High-resolution image reconstruction method and device, equipment and storage medium

Abstract

The invention provides a high-resolution image reconstruction method and device, equipment and a storage medium, and the method comprises the steps: obtaining the mixed data of a low-resolution image, and enabling the mixed data to comprise 2D data and light field data; constructing an image data block of the low-resolution image based on the mixed data, and dividing the image data block into a plurality of minimum image units; respectively calculating definition indexes of the light field data and the 2D data in each minimum image unit, and determining high-resolution data and low-resolution data in each minimum image unit according to the definition indexes of the light field data and the 2D data in each minimum image unit; and performing high-resolution reconstruction on each minimum image unit based on the high-resolution data and the low-resolution data in each minimum image unit to obtain a reconstructed image unit, and constructing a high-resolution image based on each reconstructed image unit. By adopting the method, the reconstruction efficiency and the reconstruction precision of high-resolution image reconstruction can be improved.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a high-resolution image reconstruction method, apparatus, device and storage medium. Background Art

[0002] Light field imaging technology captures the intensity and direction of light, enabling snap-first-focus and deep-depth imaging. However, traditional light field cameras sacrifice spatial resolution for angular information, resulting in a significant loss of detail in the reconstructed image. Furthermore, while conventional 2D imaging offers high spatial resolution, the lack of multi-view information makes it difficult to effectively mitigate image degradation caused by aberrations. This can include blurred images in out-of-focus areas, or ghosting and smearing caused by astigmatism and coma. With the increasing demand for high-resolution imaging, improving spatial resolution while preserving light field angular information has become a key technical challenge in this field.

[0003] Existing light field super-resolution methods (such as ResLF, LFSRDiff, etc.) mostly rely on single light field data and achieve resolution improvement through deep learning networks. However, there are problems such as distorted reconstruction results and high computational complexity (such as the difficulty of training diffusion models), and the high-resolution characteristics of 2D data are not fully utilized. Traditional 2D super-resolution algorithms (such as SRCNN and VDSR) are only designed for single-modal data, lack a mechanism for processing aberrations, and have no effective response in out-of-focus scenarios. In addition, although some multimodal methods attempt to combine light field and 2D data, they mostly rely on modeling of known hardware structures, lack effective solutions to data alignment problems in unknown hardware scenarios, and are difficult to adapt to flexible data acquisition scenarios.

[0004] The core flaws of existing technologies are: first, the data source selection lacks dynamic adaptability, and it is unable to automatically switch between light field and 2D data as high-resolution references based on image quality, resulting in imbalanced reconstruction quality in out-of-focus and in-focus areas; second, hybrid data processing is highly dependent on hardware modeling, and the accuracy of data block matching under unknown structures is low, which limits the generalization ability of the algorithm; third, the aberration suppression and resolution enhancement processes are mutually independent, and a unified optimization framework has not been formed. These flaws make it difficult for existing solutions to achieve robust and efficient reconstruction effects in complex scenarios (such as non-cooperative acquisition of multiple devices and dynamic depth of field changes) with high-resolution imaging requirements. Summary of the Invention

[0005] In view of this, an object of the present invention is to provide a high-resolution image reconstruction method, apparatus, device and storage medium to improve the reconstruction efficiency and reconstruction accuracy of high-resolution image reconstruction.

[0006] In a first aspect, an embodiment of the present application provides a high-resolution image reconstruction method, the method comprising:

[0007] Acquire mixed data of the low-resolution image, wherein the mixed data includes 2D data and light field data;

[0008] constructing an image data block of the low-resolution image based on the mixed data, and dividing the image data block into a number of minimum image units;

[0009] respectively calculating the clarity indexes of the light field data and the 2D data in each minimum image unit, and determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity indexes of the light field data and the 2D data in each minimum image unit;

[0010] Based on the high-resolution data and low-resolution data in each minimum image unit, each minimum image unit is reconstructed at high resolution to obtain a reconstructed image unit, and a high-resolution image is constructed based on each reconstructed image unit.

[0011] Optionally, constructing the image data block of the low-resolution image based on the mixed data includes:

[0012] If the hardware structure of the mixed data is known, constructing the image data block through calibration or coordinate conversion;

[0013] If the hardware structure of the mixed data is unknown, feature matching is performed based on each local area in the 2D data and the light field data to construct the image data block.

[0014] Optionally, determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes:

[0015] For each minimum image unit, a clarity index of the light field data in the minimum image unit is proportionally converted according to a size ratio between the sub-view data extracted from the light field data of the minimum image unit and the 2D data, and the clarity index of the proportionally converted light field data is compared with the clarity index of the 2D data; wherein the size refers to the physical size of the image;

[0016] If the clarity index of the scaled light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit;

[0017] If the clarity index of the scaled light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

[0018] Optionally, before respectively calculating the clarity indexes of the light field data and the 2D data in each minimum image unit, the method further includes:

[0019] Unifying the size of the sub-view data extracted from the light field data and the 2D data in each minimum image unit by up / down sampling;

[0020] Determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes:

[0021] For each minimum image unit, comparing the clarity index of the light field data in the minimum image unit with the clarity index of the 2D data;

[0022] If the clarity index of the light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit;

[0023] If the clarity index of the light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

[0024] Optionally, performing high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit includes:

[0025] For each minimum image unit, constructing a first point spread function of the minimum image unit based on the high-resolution data in the minimum image unit, and constructing a second point spread function of the minimum image unit based on the low-resolution data in the minimum image unit;

[0026] The minimum image unit is reconstructed at high resolution according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit.

