancement module to further enhance high-frequency features; and an information injection module for multi-scale feature fusion.

[0032] The model training module is used to train the slice recovery model using the high- and low-quality slice dataset, with low-quality slices as input and high-quality slices as targets, through the joint space-frequency loss function.

[0033] The inference module is used to input the slice to be restored into the trained slice restoration model and output the restored slice.

[0034] On the other hand, an electronic device is provided, the electronic device comprising:

[0035] processor;

[0036] The memory stores computer-readable instructions, which, when loaded and executed by the processor, implement the steps of the nanosecond-level laser-gated imaging slice restoration method described above.

[0037] On the other hand, a computer-readable storage medium is provided, wherein program code is stored in the computer-readable storage medium, and the program code can be called by a processor to execute the steps of the nanosecond-level laser-gated imaging slice restoration method described above.

[0038] The beneficial effects of the technical solution provided by this invention include at least the following:

[0039] (1) This invention constructs a high- and low-quality slice dataset specifically for nanosecond laser-gated slice recovery. Based on the combination of hardware platform, dataset and algorithm model, it provides a complete solution for the practical application of nanosecond laser-gated imaging technology.

[0040] (2) The present invention can achieve high-quality slice recovery with only a single slice input, without the need for multiple slices to be accumulated input, and the inference time is short, which meets the application requirements of high efficiency and real-time.

[0041] (3) Based on the data characteristics of laser-gated imaging slices, this invention constructs a slice restoration model based on GatedFormer to solve the problems of low high-frequency components and strong noise coupling, so that it can be applied to imaging scenarios in complex environments. Through the synergistic effect of multiple modules, high and low frequency details are effectively separated and enhanced, and the restored slices have higher quality. The peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) are significantly better than existing methods. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a nanosecond-level laser-gated imaging slice recovery method provided in an embodiment of the present invention;

[0044] Figure 2 (a) and (b) are schematic diagrams of the synchronous control principle of the nanosecond-level laser gating imaging technology provided in the embodiments of the present invention and comparative diagrams with existing cumulative imaging technologies, respectively.

[0045] Figure 3 This is a schematic diagram of the architecture of the GatedFormer slice recovery model provided in an embodiment of the present invention;

[0046] Figure 4 This is a schematic diagram of a nanosecond-level laser-gated imaging slice restoration system provided in an embodiment of the present invention;

[0047] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0049] In embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the term "exemplary" is intended to present concepts in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either one or the other. In embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably; it should be noted that their intended meanings are consistent unless the distinction is emphasized.

[0050] This invention provides a nanosecond-level laser-gated imaging slice restoration method, which solves the problems of existing technologies through a three-pronged technical solution of "hardware platform + dataset + algorithm model". (Reference) Figure 1 As shown, the processing flow of the method includes the following steps:

[0051] S1. Build a nanosecond-level laser-gated imaging platform and acquire nanosecond-level laser-gated imaging slices.

[0052] As an optional embodiment of the present invention, the nanosecond-level laser-gated imaging platform includes: a nanosecond-level laser-gated camera (shortest shutter width 500 picoseconds, frame rate 98 frames per second), a multi-channel synchronous timing controller (accuracy 10 picoseconds), a high-power laser emitter (wavelength 860nm, peak power 4000W), and supporting optical components (18-400mm zoom lens, beam expander, etc.), for acquiring nanosecond-level laser-gated imaging slices.

[0053] S2. Construct high- and low-quality slice datasets based on the collected slices.

[0054] In this embodiment of the invention, based on the aforementioned nanosecond-level laser-gated imaging platform, 4275 slices containing scenes such as buildings, offices, and squares were acquired. These slices were used to construct a high- and low-quality slice dataset. The high- and low-quality slice dataset includes data pairs consisting of low-quality slices and high-quality slices. The low-quality slices are single-frame slice data (e.g., single-frame slices after preprocessing from the original acquisition), and the high-quality slices are 16-frame cumulative average slice data corresponding to the low-quality slice scene (i.e., slices obtained by cumulatively averaging 16 frames acquired in the scene corresponding to the low-quality slice), achieving pixel-level correspondence between the two. Subsequently, the high- and low-quality slice dataset was divided into a training set, a validation set, and a test set in a 7:1.5:1.5 ratio to provide data support for model training.

