Super-Resolution Single-Image Construction Method and Device

Abstract

The present invention provides a method and device for constructing a super-resolution single image, including: S1, performing detail enhancement on the edge of the image; S2, using a single image self-learning algorithm to supplement and enhance the details of the non-edge part of the image; S3, image fusion enhancement, performing image weighted superposition and filtering denoising on the images output by step S1 and step S2, and obtaining a high-resolution image outputted in the end. The present invention effectively enhances the image display effect and optimizes the user's visual experience.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of image processing technology, and in particular to a method and device for constructing a super-resolution single image. Background Art

[0002] In the human visual perception system, high-resolution images (HR) are an important medium for images to clearly express their spatial structure, detail features, edge texture and other information. They have extremely wide practical value in the fields of medicine, industry, satellite remote sensing, road monitoring, security monitoring, audio and video entertainment, etc. However, many current images are low-resolution images (LR) when initially acquired. How to reconstruct low-resolution images into high-resolution images with clear detail features is a difficult problem in the field of computer vision and image processing. Therefore, the development of a super-resolution single image construction method and device that can effectively overcome the difficulties in the above-mentioned related technologies has become a technical problem that needs to be solved urgently in the industry. Summary of the invention

[0003] In view of the above problems existing in the prior art, an embodiment of the present invention provides a super-resolution single image construction method and device.

[0004] In the first aspect, an embodiment of the present invention provides a super-resolution single image construction method, including: S1, performing detail enhancement on the edge part of the image. Since the edge of the image contains the most important information in an image, the aliasing and blurring phenomena that appear after the super-resolution of the edge part of the image will seriously affect the visual quality of the image after super-resolution, so the details of the edge part of the image are supplemented and enhanced; S2, performing detail enhancement on the non-edge part of the image. Since the non-edge part of the image affects the user's visual experience, based on the self-similarity of the image details, a single image self-learning algorithm is used to supplement and enhance the details of the non-edge part of the image; S3, image fusion enhancement, performing image weighted superposition and filtering denoising on the images output by step S1 and step S2, and enhancing the image details as a whole through image fusion to obtain a high-resolution image that is finally output.

[0005] Based on the content of the above method embodiments, in the super-resolution single-image construction method provided in the embodiments of the present invention, step S1 specifically includes: S101. Obtain a low-resolution image I; S102. Use an interpolation algorithm to magnify the low-resolution image I to obtain an image A; S103. Perform an RGB image judgment on the image A. If it is a grayscale image, select S105 to execute. If it is an RGB color image, then select S104 to execute; S104. Perform an RGB color fusion operation on the image A to obtain an image B; S105. Perform an operator edge detection on the input image to obtain a binary image C; S106. Extract the edge contour of the binary image C to generate an edge pixel position database W; S107. Create a contrast enhancement filter; S108. Perform contrast stretching on the image A to obtain an image D; S109. Perform a multi-dimensional image filtering operation on the image D and the contrast enhancement filter; S110. Through the processing of S109, obtain an edge-enhanced high-resolution image Y.

[0006] Based on the content of the above method embodiments, in the super-resolution single-image construction method provided in the embodiments of the present invention, step S2 specifically includes: S201. Obtain a low-resolution image I; S202. Use an interpolation algorithm to reduce the image I to obtain an image E; S203. Perform a small-block image segmentation on the image I, and the empirical value is a 5×5 small block; S204. For the segmented image I, slide and circularly extract a small-block image L, and obtain a pixel block position T after magnification; S205. Obtain the edge pixel position database W generated in step S1; S206. In the edge pixel position database W, check whether there is an intersection with the position T; S207. If no intersecting element is found, it indicates that the position T is a non-edge part and needs to be detail-enhanced. In the image E, perform a K-nearest neighbor algorithm search on the small-block image L to find a similar small block N smaller than the threshold; S208. After the small block N is magnified, take the corresponding small block M at the corresponding position in the image I to obtain a low-high image pair; S209. Loop S207 and S208 to perform a sliding search in the image E to obtain an LR-HR image pair database U; S210. Obtain the image A generated in step S1; S211. Take out all the M in the LR-HR image pair database U corresponding to the small-block image L and perform pixel weighted fusion to obtain an image small block Q; S212. In the image A, directly copy the image small block Q to the pixel area after the small-block image L is magnified; Loop S204 to S211 to complete the detail enhancement of the non-edge part of the image A; S213. Obtain a non-edge-enhanced high-resolution image Z.