[0027] Optionally, performing high-resolution reconstruction on the minimum image unit according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit includes:

[0028] determining a constraint coefficient of the minimum image unit according to a first point spread function and a second point spread function of the minimum image unit;

[0029] updating a first point spread function and a second point spread function of the minimum image unit according to a constraint coefficient of the minimum image unit;

[0030] A reconstructed image unit corresponding to the minimum image unit is generated according to the updated first point spread function and the second point spread function of the minimum image unit.

[0031] Optionally, constructing a high-resolution image based on each reconstructed image unit includes:

[0032] The reconstructed image units are combined according to the division method of each minimum image unit to obtain the high-resolution image.

[0033] In a second aspect, an embodiment of the present application provides a high-resolution image reconstruction device, the device comprising:

[0034] A hybrid data acquisition module, configured to acquire hybrid data of the low-resolution image, wherein the hybrid data includes 2D data and light field data;

[0035] a data block division module, configured to construct an image data block of the low-resolution image based on the mixed data, and divide the image data block into a plurality of minimum image units;

[0036] a clarity analysis module, configured to calculate clarity indices of the light field data and 2D data in each minimum image unit, and determine high-resolution data and low-resolution data in each minimum image unit based on the clarity indices of the light field data and 2D data in each minimum image unit;

[0037] The image reconstruction module is used to perform high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit, and to construct a high-resolution image based on each reconstructed image unit.

[0038] Optionally, constructing the image data block of the low-resolution image based on the mixed data includes:

[0039] If the hardware structure of the mixed data is known, constructing the image data block through calibration or coordinate conversion;

[0040] If the hardware structure of the mixed data is unknown, feature matching is performed based on each local area in the 2D data and the light field data to construct the image data block.

[0041] Optionally, determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes:

[0042] For each minimum image unit, performing a proportional conversion on the clarity index of the light field data in the minimum image unit according to a size ratio between the sub-view data extracted from the light field data in the minimum image unit and the 2D data, and comparing the clarity index of the proportionally converted light field data with the clarity index of the 2D data, wherein the size refers to the physical size of the image;

[0043] If the clarity index of the scaled light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit;

[0044] If the clarity index of the scaled light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

[0045] Optionally, the apparatus includes a size unification module, configured to unify the sizes of the sub-view data extracted from the light field data and the 2D data in each minimum image unit by up / down sampling before respectively calculating the clarity indexes of the light field data and the 2D data in each minimum image unit;

[0046] Determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes:

[0047] For each minimum image unit, comparing the clarity index of the light field data in the minimum image unit with the clarity index of the 2D data;

[0048] If the clarity index of the light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit;

[0049] If the clarity index of the light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

[0050] Optionally, performing high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit includes:

[0051] For each minimum image unit, constructing a first point spread function of the minimum image unit based on the high-resolution data in the minimum image unit, and constructing a second point spread function of the minimum image unit based on the low-resolution data in the minimum image unit;

[0052] The minimum image unit is reconstructed at high resolution according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit.

[0053] Optionally, performing high-resolution reconstruction on the minimum image unit according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit includes:

[0054] determining a constraint coefficient of the minimum image unit according to a first point spread function and a second point spread function of the minimum image unit;

[0055] updating a first point spread function and a second point spread function of the minimum image unit according to a constraint coefficient of the minimum image unit;

[0056] A reconstructed image unit corresponding to the minimum image unit is generated according to the updated first point spread function and the second point spread function of the minimum image unit.

[0057] Optionally, constructing a high-resolution image based on each reconstructed image unit includes:

[0058] The reconstructed image units are combined according to the division method of each minimum image unit to obtain the high-resolution image.

[0059] In a third aspect, an embodiment of the present application provides a computer device comprising: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the high-resolution image reconstruction method described in any optional implementation of the first aspect are performed.

[0060] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, executes the steps of the high-resolution image reconstruction method described in any optional embodiment of the first aspect above.

[0061] The technical solutions provided by this application include but are not limited to the following beneficial effects:

[0062] The present application first obtains mixed data of a low-resolution image, wherein the mixed data includes 2D data and light field data, which can integrate the spatial detail advantages of 2D data with the angular information advantages of light field data, and provide richer and complementary original information for subsequent reconstruction; then, based on the mixed data, an image data block of the low-resolution image is constructed, and the image data block is divided into several minimum image units, which can decompose the global reconstruction task into local processing, reduce computational complexity, facilitate precise optimization of local features, and improve processing efficiency and reconstruction precision; next, the clarity index of the light field data and 2D data in each minimum image unit is calculated respectively, Based on the clarity indicators of the light field data and 2D data in each minimum image unit, the high-resolution data and low-resolution data in each minimum image unit are determined. The data source with greater detail advantages in each local unit can be adaptively selected, so that the reconstruction can fully utilize the strengths of the data and avoid the limitations of a single data modality. Finally, based on the high-resolution data and low-resolution data in each minimum image unit, each minimum image unit is reconstructed at high resolution to obtain a reconstructed image unit. A high-resolution image is constructed based on each reconstructed image unit. This can first accurately restore local details and then splice them into a complete image, taking into account both local accuracy and global consistency, and outputting high-quality and high-resolution results.