[0055] Figure 2Figures (a) and (b) are schematic diagrams illustrating the synchronous control principle of the nanosecond-level laser-gated imaging technology provided in this embodiment of the invention, and a comparison diagram with existing cumulative imaging technologies, respectively. In this invention, a single low-quality slice is used as input, and a high-quality slice is used as the target to train the model. During the inference phase, there is no need for multi-slice cumulative input; only a single slice needs to be passed through the model to achieve high-quality reconstruction, resulting in higher efficiency and convenience.

[0056] S3. Construct a slice recovery model based on GatedFormer. (Reference) Figure 3 As shown, the GatedFormer-based slice restoration model includes: a Wavelet-Mixed Attention (WMA) module for enhancing high-frequency and low-frequency features of slices to achieve texture restoration; a Multi-Scale Gated Feed-Forward Neural Network (MSGFFN) for multi-scale feature extraction and information flow control; a Global High-Frequency Information Enhancement Module (GHFIEM) for further enhancing high-frequency features; and an Information Injection Module (IM) for multi-scale feature fusion.

[0057] Specifically, the wavelet hybrid attention module includes:

[0058] Let the input be Low-frequency characteristic components were obtained by applying Haar wavelet transform. and high-frequency characteristic components The high-frequency characteristic component H is divided into three sub-bands. Formed by splicing along the channel dimension The low-frequency characteristic component L is used as .

[0059] After decomposing the slice into low-frequency (LF) and high-frequency (HF) features using wavelet transform, the HF features are used to extract horizontal / vertical texture through 1×3 and 3×1 depthwise separable convolutions. Then, GELU activation is applied to generate enhanced HF features. In the low-frequency feature branch, the query, key, and value matrix are... The sequence length is N, and the feature dimension is C. H and W represent the height and width, respectively, i.e., the number of pixels in the feature map in the vertical and horizontal directions. Standard Softmax attention has a complexity of O(N²) for calculating pairwise similarity, which is costly for high-resolution restoration. To address this issue, this invention proposes Decoupled Linear Attention (DLA), which normalizes the sequence using Softmax, reducing the complexity to O(N). To compensate for the potential low-rank problem of DLA, a depthwise separable convolution is applied to V. (Projection matrix...) The input LF is mapped to Q, K, and V to obtain the enhanced low-frequency feature component (LFR). To reconstruct the enhanced LFR and high-frequency feature component (HFR), the HFR is first segmented into directional components along the channel. Reshape and piece together Inverse wavelet transform generates fused features. The high-frequency enhancement branch is achieved through a 1×1 convolution. Dimensional reduction Applying depthwise separable convolutions with 1×5 and 5×1 kernels, followed by sigmoid activation, yields... After bilinear interpolation, and Align and multiply element by element to produce the final output. This enables the coordinated recovery of high-frequency details and low-frequency structures.

[0060] The multi-scale gated feedforward neural network specifically includes:

[0061] A dual-branch structure is adopted, which extracts multi-scale local features through 1×3+3×1 and 1×5+5×1 depthwise separable convolutions respectively. After GELU activation, the information flow is dynamically adjusted through element-wise multiplication gating mechanism to achieve more efficient feature representation and information fusion.

[0062] The global high-frequency information enhancement module specifically includes:

[0063] Given encoder features , and The fusion features, along with their corresponding terms C1, C2, and C3, are extracted using Fourier transform by GHFIEM. After aligning the sizes through global average pooling and bilinear interpolation, the high-frequency features are concatenated and dimensionality-reduced using 3×3 convolution to obtain the fused features. Subsequently, after convolutional blocks, nonlinear transformations, and sigmoid activation enhancement, enhanced global high-frequency features are obtained through inverse Fourier transform. This addresses the issue of insufficient cross-layer information exchange.