[0007] Based on the content of the above method embodiments, the super-resolution single-image construction method provided in the embodiments of the present invention, step S3 specifically includes: S301. Obtain the edge-enhanced high-resolution image Y generated in step S1; S302. Obtain the non-edge-enhanced high-resolution image Z generated in step S2; S303. Use the pixel weighted fusion method to perform multi-image fusion reconstruction on image Y and image Z; S304. Add median filtering to the image after multi-image fusion reconstruction to remove noise; S305. Generate the high-resolution image O as the final output.

[0008] In a second aspect, an embodiment of the present invention provides a super-resolution single-image construction device, including: a first main module for implementing S1, enhancing details for the edge part of the image; a second main module for implementing S2, using a single-image self-learning algorithm to supplement and enhance details for the non-edge part of the image; a third main module for implementing S3, image fusion enhancement, performing image weighted superposition and filtering denoising on the images output in step S1 and step S2 to obtain the final output high-resolution image.

[0009] In a third aspect, an embodiment of the present invention provides an electronic device, including:

[0010] At least one processor; and

[0011] At least one memory communicatively connected to the processor, wherein:

[0012] The memory stores program instructions executable by the processor, and the processor can execute the super-resolution single-image construction method provided by any one of the various implementation manners in the first aspect by calling the program instructions.

[0013] In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium, and the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions cause the computer to execute the super-resolution single-image construction method provided by any one of the various implementation manners in the first aspect.

[0014] The super-resolution single-image construction method and device provided in the embodiments of the present invention effectively solve the sawtooth phenomenon and blur phenomenon after image super-resolution and supplement key details of the image through edge detail reconstruction optimization and single-image feature self-learning. It can be used on various media devices. For example, adding a single-image super-resolution method to the application program of the set-top box media center, and the set-top box outputs the image processed by the super-resolution algorithm to the TV for direct display, thereby effectively enhancing the image display effect and optimizing the user visual experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0016] Figure 1 Flowchart of the super-resolution single-image construction method provided by the embodiment of the present invention;

[0017] Figure 2 Schematic structural diagram of the super-resolution single-image construction device provided by the embodiment of the present invention;

[0018] Figure 3 Schematic physical structure diagram of the electronic device provided by the embodiment of the present invention;

[0019] Figure 4 Schematic diagram of the specific process of step S1 provided by the embodiment of the present invention;

[0020] Figure 5 Schematic diagram of the specific process of step S2 provided by the embodiment of the present invention;

[0021] Figure 6 Schematic diagram of the specific process of step S3 provided by the embodiment of the present invention. Specific embodiments

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention. In addition, the technical features in each embodiment or a single embodiment provided by the present invention can be combined with each other arbitrarily to form a feasible technical solution. Such combination is not restricted by the order of steps and / or the mode of structural composition, but must be based on the fact that those of ordinary skill in the art can implement it. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0023] The embodiment of the present invention provides a super-resolution single-image construction method. See Figure 1, the method includes: S1. Perform detail enhancement on the edge part of the image. Since the edge of the image contains the most important information in an image, the jagged and blurred phenomena that appear after super-resolution of the edge part of the image will seriously affect the visual quality of the super-resolved image. Therefore, detail supplementation and enhancement are performed on the edge part of the image; S2. Perform detail enhancement on the non-edge part of the image. Since the non-edge part of the image affects the user's visual experience, based on the self-similarity of image details, a single-image self-learning algorithm is used to perform detail supplementation and enhancement on the non-edge part of the image; S3. Image fusion enhancement. For the images output in step S1 and step S2, perform image weighted superposition and filtering denoising, and overall enhance the image details through image fusion to obtain the finally output high-resolution image.