[0063] In summary, this application improves the reconstruction efficiency and reconstruction accuracy of high-resolution image reconstruction by fusing multimodal data, disassembling local optimization, dynamically selecting optimal data, and performing local reconstruction and global splicing.

[0064] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0066] Figure 1 A flowchart of a high-resolution image reconstruction method provided by the first embodiment of the present invention is shown;

[0067] Figure 2 FIG2 shows a schematic diagram of a minimum image unit provided by the first embodiment of the present invention;

[0068] Figure 3A flow chart of a method for determining high and low resolution data provided by the first embodiment of the present invention is shown;

[0069] Figure 4 A flow chart of a second method for determining high and low resolution data provided by the first embodiment of the present invention is shown;

[0070] Figure 5 A flowchart of a high-resolution image generation method provided by the first embodiment of the present invention is shown;

[0071] Figure 6 A flowchart of a method for generating a reconstructed image unit provided in the first embodiment of the present invention is shown;

[0072] Figure 7 A schematic structural diagram of a high-resolution image reconstruction device provided by a second embodiment of the present invention is shown;

[0073] Figure 8 A schematic structural diagram of a computer device provided in the third embodiment of the present invention is shown. DETAILED DESCRIPTION

[0074] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0075] Example 1

[0076] To facilitate understanding of this application, Figure 1 The flowchart of a high-resolution image reconstruction method provided by the first embodiment of the present invention is shown to describe the content of the first embodiment of the present application in detail.

[0077] See also Figure 1 As shown, Figure 1 The flowchart of a high-resolution image reconstruction method provided by the first embodiment of the present invention is shown, wherein the method includes steps S101 to S104:

[0078] S101: Acquire mixed data of a low-resolution image, wherein the mixed data includes 2D data and light field data.

[0079] Specifically, there are two ways to acquire mixed data: one is direct acquisition through a sparse light field structure, that is, the microlens units are non-densely arranged at periodic intervals on the sensor, and sparse light field data and 2D data are collected at the same time; the other is joint acquisition through a traditional light field camera and an ordinary wide-field camera, with the light field camera acquiring light field data and the wide-field camera acquiring 2D data of the same scene. Among them, light field data contains angular information from multiple perspectives, while 2D data provides higher spatial resolution. The combination of the two forms a complementary low-resolution mixed data source. Light field data is multi-perspective data, and the acquisition method can be hardware acquisition or algorithm generation. The acquisition methods include but are not limited to light field cameras, camera arrays composed of multiple cameras in space, or single cameras moving to shoot multi-perspective images. 2D data refers to planar image data acquired by ordinary imaging methods, which only contains light intensity information and does not contain angle information.

[0080] S102: Constructing an image data block of the low-resolution image based on the mixed data, and dividing the image data block into a number of minimum image units.

[0081] Specifically, based on the mixed data, the full-size low-resolution image is converted into an image data block, and then different minimum units are preset according to different data source types, and the image data block is divided into several minimum image units: based on the data acquisition method, if there is a microlens array structure, the period of the microlens arrangement is set as the minimum area; if there is no microlens array structure, based on the data characteristics, if there is a spatial periodic overlapping area between two data representing the same area in the mixed data, the area within the overlapping period is set as the minimum area.

[0082] See also Figure 2 As shown, Figure 2 A schematic diagram of a minimum image unit provided in Example 1 of the present invention is shown, wherein the image data block corresponding to the full-size image is divided into M×N minimum image unit patches. Each minimum image unit patch contains light field data and 2D data, forming a set of high- and low-resolution data pairs. The figure shows a local 3×3 minimum image unit, which provides structured input for subsequent image reconstruction.

[0083] S103: respectively calculating clarity indexes of the light field data and the 2D data in each minimum image unit, and determining high-resolution data and low-resolution data in each minimum image unit according to the clarity indexes of the light field data and the 2D data in each minimum image unit.

[0084] Specifically, the information entropy, discreteness, average edge strength MSE, discreteness D(I LLightweight clarity metrics such as the image quality indicator (x,y) or modulation transfer function (MTF) are used to determine the impact of aberrations on image quality. If the aberrations are large (e.g., outside the lens depth of field), the light field data, after reconstruction and restoration of multi-view information, can suppress the blur caused by aberrations, and its clarity is superior. Therefore, it is set as high-resolution data, and the 2D data is used as low-resolution data. If the aberrations are small (e.g., within the depth of field), the 2D data, due to its clear focus and superior clarity, is set as high-resolution data, and the light field data is used as low-resolution data. This dynamic selection mechanism enables adaptive adaptation to scenes with different aberrations.