[0064] The information injection module specifically includes:

[0065] Adjust by average pooling or bilinear interpolation To match and fuse features The size. After sigmoid activation. and Element-wise multiplication yields the feature representation Then, the output is obtained by processing through convolutional blocks to enhance the feature representation. This adjusts the global high-frequency features output by the global high-frequency information enhancement module to a size that matches the decoder features, and then injects them into the decoder through element-wise multiplication to enhance the global consistency of high-frequency details.

[0066] Furthermore, the slice recovery model based on GatedFormer also includes a gated block, which improves information utilization by combining the output of the wavelet hybrid attention module and the multi-scale gated feedforward neural network through feature decomposition.

[0067] S4. During the training phase, the high- and low-quality slice datasets are used, with low-quality slices as input and high-quality slices as targets, and the slice recovery model is trained using the space-frequency joint loss function.

[0068] In this embodiment of the invention, a dedicated high- and low-quality slice dataset is constructed to train the model. While using the accumulated high-quality slices as the real target for model training, only a single low-quality slice is used as the model input. This achieves the effect of obtaining the recovery of multiple slices from a single slice during the inference stage, while overcoming the problems of poor quality of single-frame slices and difficulty in collecting accumulated slices.

[0069] L1 / L2 distance and SSIM loss are widely used in image restoration, but they mainly operate in the spatial domain and struggle to effectively separate high-frequency and low-frequency information, thus hindering the simultaneous optimization of global structure and local details. Furthermore, they are sensitive to noise and outliers. Nanosecond-level laser-gated slicing, due to its longer wavelength, is less sensitive to atmospheric scattering, has fewer high-frequency components, and is more susceptible to noise and outliers.

[0070] To address these issues, a joint spatial-frequency loss function is proposed. This function separates high-frequency and low-frequency features using both spatial and frequency domains, while preserving global structure and local details, thus improving the model's robustness and generalization ability. Given M training pairs... (i=1,...,M), extract the high-frequency feature components of the slices using Fourier transform. Represents edges and details, as well as low-frequency feature components. This represents the smooth region and structure. Here, F(Y) is the Fourier transform of slice Y, and |F(Y)| is its amplitude spectrum.

[0071]

[0072]

[0073]

[0074]

[0075]

[0076] in, Optimize high-frequency components to reduce detail differences. Constrain the spectrum to preserve the overall frequency distribution and minimize artifacts. Targeting low-frequency components to reduce differences in smoothing regions and structures. Minimize pixel-level differences, ε=10 -3 To ensure stability.

[0077] This invention combines spatial domain constraints and frequency domain constraints to balance the optimization of high and low frequency components, thereby improving the model's ability to recover details and structure.

[0078] S5. In the inference phase, the slice to be restored is input into the trained slice restoration model, and the restored slice is output.

[0079] The specific workflow of the method is as follows:

[0080] Training phase: Using the constructed high and low quality slice dataset, with low quality slices as input and 16 frames of accumulated real values ​​of high quality slices as the objective, a slice recovery model based on GatedFormer is trained through the joint space-frequency loss function;

[0081] Inference phase: Input the single low-quality slice to be restored into the trained model, and output the restored slice to achieve efficient single-slice restoration.

[0082] Compared with existing technologies, this invention constructs a high- and low-quality slice dataset specifically for nanosecond-level laser-gated slice restoration. Based on the combination of hardware platform, dataset, and algorithm model, it provides a complete solution for the practical application of nanosecond-level laser-gated imaging technology. In the embodiments of this invention, high-quality slice restoration can be achieved with only a single slice input model, and the inference time is short (the basic GatedFormer model inference time is only about 0.411 seconds), which can meet the application requirements of high efficiency and real-time performance.

[0083] Meanwhile, based on the data characteristics of laser-gated imaging slices, this invention constructs a slice restoration model based on GatedFormer to address the issues of low high-frequency components and strong noise coupling. Through the collaborative action of multiple modules, it effectively separates and enhances high- and low-frequency details, resulting in higher quality restored slices with significantly better peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) than existing methods. Table 1 compares the performance of the method of this invention with existing image enhancement methods on a nanosecond-level laser-gated imaging slice restoration dataset.