[0024] Based on the content of the above method embodiment, as an optional embodiment, in the super-resolution single-image construction method provided in the embodiment of the present invention, step S1 specifically includes: S101. Obtain a low-resolution image I; S102. Use an interpolation algorithm to magnify the low-resolution image I to obtain an image A; S103. Judge the RGB image of the image A. If it is a grayscale image, select S105 to execute. If it is an RGB color image, then select S104 to execute; S104. Perform an RGB color fusion operation on the image A to obtain an image B; S105. Perform an operator edge detection on the input image to obtain a binary image C; S106. Extract the edge contour of the binary image C to generate an edge pixel position database W; S107. Create a contrast enhancement filter; S108. Perform contrast stretching on the image A to obtain an image D; S109. Perform a multi-dimensional image filtering operation on the image D and the contrast enhancement filter; S110. Through the processing of S109, obtain an edge-enhanced high-resolution image Y. It should be noted that RGB color fusion is an existing technology in the industry, which converts an RGB image into a grayscale image by eliminating hue and saturation information while retaining brightness.

[0025] Specifically, it can be referred to Figure 4, S101: Obtain the input low-resolution image I; S102: Use a traditional interpolation algorithm to magnify the image I by X times to obtain image A; S103: Judge the RGB (Red, Green, Blue, red, green, blue) image of image A. If it is a grayscale image, select S105 to execute. If it is an RGB color image, then select S104 to execute; S104: Perform an RGB color fusion operation on image A to obtain image B; S105: Perform an operator edge detection on the input image to obtain a binary image C; S106: Extract the edge contour of the binary image C to generate an edge pixel position database W; S107: Create a contrast enhancement filter; S108: Perform contrast stretching on image A to obtain image D; S109: Perform an N-dimensional filtering operation on the multi-dimensional image of image D and the contrast enhancement filter; S110: Through the processing of S109, obtain an edge-enhanced high-resolution image Y (magnified by X times).

[0026] Based on the content of the above method embodiments, as an alternative embodiment, the super-resolution single-image construction method provided in the embodiments of the present invention, step S2 specifically includes: S201, obtaining a low-resolution image I; S202, using an interpolation algorithm to downscale the image I to obtain an image E; S203, performing small-block image segmentation on the image I, and the empirical value is small blocks of 7×7 to 9×9 (in another embodiment, it can be small blocks of 5×5); S204, for the segmented image I, sliding and cyclically taking out small-block images L, and obtaining pixel block positions T after magnification; S205, obtaining the edge pixel position database W generated in step S1; S206, in the edge pixel position database W, checking whether there is an intersection with the position T; S207, if no intersecting element is found, it indicates that the position T is a non-edge part and needs to be subjected to detail enhancement. In the image E, perform a K-nearest neighbor algorithm search on the small-block image L to find a similar small block N smaller than the threshold; S208, after magnifying the small block N, take the corresponding small block M at the corresponding position in the image I to obtain a low-high image pair (it should be noted that this small block L, corresponding to L(LR) is a low-resolution image small block; this small block M, corresponding to M(HR) is a high-resolution image small block; the two are a set of low-high resolution image pairs; the image E is the image I downscaled by X times. Find a similar small block N in the image E; obtain the corresponding image small block coordinates, and after magnifying by X times, the corresponding small block M can be found at the corresponding position in the image I, so as to obtain a low-high image pair of L-M); S209, loop S207 and S208 to perform a sliding search in the image E to obtain an LR-HR image pair database U (through the sliding search, a low-high image pair of one L and multiple Ms can be obtained. After traversing the image E in this way, an LR-HR image pair database of one L and multiple Ms can be generated); S210, obtaining the image A generated in step S1; S211, taking out all Ms in the LR-HR image pair database U corresponding to the small-block image L, and performing pixel weighted fusion to obtain an image small block Q; S212, in the image A, directly copy the image small block Q to the pixel area after magnifying the small-block image L; loop S204 to S211 to complete the detail enhancement of the non-edge part of the image A; S213, obtaining a non-edge enhanced high-resolution image Z.