[0085] Calculate the information entropy and discrete degree D(I L Lightweight clarity metrics include information entropy (MSE), mean edge strength (MSE), and modulation transfer function (MTF). Information entropy reflects the degree of image blur by calculating the uncertainty of pixel grayscale distribution. Lower values ​​indicate more uniform pixel distribution (higher blur). MSE assesses detail richness by accumulating edge gradient magnitudes. Higher values ​​indicate sharper edges.

[0086] Degree of dispersion D(I L (x, y)) is calculated as follows: convolve the spatial information image I(x, y) in any window with the Laplace filter mask to obtain the texture information I of the input value L (x,y), that is, the information of rapid grayscale changes. Statistics I L The discrete degree of high-frequency texture D(I L (x,y)), that is: within the statistical window area, if the original input texture is clear and the grayscale changes rapidly, the more high-frequency components there are, the more discrete the high frequency is compared to the low frequency. The specific formula is as follows: D(I L (x,y))=mean[I L (x,y)-mean(I L (x,y))] 2 .

[0087] The modulation transfer function (MTF) is calculated by roughly estimating the point spread function (PSF) of each data block from the same data source, and calculating the MTF value from the PSF to obtain a quantifiable image clarity index. Specifically, the method for estimating the PSF is to input the blur kernel using blind deconvolution, such as by using RL iterative optimization. During the iterative image processing, the blur kernel used in the last round of iteration is output as the PSF, or an existing lightweight network for end-to-end PSF estimation is designed or used. The method for obtaining the MTF quantitative clarity is to perform a Fourier transform on the PSF to obtain the OTF (Optical Transfer Function), and then modulo the OTF to obtain the MTF value. The resolution can be determined from the size of the MTF value.

[0088] S104: performing high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit, and constructing a high-resolution image based on each reconstructed image unit.

[0089] Specifically, before the high-resolution reference group is used to guide the low-resolution to-be-restored group to perform super-resolution reconstruction to obtain a high-resolution image, the method further includes:

[0090] The data format of each minimum image unit is normalized, including color space conversion, channel dimension adjustment, etc.; the image edges are mirror-filled, zero-filled, or the boundary is extended; the data of each minimum image unit is exposed compensated (linear or nonlinear transformation of the brightness value of the image pixel to enhance the details of the dark area in the image or compress the highlight area information) using algorithms including but not limited to Retinex; the data of each minimum image unit is smoothed or denoised using methods including spatial filtering or frequency domain filtering of the image pixel matrix, or data standardization by subtracting the mean and removing zero points.

[0091] When performing high-resolution reconstruction of each minimum image unit based on the high-resolution and low-resolution data within each minimum image unit to obtain a reconstructed image unit, the following reconstruction process is performed for each minimum image unit: first, a first point spread function corresponding to the high-resolution data is constructed in the frequency domain based on the high-resolution data, and a second point spread function corresponding to the low-resolution data is constructed based on the low-resolution data. The point spread function consists of a set of orthogonal bases and orthogonal basis coefficients. The constraint coefficients are solved using a greedy algorithm, a convex relaxation optimization algorithm, or a statistical optimization method, so that the high-resolution and low-resolution data satisfy the objective function minimization. Then, the first and second point spread functions are iteratively updated to optimize the point spread functions' ability to represent data features. If the first and second point spread functions differ in structure, an additional mapping matrix must be solved. Finally, a high-resolution reconstructed block is generated based on the mapping matrix. After all the reconstructed blocks are spliced ​​together, a complete high-resolution image is obtained. The entire process is divided into three modules: data construction, preprocessing, and data reconstruction, achieving complementary reconstruction of mixed data.

[0092] In an optional embodiment, dividing the image data block into a number of minimum image units includes:

[0093] If the hardware structure of the mixed data is known, the image data block is divided into a number of minimum image units through calibration or coordinate conversion.

[0094] Specifically, if the hardware structure of the mixed data is known, and the hardware structures of the devices from different data sources are different, different methods can be used to divide the minimum image units. For the case where a 2D camera and a light field camera are known, the intrinsic parameters (such as focal length, principal point), extrinsic parameters (such as rotation matrix, translation vector) and distortion parameters of the two can be obtained through calibration. These parameters are used to transform the coordinates of the same target, and the multi-view image of the light field data is aligned with the corresponding area of ​​the 2D data through coordinate mapping to ensure that the local areas of the two are strictly corresponding in physical space. The aligned image is then divided into the minimum image units containing the same target area. For the case where the sparse light field arrangement structure is known, the position of the microlens and the local area corresponding to each microlens (such as the central area as the 2D data acquisition point and the surrounding area as the light field angle sampling point) can be determined according to the periodic spacing and non-dense arrangement specifications of its microlens units. The area range of the 2D data and light field data corresponding to each microlens is calculated according to the specifications, and it is divided into the minimum image units containing the same target area or adjacent target areas.

[0095] If the hardware structure of the mixed data is unknown, feature matching is performed on the 2D data and the light field data in the image data block to obtain a number of minimum image units.