[0084] Table 1

[0085]

[0086] Accordingly, embodiments of the present invention also provide a nanosecond-level laser-gated imaging slice restoration system. Figure 4 This is a structural block diagram illustrating a nanosecond-level laser-gated imaging slice restoration system according to an exemplary embodiment. Figure 4 As shown, the system includes:

[0087] The data acquisition module is used to build a nanosecond-level laser-gated imaging platform and acquire nanosecond-level laser-gated imaging slices.

[0088] The dataset building module is used to construct high- and low-quality slice datasets based on the collected slices;

[0089] The model building module is used to build a slice restoration model based on GatedFormer, including: a wavelet hybrid attention module to enhance high-frequency and low-frequency features of slices to achieve texture restoration; a multi-scale gated feedforward neural network for multi-scale feature extraction and information flow control; a global high-frequency information enhancement module to further enhance high-frequency features; and an information injection module for multi-scale feature fusion.

[0090] The model training module is used to train the slice recovery model using the high- and low-quality slice dataset, with low-quality slices as input and high-quality slices as the target, through the joint space-frequency loss function.

[0091] The inference module is used to input the slice to be restored into the trained slice restoration model and output the restored slice.

[0092] For ease of explanation, Figure 4 Only the main components of the system are shown. The system of this embodiment can be used to perform... Figure 1 The technical solutions of the method embodiments shown are similar in principle and in effect, and will not be described again here.

[0093] In an exemplary embodiment, the present invention also provides an electronic device, the electronic device comprising:

[0094] processor;

[0095] The memory stores computer-readable instructions, which, when loaded and executed by the processor, implement the steps of the nanosecond-level laser-gated imaging slice restoration method described above.

[0096] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 5 As shown, the electronic device 300 may include a first processor 3001 and a memory 3002. Optionally, the electronic device 300 may also include a transceiver 3003. The first processor 3001, memory 3002, and transceiver 3003 may be connected via a communication bus. The memory 3002 stores computer-readable instructions, which, when executed by the first processor 3001, implement the steps of the nanosecond-level laser-gated imaging slice restoration method described above.

[0097] In a specific implementation, as one example, the first processor 3001 may include one or more CPUs, for example... Figure 5 CPU0 and CPU1 are shown in the diagram.

[0098] In a specific implementation, as one example, the electronic device 300 may also include multiple processors, for example... Figure 5 The first processor 3001 and the second processor 3004 are shown in the diagram. Each of these processors can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0099] The memory 3002 is used to store the software program that executes the present invention, and is controlled by the first processor 3001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0100] The transceiver 3003 is used to communicate with network devices or with terminal devices.

[0101] Optionally, the transceiver 3003 may include a receiver and a transmitter. The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0102] Optionally, the transceiver 3003 can be integrated with the first processor 3001 or exist independently and be coupled to the first processor 3001 through the interface circuit of the electronic device 300. This embodiment of the invention does not specifically limit this.

[0103] It should be noted that, Figure 5 The structure of the electronic device 300 shown is not intended to limit the electronic device. Actual electronic devices may include more or fewer components than shown, or combine certain components, or have different component arrangements. Furthermore, the technical effects of the electronic device 300 can be understood by referring to the technical effects of the above-described method embodiments, and will not be repeated here.

[0104] In an exemplary embodiment, the present invention also provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the steps of the nanosecond-level laser-gated imaging slice restoration method described above. For example, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc.