[0027] For details, please refer to Figure 5, S201: Obtain the input low-resolution image I; S202: Use a traditional interpolation algorithm to reduce the image I by X times to obtain the image E; S203: Segment the image I into small image blocks. The empirical value is 5×5 small blocks; S204: For the segmented image I, slide and cyclically extract the small image block L, and magnify it by X times to obtain the pixel block position T; S205: Obtain the edge pixel position database W generated in the S1 process; S206: In the edge pixel position database W, check whether there is an intersection with the position T; S207: If no intersecting element is found, it indicates that the position T is a non-edge part and needs detail enhancement. In the image E, use the K-nearest neighbor algorithm to search for the small block L, and find the similar small block N smaller than the threshold; S208: After magnifying the small block N by X times, take the corresponding small block M at the corresponding position in the image I to obtain the (L, M) low-high image pair; S209: Loop S207 and S208 to perform a sliding search in the image E to obtain the LR-HR image pair database U; S210: Obtain the image A (magnified by X times) generated in the S1 process; S211: Take out all the Ms corresponding to the L in the LR-HR image pair database U, perform pixel weighted fusion to obtain the image small block Q; S212: In the image A, directly copy the image small block Q to the pixel area after magnifying L by X times; Loop S204~S211 to complete the detail enhancement of the non-edge part of the image A; S213: Obtain the non-edge enhanced high-resolution image Z (magnified by X times).

[0028] Based on the content of the above method embodiment, as an optional embodiment, the super-resolution single-image construction method provided in the embodiment of the present invention, step S3 specifically includes: S301. Obtain the edge-enhanced high-resolution image Y generated in step S1; S302. Obtain the non-edge-enhanced high-resolution image Z generated in step S2; S303. Use the pixel weighted fusion method to perform multi-image fusion reconstruction on the image Y and the image Z; S304. Add median filtering to the multi-image fusion reconstructed image to remove noise; S305. Generate the high-resolution image O as the final output.

[0029] Specifically, it can be seen Figure 6 , S301: Obtain the edge-enhanced high-resolution image Y (magnified by X times) generated in the S1 process; S302: Obtain the non-edge-enhanced high-resolution image Z (magnified by X times) generated in the S2 process; S303: Use the pixel weighted fusion method to perform multi-image fusion reconstruction on the image Y and the image Z; S304: Add median filtering to the multi-image fusion reconstructed image to remove noise; S305: Generate the high-resolution image 0 (output, magnified by X times) as the final output of the solution.

[0030] Based on the content of the above method embodiments, as an alternative embodiment, in the super-resolution single-image construction method provided in the embodiments of the present invention, in the loop of S207 and S208, a sliding search is performed on the image E to obtain the LR-HR image pair database U, including: using the K-nearest neighbor algorithm to perform a sliding search in the image reduced by X times, so as to cyclically obtain the LR-HR high and low image pair database.

[0031] Based on the content of the above method embodiments, as an alternative embodiment, in the super-resolution single-image construction method provided in the embodiments of the present invention, steps S211 to S213 specifically include: by taking out all the HR high images corresponding to an LR in the LR-HR high and low image pair database and performing weighted fusion, a high-quality image block is finally obtained, and the low-quality image block is replaced to enhance the non-edge part of the image.

[0032] Based on the content of the above method embodiments, as an alternative embodiment, in the super-resolution single-image construction method provided in the embodiments of the present invention, after obtaining the final output high-resolution image, it further includes: according to the image corresponding scene mode selected by the user, using a preset multi-image fusion method template to enhance the multi-image fusion effect. For example: in portrait mode, Gaussian filtering and grayscale matrix are added to perform filtering processing on the three RGB channels to ensure that the skin color of the image object is soft and natural. Another example: in night scene mode, double histogram equalization processing is added to enhance the low-light image object.