[0096] Specifically, if the hardware structure of the mixed data is unknown, first extract local feature points (such as SIFT and ORB features) from the 2D data and light field data respectively and calculate the feature descriptors. Then, use a feature matching algorithm (such as FLANN matching or brute force matching) to find local area matching pairs between the two. During the process, it is necessary to ensure the global consistency of the matching through epipolar constraints, homography transformations, etc., and use indicators such as SSIM and ZNCC to screen matching pairs with high local similarity to determine a unique matching relationship. Then, according to the matching results, the matching areas of the 2D data and the light field data are divided into corresponding minimum image units, so that the light field data and 2D data in each data block represent the same target area or adjacent target areas.

[0097] In an alternative embodiment, see Figure 3 As shown, Figure 3 A flowchart of a method for determining high-resolution and low-resolution data provided by a first embodiment of the present invention is shown, wherein the method determines high-resolution data and low-resolution data in each minimum image unit based on the clarity index of the light field data and 2D data in each minimum image unit, including steps S301 to S305:

[0098] S301: For each minimum image unit, proportionally convert the clarity index of the light field data in the minimum image unit according to the size ratio of the sub-view data extracted from the light field data of the minimum image unit to the 2D data, and compare the clarity index of the proportionally converted light field data with the clarity index of the 2D data, wherein the size refers to the physical size of the image.

[0099] Specifically, for each minimum image unit, the sub-view data in the light field data is first extracted, and the size ratio between it and the 2D data is calculated (e.g., if the image height of a certain object point in the sub-view data of the light field data is 1 / 4 of the image height of the object point in the 2D data, then the ratio is 0.25). Since the resolution of the light field data and the 2D data may differ, directly comparing the clarity index will lead to errors. Therefore, the clarity index of the light field data needs to be linearly converted according to this ratio (for example, the index value of the light field data is multiplied by the scale factor). After the conversion, the clarity index of the light field data is numerically compared with the index of the 2D data to eliminate the impact of the resolution difference on the index judgment.

[0100] S302: If the clarity index of the scaled light field data in the minimum image unit exceeds the clarity index of the 2D data, the light field data in the minimum image unit is determined as high-resolution data in the minimum image unit, and the 2D data in the minimum image unit is determined as low-resolution data in the minimum image unit.

[0101] Specifically, the scaled light field data has higher clarity than the 2D data, indicating that the light field data contains richer detail or better aberration suppression within the current minimum image unit. In this case, the light field data is considered high-resolution data and serves as a reference for subsequent reconstruction. The 2D data, however, is considered low-resolution data due to its lower clarity, requiring guidance from the light field data along with the high-resolution data for super-resolution reconstruction.

[0102] S303: If the clarity index of the scaled light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

[0103] Specifically, if the clarity index of the light field data after proportional conversion is not higher than that of the 2D data, it means that the 2D data has better focus clarity in the current area. In this case, the 2D data is determined to be high-resolution data and used to guide reconstruction. The light field data is determined to be low-resolution data due to its poor resolution or detail, and its resolution needs to be improved through feature transfer from the 2D data.

[0104] In an optional embodiment, before separately calculating the clarity indexes of the light field data and the 2D data in each minimum image unit, the method further includes: unifying the sizes of the sub-view data extracted from the light field data and the 2D data in each minimum image unit by up / down sampling.

[0105] Specifically, if the spatial resolution of the sub-perspective data of the light field data is lower than that of the 2D data, upsampling is used to enlarge it to the size of the 2D data; or the 2D data is downsampled to reduce it to the size of the light field sub-perspective data; if the spatial resolution of the sub-perspective data of the light field data is higher than that of the 2D data, downsampling is used to reduce it to the size of the 2D data; or the 2D data is upsampled to enlarge it to the size of the light field sub-perspective data; ensure that the two are strictly aligned in the image space. This operation ensures that the clarity of the two data is compared at the same image height, avoids indicator deviation due to resolution difference, and provides a unified benchmark for subsequent judgment.

[0106] See also Figure 4 As shown, Figure 4 A flowchart of a second method for determining high-resolution and low-resolution data provided by the first embodiment of the present invention is shown, wherein the method determines high-resolution data and low-resolution data in each minimum image unit based on the clarity index of the light field data and 2D data in each minimum image unit, including steps S401 to S403:

[0107] S401: For each minimum image unit, compare the clarity index of the light field data in the minimum image unit with the clarity index of the 2D data.

[0108] Specifically, the same type of clarity index is calculated for the light field data and 2D data after the size is unified, and the numerical comparison is directly performed.

[0109] S402: If the clarity index of the light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit.

[0110] Specifically, if the clarity index of the light field data is higher than that of the 2D data, it indicates that the light field data contains richer details or better aberration suppression within the current minimum image unit. In this case, the light field data is determined to be high-resolution data and serves as a reference for subsequent reconstruction. The 2D data, however, is determined to be low-resolution data due to its lower clarity index and requires the guidance of the light field data for super-resolution reconstruction along with the high-resolution data.

[0111] S403: If the clarity index of the light field data in the minimum image unit does not exceed the clarity index of the 2D data, determining the 2D data in the minimum image unit as high-resolution data in the minimum image unit, and determining the light field data in the minimum image unit as low-resolution data in the minimum image unit.