[0105] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0106] The use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0107] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0108] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0109] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0110] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0111] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0112] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0113] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0114] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0115] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A nanosecond laser-gated imaging slice recovery method, characterized in that, The method comprises the following steps: S1, a nanosecond laser gated imaging platform is built, and nanosecond laser gated imaging slices are collected; S2, a high-low quality slice dataset is constructed based on the collected slices; S3, a slice restoration model based on GatedFormer is constructed, comprising: a wavelet mixed attention module for enhancing slice high-frequency features and low-frequency features to realize texture restoration; a multi-scale gated feedforward neural network for multi-scale feature extraction and information flow control; a global high-frequency information enhancement module for further enhancing high-frequency features; and an information injection module for multi-scale feature fusion; The wavelet mixed attention module specifically comprises: Wavelet transform is used to decompose the slice into low-frequency features and high-frequency features, the high-frequency features are extracted through 1x3 and 3x1 depth separable convolution to obtain horizontal / vertical texture, and the low-frequency features are modeled through decoupled linear attention to obtain global dependence; then, the enhanced high-frequency features and low-frequency features are fused through inverse wavelet transform to realize collaborative restoration of high-frequency details and low-frequency structures; The multi-scale gated feedforward neural network specifically comprises: A double-branch structure is adopted, and 1x3+3x1 and 1x5+5x1 depth separable convolution are used to extract multi-scale local features, respectively; after GELU activation, the information flow is dynamically adjusted through the gating mechanism of element multiplication to realize more efficient feature representation and information fusion; The global high-frequency information enhancement module specifically comprises: Fourier transform is performed on the local features of different scales of the encoder, high-frequency features are extracted, and the high-frequency features are enhanced through a convolution block, a nonlinear transformation and a sigmoid activation, and then inverse Fourier transform is performed to obtain enhanced global high-frequency features, thereby making up for the problem of insufficient cross-layer information interaction; The information injection module specifically comprises: The global high-frequency features output by the global high-frequency information enhancement module are adjusted to a size matching the features of the decoder, and the global high-frequency features are injected into the decoder through element multiplication to enhance the global consistency of high-frequency details; S4, in the training stage, the high-low quality slice dataset is used, the low-quality slice is taken as the input, and the high-quality slice is taken as the target; the slice restoration model is trained through a space-frequency joint loss function; S5, in the inference stage, the slice to be restored is input into the trained slice restoration model, and the restored slice is output.

2. The nanosecond laser-gated imaging slice recovery method of claim 1, wherein, The nanosecond laser gated imaging platform in step S1 comprises a nanosecond laser gated camera, a multi-channel synchronous timing controller, a high-power laser emitter and a matching optical assembly.

3. The nanosecond laser-gated imaging slice recovery method of claim 1, wherein, The high-low quality slice dataset in step S2 comprises a data pair composed of a low-quality slice and a high-quality slice, wherein the low-quality slice is single-frame slice data, the high-quality slice is 16-frame cumulative average slice data corresponding to the scene of the low-quality slice, and the two realize pixel-level correspondence; the high-low quality slice dataset is divided into a training set, a verification set and a test set according to a ratio of 7:1.5:1.

5.

4. The nanosecond laser-gated imaging slice recovery method of claim 1, wherein, The slice restoration model based on GatedFormer in step S3 further comprises a gating block, which combines the outputs of the wavelet mixed attention module and the multi-scale gated feedforward neural network through feature splitting to improve information utilization.

5. The nanosecond laser-gated imaging slice recovery method of claim 1, wherein, The space-frequency joint loss function in step S4 combines a spatial domain constraint and a frequency domain constraint.

6. A nanosecond laser-gated imaging slice recovery system for implementing the method of any one of claims 1 to 5, characterized in that, The system comprises: A data acquisition module is configured to build a nanosecond laser gating imaging platform and acquire nanosecond laser gating imaging slices. A data set construction module is configured to construct a high-low quality slice data set based on the acquired slices. A model construction module is configured to construct a slice restoration model based on GatedFormer, including: a wavelet hybrid attention module configured to enhance high-frequency features and low-frequency features of the slices to achieve texture restoration; a multi-scale gated feedforward neural network configured to perform multi-scale feature extraction and information flow control; a global high-frequency information enhancement module configured to further enhance high-frequency features; and an information injection module configured to perform multi-scale feature fusion. A model training module is configured to train the slice restoration model by using the high-low quality slice data set, taking low-quality slices as input and high-quality slices as target, and through a space-frequency joint loss function. An inference module is configured to input a slice to be restored into the trained slice restoration model and output a restored slice.

Images