[0033] The super-resolution single-image construction method provided in the embodiments of the present invention effectively solves the sawtooth phenomenon and blur phenomenon after image super-resolution, supplements the key details of the image through edge detail reconstruction optimization and single-image feature self-learning, thereby effectively enhancing the image display effect and optimizing the user visual experience.

[0034] The implementation basis of each embodiment of the present invention is achieved through programmed processing by a device with a processor function. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention can be encapsulated into various modules. Based on this actual situation, on the basis of the above embodiments, the embodiments of the present invention provide a super-resolution single-image construction device, and this device is used to execute the super-resolution single-image construction method in the above method embodiments. See Figure 2 This device includes: a first main module for implementing S1, enhancing the details of the edge part of the image; a second main module for implementing S2, using a single-image self-learning algorithm to supplement and enhance the details of the non-edge part of the image; a third main module for implementing S3, image fusion enhancement, performing image weighted superposition and filtering denoising on the images output by steps S1 and S2 to obtain the finally output high-resolution image.

[0035] The super-resolution single-image construction device provided by the embodiment of the present invention adopts Figure 2 several modules therein to add a single-image super-resolution method to the set-top box media center application. Through edge detail reconstruction optimization and single-image feature self-learning, it effectively solves the sawtooth phenomenon and blurring phenomenon after image super-resolution, supplements key details of the image, and the set-top box outputs the image processed by the super-resolution algorithm to the TV for direct display, thereby effectively enhancing the image display effect and optimizing the user's visual experience.

[0036] It should be noted that the device in the device embodiment provided by the present invention can be used not only to implement the method in the above method embodiment, but also to implement the methods in other method embodiments provided by the present invention. The difference is only in setting corresponding functional modules, and its principle is basically the same as that of the above device embodiment provided by the present invention. As long as those skilled in the art, based on the above device embodiment, refer to the specific technical solutions in other method embodiments, obtain corresponding technical means by combining technical features, and the technical solutions composed of these technical means, and on the premise of ensuring the practicality of the technical solutions, the device in the above device embodiment can be improved to obtain corresponding device-type embodiments for implementing the methods in other method-type embodiments. For example:

[0037] Based on the content of the above device embodiment, as an optional embodiment, the super-resolution single-image construction device provided by the embodiment of the present invention further includes: a first sub-module, and the specific steps for implementing step S1 include: S101, obtaining a low-resolution image I; S102, using an interpolation algorithm to magnify the low-resolution image I to obtain an image A; S103, performing an RGB image judgment on the image A. If it is a grayscale image, select S105 to execute. If it is an RGB color image, then select S104 to execute; S104, performing an RGB color fusion operation on the image A to obtain an image B; S105, performing an operator edge detection on the input image to obtain a binary image C; S106, extracting the edge contour of the binary image C to generate an edge pixel position database W; S107, creating a contrast enhancement filter; S108, performing a contrast stretch on the image A to obtain an image D; S109, performing a multi-dimensional image filtering operation on the image D and the contrast enhancement filter; S110, through the processing of S109, obtaining an edge-enhanced high-resolution image Y.

[0038] Based on the content of the above device embodiments, as an alternative embodiment, the super-resolution single-image construction device provided in the embodiments of the present invention further includes: a second sub-module, and the implementation of step S2 specifically includes: S201. Obtain a low-resolution image I; S202. Use an interpolation algorithm to downscale the image I to obtain an image E; S203. Perform small-block image segmentation on the image I, and the empirical value is a 5×5 small block; S204. For the segmented image I, slide and cyclically extract a small-block image L, and obtain a pixel-block position T after magnification; S205. Obtain the edge pixel position database W generated in step S1; S206. In the edge pixel position database W, check whether there is an intersection with the position T; S207. If no intersecting element is found, it indicates that the position T is a non-edge part and needs detail enhancement. In the image E, perform a K-nearest neighbor algorithm search on the small-block image L to find a similar small block N smaller than the threshold; S208. After the small block N is magnified, take the corresponding small block M at the corresponding position in the image I to obtain a low-high image pair; S209. Cycle S207 and S208 to perform a sliding search in the image E to obtain an LR-HR image pair database U; S210. Obtain the image A generated in step S1; S211. Take out all Ms corresponding to the small-block image L in the LR-HR image pair database U and perform pixel weighted fusion to obtain an image small block Q; S212. In the image A, directly copy the image small block Q to the pixel area after magnification of the small-block image L; cycle S204 to S211 to complete the detail enhancement of the non-edge part of the image A; S213. Obtain a non-edge enhanced high-resolution image Z.