[0112] Specifically, if the clarity index of the light field data is not higher than that of the 2D data, it means that the 2D data has better focus clarity in the current area. In this case, the 2D data is determined to be high-resolution data and used to guide reconstruction. The light field data is determined to be low-resolution data due to its poor resolution or detail, and its resolution needs to be improved by transferring features from the 2D data.

[0113] In an alternative embodiment, see Figure 5 As shown, Figure 5 A flowchart of a high-resolution image generation method provided in the first embodiment of the present invention is shown, wherein the method of performing high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit includes steps S501 to S503:

[0114] S501: For each minimum image unit, construct a first point spread function of the minimum image unit based on high-resolution data in the minimum image unit, and construct a second point spread function of the minimum image unit based on low-resolution data in the minimum image unit.

[0115] Specifically, construct a point spread function psf (including the first point spread function psf H and the second point spread function psf L ), first extract high-resolution data y H As the first point spread function psf H training samples; at the same time, extract low-resolution data y L As the second point spread function psf L The specific implementation includes: initializing the PSF by randomly sampling the minimum image unit, which is composed of a set of orthogonal bases and initialization coefficients in the frequency domain, and iteratively optimizing the coefficients of each orthogonal basis in the PSF using an optimization algorithm so that the PSF can represent the data features of the corresponding resolution. For example, performing wavelet transform on high-resolution data, or performing singular value decomposition (SVD) or feature extraction on the image to generate a PSF covering its texture and edge patterns. H , ensure psf H Capable of accurately characterizing high-resolution details.

[0116] S502: Perform high-resolution reconstruction on the minimum image unit according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit.

[0117] Specifically, solve the shared constraint coefficient α for each minimum image unit:

[0118] Make low resolution data L and high-resolution data H Meet the optimization goal λ is the regularization parameter.

[0119] By iteratively solving the constraint coefficients through the optimization algorithm, it is ensured that α contains only a small number of non-zero elements, achieving sparse representation of the data and reducing the complexity of the operation. If there is a difference in the structure of the first point spread function and the second point spread function (for example, the orthogonal basis selected for the initialization of the point spread function is not completely consistent), it is necessary to additionally solve the mapping matrix W (by minimizing ), and finally generate the reconstructed image unit K = psf H ·α·W, maps low-resolution data to high-resolution space.

[0120] In an alternative embodiment, see Figure 6 As shown, Figure 6 A flowchart of a method for generating a reconstructed image unit provided by the first embodiment of the present invention is shown, wherein the method of performing high-resolution reconstruction on the minimum image unit according to the first point spread function and the second point spread function of the minimum image unit to obtain the reconstructed image unit includes steps S601 to S603:

[0121] S601: Determine a constraint coefficient of the minimum image unit according to a first point spread function and a second point spread function of the minimum image unit.

[0122] Specifically, for each minimum image unit low-resolution data y L and high resolution reference y H , in the constructed second point spread function psf L and the first point spread function psf H Next, solve the objective function The constraint coefficient α is minimized. λ is a regularization parameter used to control the sparsity constraint. By iteratively selecting the point spread function most relevant to the residual, the constraint coefficient vector is gradually constructed to ensure that α contains only a small number of non-zero elements, achieving a sparse representation of the shared features of high- and low-resolution data while reducing computational complexity.

[0123] S602: Update the first point spread function and the second point spread function of the minimum image unit according to the constraint coefficient of the minimum image unit.

[0124] Specifically, the constraint coefficient α is fixed, and the reconstruction error matrix E = y-psf·α (y is taken as y as needed) is calculated in groups. L or y H , corresponding to psf L or psf H Then perform singular value decomposition (SVD) on the error matrix E, take the left singular vector corresponding to the maximum singular value, and update the psf L or psf H ; Adjust α to maintain sparsity and repeat the iteration until psf L 、psf H Convergence means optimizing the orthogonal basis coefficients that constitute the point spread function so that the point spread function can more accurately cover the texture, edge and other characteristic patterns of the smallest image unit, thereby improving the ability to represent high and low resolution data.

[0125] S603: Generate a reconstructed image unit corresponding to the minimum image unit according to the updated first point spread function and the second point spread function of the minimum image unit.

[0126] Specifically, the reconstructed image unit is generated based on the updated point spread function and the constraint coefficient α. If the first point spread function and the second point spread function have the same structure (for example, the orthogonal basis selected by the initialized first point spread function and the second point spread function are completely consistent), directly through K = psf H · α calculates the reconstructed image unit K; if there is a difference between the first point spread function and the second point spread function structure (such as the orthogonal basis selected for the initialized point spread function is not completely consistent), it is necessary to additionally solve the mapping matrix W (by minimizing To achieve feature space alignment), the final reconstructed image unit is K = psf H ·α·W. This step converts low-resolution data blocks into high-resolution details through point spread function mapping and structural alignment, completes single-block reconstruction, and provides a basis for subsequent global image stitching.

[0127] In an optional embodiment, constructing a high-resolution image based on each reconstructed image unit includes:

[0128] The reconstructed image units are combined according to the division method of each minimum image unit to obtain the high-resolution image.

