A burr edge simulation quantitative method, device, electronic equipment and storage medium

Abstract

The application discloses a burr edge simulation quantitative method and device, electronic equipment and a storage medium. The method comprises the following steps: traversing all pixels in an original image, and performing edge detection on the original image; then, based on strong edge pixel points, a region where a burr is located is simulated; then, according to the original image, a burr component is constructed; finally, a burr edge simulation image is obtained by fusing the original image and the burr component. The application accurately simulates the morphological characteristics of the burr edge, and is helpful for accurately identifying and removing burr noise.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and in particular to a burr edge simulation quantitative method, device, electronic equipment and storage medium. BACKGROUND

[0002] In actual applications, image / video often appears burr phenomenon in the edge area after low code rate compression transmission and related processing procedures, resulting in a decline in visual experience. Conventional burr edge removal algorithms have many shortcomings due to the limitations of algorithm principles and design logic. For example, median filtering, Gaussian filtering and other algorithms perform global unified processing on the entire image, lack of pertinence, and do not consider the "local isolation" feature of burr edge. For another example, mean filtering and other algorithms reduce high-frequency components to reduce noise, but texture details (hair, fabric lines, etc.) in the image are also high-frequency information and are filtered out at the same time, which easily causes damage to texture details. For another example, some traditional algorithms (such as non-local mean filtering, NLM) retain the structure by global pixel matching, but the calculation amount increases exponentially with the image resolution, making it difficult to apply to real-time video processing. The essential defect of conventional burr edge removal algorithms is "extensive processing based on bottom-layer pixel statistics", which lacks accurate identification of burr shape and understanding of image semantics. This makes it difficult for them to meet the needs of "efficient noise reduction" and "detail fidelity" at the same time when facing diversified burr edges. In recent years, deep learning algorithms can theoretically make up for the shortcomings of conventional algorithms by learning the shape features of burrs and the context relationship on the image. However, similarly, it still needs to train the precise burr shape as the input training set.

[0003] Therefore, it is crucial to accurately simulate the shape features of burr edges. This helps our designed algorithm (especially deep learning network) accurately identify the pattern features of burrs, so as to conduct targeted suppression and avoid damaging other texture details to the greatest extent, ensuring that the original enhancement or sharpening effect is retained. SUMMARY

[0004] Since the existing method has the above problems, the present application embodiment proposes a burr edge simulation quantitative method, device, electronic equipment and storage medium.

[0005] Specifically, the present application embodiment provides the following technical solutions:

[0006] In a first aspect, the present application embodiment provides a burr edge simulation quantitative method, comprising:

[0007] Edge detection is performed on all pixels in the original image, and then a region where burrs are located is simulated based on strong edge pixel points; then a burr component is constructed according to the original image; finally, a burr edge simulation image is obtained by fusing the original image and the burr component.

[0008] In a second aspect, an embodiment of the present application provides a burr edge simulation and quantification device, comprising:

[0009] An edge detection module is configured to perform edge detection on all pixels in the original image.

[0010] A burr region determination module is configured to simulate a region where burrs are located based on strong edge pixel points.

[0011] A burr component construction module is configured to construct a burr component according to the original image.

[0012] An image fusion module is configured to fuse the original image and the burr component to obtain a burr edge simulation image.

[0013] In a third aspect, an embodiment of the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the burr edge simulation and quantification method according to the first aspect when executing the computer program.

[0014] In a fourth aspect, an embodiment of the present application further provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the burr edge simulation and quantification method according to the first aspect.

[0015] According to the above technical solution, the technical solution provided by the present application can accurately simulate the morphological characteristics of burr edges, help subsequent design to accurately identify the pattern characteristics of burrs by denoising algorithm (especially deep learning network), and thus perform targeted inhibition, thereby avoiding damage to other texture details to the greatest extent and ensuring that the original enhancement or sharpening effect is retained. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0017] Figure 1 A flowchart of the burr edge simulation and quantification method provided by an embodiment of the present application.

[0018] Figure 2A flow chart of the edge detection method provided by an embodiment of the present application.

[0019] Figure 3 An example diagram of a rectangular region of burr edge.

[0020] Figure 4 A structural schematic diagram of the burr edge simulation and quantification device provided by an embodiment of the present application.

[0021] Figure 5 A structural schematic diagram of the electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0022] The specific embodiments of the present application are further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, and cannot be used to limit the protection scope of the present application.

[0023] Figure 1 A flow chart of the burr edge simulation and quantification method provided by an embodiment of the present application is shown in FIG. 1. Figure 1 As shown in FIG. 1, the burr edge simulation and quantification method provided by the embodiment of the present application specifically includes the following contents:

[0024] Step 1, traverse all pixels in the original image, and perform edge detection on the original image.

[0025] In the image processing task, the edge detection is generally based on identifying the positions where the brightness or color in the image changes significantly to detect the edge pixels of the image. The commonly used methods include: 1. Gradient-based method, such as Roberts operator, Prewitt operator, Sobel operator, etc. Although this kind of method is simple and efficient, it has poor noise resistance and is sensitive to noise, and is easy to produce false edges; 2. Second-order derivative-based method, such as Laplacian operator, LoG operator, etc. This kind of method is relatively complex in calculation, accurate in positioning, but slow in processing speed and sensitive to noise; 3. Canny operator. This method has strong noise resistance and good connectivity, but weak edges are lost.