[0039] Based on the content of the above device embodiments, as an alternative embodiment, the super-resolution single-image construction device provided in the embodiments of the present invention further includes: a third sub-module, and the implementation of step S3 specifically includes: S301. Obtain the edge-enhanced high-resolution image Y generated in step S1; S302. Obtain the non-edge enhanced high-resolution image Z generated in step S2; S303. Use a pixel weighted fusion method to perform multi-image fusion reconstruction on the image Y and the image Z; S304. Add median filtering to the multi-image fusion reconstructed image to remove noise; S305. Generate a high-resolution image O as the final output.

[0040] The method of the embodiments of the present invention is implemented relying on an electronic device. Therefore, it is necessary to introduce the relevant electronic device. For this purpose, the embodiments of the present invention provide an electronic device, such as Figure 3As shown in the figure, the electronic device includes: at least one processor, a communications interface, at least one memory, and a communication bus. Among them, the at least one processor, the communications interface, and the at least one memory complete communication with each other through the communication bus. The at least one processor can call the logic instructions in the at least one memory to execute all or part of the steps of the methods provided in the foregoing method embodiments.

[0041] In addition, when the logic instructions in the foregoing at least one memory 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 such an understanding, the technical solution of the present invention, in essence, 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. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the method embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0042] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0043] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the technical solution, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0044] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. With this understanding, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur in a different order than that noted in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, or sometimes in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or by a combination of dedicated hardware and computer instructions.

[0045] It should be noted that the term "comprising", "including", or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a series of elements includes not only those elements but also other elements not expressly listed, or elements that are inherent to such process, method, article, or apparatus. Without further limitation, the elements defined by the statement "comprising..." do not exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the said elements.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. However, such modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A super-resolution single-image construction method, characterized in that, it includes: S1. Perform detail enhancement on the edge part of the image to obtain an edge-enhanced high-resolution image Y; S2. Use a single-image self-learning algorithm to perform detail supplementation and enhancement on the non-edge part of the image; Step S2 specifically includes: S201. Obtain a low-resolution image I; S202. Use an interpolation algorithm to downscale the image I to obtain an image E; S203. Segment the image I into small image blocks; S204. For the segmented image I, slide and cyclically extract a small image block L, and after magnification, obtain the pixel block position T; S205. Obtain the edge pixel position database W generated in step S1; S206. In the edge pixel position database W, check whether there is an intersection with the position T; S207. If no intersecting element is found, it indicates that the position T is a non-edge part and needs detail enhancement. In the image E, perform a K-nearest neighbor algorithm search on the small image block L to find a similar small block N smaller than the threshold; S208. After magnifying the small block N, take the corresponding small block M at the corresponding position in the image I to obtain a low-high image pair; S209. Cycle through S207 and S208 to perform a sliding search in the image E to obtain an LR-HR image pair database U; S210. Obtain the image A generated in step S1; S211. Take out all the Ms corresponding to the small image block L in the LR-HR image pair database U and perform pixel weighted fusion to obtain an image small block Q; S212. In the image A, directly copy the image small block Q to the pixel area after magnification of the small image block L; Cycle through S204 to S211 to complete the detail enhancement of the non-edge part of the image A; S213. Obtain a non-edge-enhanced high-resolution image Z; S3. Image fusion enhancement. Perform image weighted superposition and filter denoising on the images output in steps S1 and S2 to obtain the finally output high-resolution image.