[0129] Specifically, the reconstruction results of all minimum image units are stitched together according to the original segmentation coordinates to form a complete high-resolution image. During stitching, weighted fusion of overlapping regions (e.g., Gaussian weights) is used to eliminate seams between blocks. The weights are dynamically adjusted based on the distance from the pixel to the block center to ensure smooth edge transitions. If further optimization is required, the stitched image can be post-processed, such as using non-local mean filtering for denoising, unsharp masking for edge enhancement, or leveraging light field angle information for global consistency constraints. The final output is a high-resolution image with rich detail and continuous structure.

[0130] Example 2

[0131] See also Figure 7 As shown, Figure 7 FIG. 1 is a schematic structural diagram of a high-resolution image reconstruction device provided in a second embodiment of the present invention, wherein the device comprises:

[0132] A mixed data acquisition module 701 is used to acquire mixed data of a low-resolution image, wherein the mixed data includes 2D data and light field data;

[0133] A data block division module 702 is configured to construct an image data block of the low-resolution image based on the mixed data, and divide the image data block into a plurality of minimum image units;

[0134] a clarity analysis module 703 for respectively calculating clarity indices of the light field data and 2D data in each minimum image unit, and determining high-resolution data and low-resolution data in each minimum image unit based on the clarity indices of the light field data and 2D data in each minimum image unit;

[0135] The image reconstruction module 704 is configured to perform high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit, and construct a high-resolution image based on each reconstructed image unit.

[0136] In an optional embodiment, a hybrid data acquisition module is configured to acquire hybrid data of the low-resolution image, wherein the hybrid data includes 2D data and light field data;

[0137] a data block division module, configured to construct an image data block of the low-resolution image based on the mixed data, and divide the image data block into a plurality of minimum image units;

[0138] a clarity analysis module, configured to calculate clarity indices of the light field data and 2D data in each minimum image unit, and determine high-resolution data and low-resolution data in each minimum image unit based on the clarity indices of the light field data and 2D data in each minimum image unit;

[0139] The image reconstruction module is used to perform high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit, and to construct a high-resolution image based on each reconstructed image unit.

[0140] In an optional embodiment, determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes:

[0141] For each minimum image unit, performing a proportional conversion on the clarity index of the light field data in the minimum image unit according to a size ratio between the sub-view data extracted from the light field data in the minimum image unit and the 2D data, and comparing the clarity index of the proportionally converted light field data with the clarity index of the 2D data, wherein the size refers to the physical size of the image;

[0142] If the clarity index of the scaled light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit;

[0143] If the clarity index of the scaled light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

[0144] In an optional embodiment, the apparatus includes a size unification module, the size unification module being configured to unify the sizes of the sub-view data extracted from the light field data and the 2D data in each minimum image unit by up / down sampling before respectively calculating the clarity index of the light field data and the 2D data in each minimum image unit;

[0145] Determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes:

[0146] For each minimum image unit, comparing the clarity index of the light field data in the minimum image unit with the clarity index of the 2D data;

[0147] If the clarity index of the light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit;

[0148] If the clarity index of the light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

[0149] In an optional embodiment, performing high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit includes:

[0150] For each minimum image unit, constructing a first point spread function of the minimum image unit based on the high-resolution data in the minimum image unit, and constructing a second point spread function of the minimum image unit based on the low-resolution data in the minimum image unit;

[0151] The minimum image unit is reconstructed at high resolution according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit.

[0152] In an optional embodiment, performing high-resolution reconstruction on the minimum image unit according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit includes:

[0153] determining a constraint coefficient of the minimum image unit according to a first point spread function and a second point spread function of the minimum image unit;

[0154] updating a first point spread function and a second point spread function of the minimum image unit according to a constraint coefficient of the minimum image unit;

[0155] A reconstructed image unit corresponding to the minimum image unit is generated according to the updated first point spread function and the second point spread function of the minimum image unit.

[0156] In an optional embodiment, constructing a high-resolution image based on each reconstructed image unit includes:

[0157] The reconstructed image units are combined according to the division method of each minimum image unit to obtain the high-resolution image.

[0158] Example 3

[0159] Based on the same application concept, see Figure 8 As shown, Figure 8 FIG. 1 shows a schematic diagram of the structure of a computer device provided by the third embodiment of the present invention, wherein Figure 8 As shown, a computer device 800 provided in the third embodiment of the present application includes:

[0160] A processor 801, a memory 802, and a bus 803. The memory 802 stores machine-readable instructions executable by the processor 801. When the computer device 800 is running, the processor 801 communicates with the memory 802 via the bus 803. When the processor 801 is running, the machine-readable instructions execute the steps of the high-resolution image reconstruction method shown in the above-mentioned embodiment 1.

[0161] Example 4

[0162] Based on the same application concept, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the high-resolution image reconstruction method described in any one of the above embodiments are executed.

[0163] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems and devices can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0164] The computer program product for high-resolution image reconstruction provided in an embodiment of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the method described in the previous method embodiment. The specific implementation can be found in the method embodiment and will not be repeated here.