2. The super-resolution single-image construction method according to claim 1, characterized in that, Step S1 specifically includes: S101. Obtain a low-resolution image I; S102. Use an interpolation algorithm to upscale the low-resolution image I to obtain an image A; S103. Perform an RGB image judgment on the image A. If it is a grayscale image, select S105 to execute. If it is an RGB color image, then select S104 to execute; S104. Perform an RGB color fusion operation on the image A to obtain an image B; S105. Perform an operator edge detection on the input image to obtain a binary image C; S106. Extract the edge contour of the binary image C to generate an edge pixel position database W; S107. Create a contrast enhancement filter; S108. Perform contrast stretching on the image A to obtain an image D; S109. Perform a multi-dimensional image filtering operation on the image D and the contrast enhancement filter; S110. Through the processing of S109, obtain an edge-enhanced high-resolution image Y.

3. The super-resolution single-image construction method according to claim 1, characterized in that, Step S3 specifically includes: S301, obtaining the edge-enhanced high-resolution image Y generated in step S1; S302, obtaining the non-edge-enhanced high-resolution image Z generated in step S2; S303, performing multi-image fusion reconstruction on image Y and image Z using the pixel weighted fusion method; S304, adding median filtering to the image after multi-image fusion reconstruction to remove noise; S305, generating the high-resolution image O as the final output.

4. The super-resolution single image construction method according to claim 1, wherein, in the loop S207, S208, a sliding search is performed on the image E to obtain the LR-HR image pair database U, including: using the K-nearest neighbor algorithm, performing a sliding search in the image reduced by X times, so as to cyclically obtain the LR-HR high and low image pair database.

5. The super-resolution single image construction method according to claim 1, wherein, Steps S211 to S213 specifically include: by taking out all the HR high images corresponding to one LR in the LR-HR high and low image pair database and performing weighted fusion, finally obtaining a high-quality image block, and enhancing the non-edge part of the image by replacing the low-quality image block.

6. The super-resolution single image construction method according to claim 1, wherein, after obtaining the final output high-resolution image, it further includes: according to the image corresponding mode scenario selected by the user, using the preset multi-image fusion weight value to enhance the multi-image fusion effect.

7. A super-resolution single image construction device, wherein, it includes: The first main module is used to implement detail enhancement for the edge part of the image to obtain the edge-enhanced high-resolution image Y; The second main module is used to implement the detail supplement and enhancement for the non-edge part of the image using the single-image self-learning algorithm. The specific steps of using the single-image self-learning algorithm to supplement and enhance the details of the non-edge part of the image include: S201, obtaining a low-resolution image I; S202, using an interpolation algorithm to downscale the image I to obtain an image E; S203, performing small-block image segmentation on the image I; S204, for the segmented image I, sliding and cyclically taking out small-block images L, and obtaining pixel block positions T after magnification; S205, obtaining the edge pixel position database W generated in step S1; S206, checking in the edge pixel position database W whether there is an intersection with the position T; S207, if no intersecting element is found, it indicates that the position T is a non-edge part and needs detail enhancement. In the image E, perform a K-nearest neighbor algorithm search for the small-block image L to find a similar small-block N smaller than the threshold; S208, after magnifying the small-block N, take the corresponding small-block M at the corresponding position in the image I to obtain a low-high image pair; S209, cycle S207 and S208 to perform a sliding search in the image E to obtain an LR-HR image pair database U; S210, obtaining the image A generated in step S1; S211, taking out all Ms corresponding to the small-block image L in the LR-HR image pair database U and performing pixel weighted fusion to obtain an image small-block Q; S212, in the image A, directly copy the image small-block Q to the pixel area after magnifying the small-block image L; cycle S204 to S211 to complete the detail enhancement of the non-edge part of the image A; S213, obtaining a non-edge enhanced high-resolution image Z. The third main module is used to implement image fusion enhancement. For the edge-enhanced high-resolution image Y and the non-edge-enhanced high-resolution image Z, perform image weighted superposition and filtering denoising to obtain the finally output high-resolution image.

8. An electronic device characterized in that it includes at least one processor, at least one memory, and a communication interface; wherein the processor, the memory, and the communication interface communicate with each other; the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium characterized in that the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions cause the computer to execute the method according to any one of claims 1 to 6.

Images