[0165] The high-resolution image reconstruction device provided in the embodiments of the present invention can be specific hardware on a device or software or firmware installed on the device. The implementation principles and technical effects of the device provided in the embodiments of the present invention are the same as those of the aforementioned method embodiments. For the sake of brief description, any matters not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiments. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can all refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0166] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, the indirect coupling or communication connection of the device or unit may be electrical, mechanical or other forms.

[0167] The units described as separate components may or may not be physically separate, and 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 these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0168] In addition, each functional unit in the embodiment provided by the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0169] If the functions are implemented in the form of 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 the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0170] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0171] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. However, such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. They should all be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A high-resolution image reconstruction method, characterized in that: The method comprises: Acquire mixed data of the low-resolution image, wherein the mixed data includes 2D data and light field data; constructing an image data block of the low-resolution image based on the mixed data, and dividing the image data block into a number of minimum image units; respectively calculating the clarity indexes of the light field data and the 2D data in each minimum image unit, and determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity indexes of the light field data and the 2D data in each minimum image unit; Based on the high-resolution data and low-resolution data in each minimum image unit, each minimum image unit is reconstructed at high resolution to obtain a reconstructed image unit, and a high-resolution image is constructed based on each reconstructed image unit.

2. The method according to claim 1, characterized in that The step of dividing the image data block into a plurality of minimum image units comprises: If the hardware structure of the mixed data is known, the image data block is divided into a number of minimum image units by calibration or coordinate conversion; If the hardware structure of the mixed data is unknown, feature matching is performed on the 2D data and the light field data in the image data block to obtain a number of minimum image units.

3. The method according to claim 1, characterized in that Determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes: For each minimum image unit, performing a proportional conversion on the clarity index of the light field data in the minimum image unit according to a size ratio between the sub-view data extracted from the light field data in the minimum image unit and the 2D data, and comparing the clarity index of the proportionally converted light field data with the clarity index of the 2D data, wherein the size refers to the physical size of the image; If the clarity index of the scaled light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit; If the clarity index of the scaled light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

4. The method according to claim 1, wherein Before respectively calculating the clarity indexes of the light field data and the 2D data in each minimum image unit, the method further includes: Unifying the size of the sub-view data extracted from the light field data and the 2D data in each minimum image unit by up / down sampling; Determining the high-resolution data and the low-resolution data in each minimum image unit according to the clarity index of the light field data and the 2D data in each minimum image unit includes: For each minimum image unit, comparing the clarity index of the light field data in the minimum image unit with the clarity index of the 2D data; If the clarity index of the light field data in the minimum image unit exceeds the clarity index of the 2D data, determining the light field data in the minimum image unit as high-resolution data in the minimum image unit, and determining the 2D data in the minimum image unit as low-resolution data in the minimum image unit; If the clarity index of the light field data in the minimum image unit does not exceed the clarity index of the 2D data, the 2D data in the minimum image unit is determined as the high-resolution data in the minimum image unit, and the light field data in the minimum image unit is determined as the low-resolution data in the minimum image unit.

5. The method according to claim 1, wherein The step of performing high-resolution reconstruction on each minimum image unit based on the high-resolution data and the low-resolution data in each minimum image unit to obtain a reconstructed image unit includes: For each minimum image unit, constructing a first point spread function of the minimum image unit based on the high-resolution data in the minimum image unit, and constructing a second point spread function of the minimum image unit based on the low-resolution data in the minimum image unit; The minimum image unit is reconstructed at high resolution according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit.

6. The method according to claim 5, characterized in that The step of performing high-resolution reconstruction on the minimum image unit according to the first point spread function and the second point spread function of the minimum image unit to obtain a reconstructed image unit includes: determining a constraint coefficient of the minimum image unit according to a first point spread function and a second point spread function of the minimum image unit; updating a first point spread function and a second point spread function of the minimum image unit according to a constraint coefficient of the minimum image unit; A reconstructed image unit corresponding to the minimum image unit is generated according to the updated first point spread function and the second point spread function of the minimum image unit.

7. The method according to claim 1, characterized in that The constructing of a high-resolution image based on each reconstructed image unit includes: The reconstructed image units are combined according to the division method of each minimum image unit to obtain the high-resolution image.

8. A high-resolution image reconstruction device, characterized in that: The device comprises: A hybrid data acquisition module, configured to acquire hybrid data of the low-resolution image, wherein the hybrid data includes 2D data and light field data; a data block division module, configured to construct an image data block of the low-resolution image based on the mixed data, and divide the image data block into a plurality of minimum image units; a clarity analysis module, configured to calculate clarity indices of the light field data and 2D data in each minimum image unit, and determine high-resolution data and low-resolution data in each minimum image unit based on the clarity indices of the light field data and 2D data in each minimum image unit; The image reconstruction module is used to perform high-resolution reconstruction on each minimum image unit based on the high-resolution data and low-resolution data in each minimum image unit to obtain a reconstructed image unit, and to construct a high-resolution image based on each reconstructed image unit.

9. A computer device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the high-resolution image reconstruction method according to any one of claims 1 to 7 are performed.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the high-resolution image reconstruction method according to any one of claims 1 to 7.

Images