[0026] It should be noted that the method of edge detection is not limited by the embodiments of the present application, and the corresponding edge detection method can be selected according to the actual situation.

[0027] In one embodiment, as shown in FIG. 2, the traversing all pixels in the original image and performing edge detection on the original image includes: Figure 2 Divide a square first pixel region centered on the current pixel, and the size of the first pixel region is K*K.

[0028]

[0029] ​Wherein, K=2r+1;r is a natural number, representing the radius of the first pixel region.

[0030] Shrinking or enlarging the first pixel region to a second pixel region of a preset size.

[0031] The size of the second pixel region is not less than 5*5.

[0032] According to the pixels in the second pixel region, the second derivative of the current pixel point is calculated.

[0033] It should be noted that the method for calculating the second derivative is not limited in the embodiment of the application, and the derivative method can be selected according to the actual situation.

[0034] Different sizes of the first pixel region are selected, and the above steps are repeated to obtain a multiscale second derivative.

[0035] The function relationship f(r) between the multiscale second derivative and the radius of the first pixel region is fitted.

[0036] According to the function relationship f(r), one or more first pixel region radius sizes r_argmax corresponding to the maximum value of the second derivative of the current pixel are counted.

[0037] A threshold T is preset for the first pixel region radius, if the first pixel region radius r_argmax is less than the threshold T, the current pixel is a strong edge pixel, if the first pixel region radius r_argmax is greater than the threshold T, the current pixel is a weak edge pixel.

[0038] Step 2, based on the strong edge pixel point, the area where the burr is located is simulated.

[0039] We can imagine the burr area as a "ink point model", and the strong edge pixel is equivalent to the center position of the ink point, and the burr area is equivalent to the trace of the ink spreading from the center position to the surrounding. The way of ink spreading is random but follows certain rules, and the burr is the same.

[0040] The simulation of the area where the burr is located based on the strong edge pixel point comprises:

[0041] The range of the burr generated is not more than a square region of R*R centered on the current strong edge pixel, and one or more rectangular regions containing the current strong edge pixel are randomly selected in the square region.

[0042] It should be noted that the size of the square region is not limited in the embodiment of the application, and can be preset according to the actual situation.

[0043] For example, the square region and the rectangular region are as shown in Figure 3 .

[0044] Superimpose the rectangular regions to obtain a collection region, which is the region where the burr is located.

[0045] Traverse all strong edge pixels to obtain the corresponding burr region.

[0046] Step 3, constructing a burr component according to the original image.

[0047] The burr component is constructed according to the original image, comprising:

[0048] Marking the pixels in the burr region as 1 and marking other pixels as 0 to construct a burr region statistical component C of the current image.

[0049] Constructing a rectangular window for each pixel.

[0050] The size of the rectangular window is not more than a square region of R' * R' centered on the current pixel, and the rectangular window contains the current pixel.

[0051] It should be noted that the size of the square region is not limited in the embodiment of the application, and can be preset according to actual conditions.

[0052] Filtering the original image with the rectangular window to obtain a second image.

[0053] Preferably, in order to reduce randomness, when constructing a rectangular window for each pixel, the filtering rectangular window of adjacent pixels should be similar, otherwise, the final burr generation result will tend to be disordered noise characteristics.

[0054] It should be noted that the selection of the filtering method is not limited in the embodiment of the application, and can be selected according to actual conditions. For example, frequency domain filtering, median filtering, average filtering, Gaussian filtering, etc.

[0055] Calculating the residual error according to the original image and the second image to obtain a burr component.

[0056] Y=C * ( Y2-Y1) .

[0057] Wherein, C is a burr region statistical component, Y1 represents the original image, and Y2 represents the second image.

[0058] Step 4, fusing the original image and the burr component to obtain a burr edge simulation image.

[0059] The calculation formula of the obtained burr edge simulation image is as follows:

[0060] result=Y1+α * Y.

[0061] Wherein, Y1 represents an original image, Y represents a burr component, and a represents a fusion coefficient, and a >= 0.

[0062] It should be noted that the specific value of a as the fusion coefficient is not limited by the embodiment of the present application, and in actual application, the strength of the burr can be adjusted by adjusting the size of the fusion coefficient. The greater the fusion coefficient, the stronger the burr; the smaller the fusion coefficient, the weaker the burr.

[0063] Figure 4 The structure schematic diagram of the burr edge simulation quantitative device provided by the embodiment of the present application is shown as follows, Figure 4 The burr edge simulation quantitative device provided by the embodiment of the present application specifically includes the following contents:

[0064] The edge detection module is used for traversing all pixels in the original image and performing edge detection on the original image.

[0065] It should be noted that the method of edge detection is not limited by the embodiment of the present application, and the corresponding edge detection method can be selected according to the actual situation.

[0066] In one embodiment, the traversing all pixels in the original image and performing edge detection on the original image includes:

[0067] A square first pixel region centered on the current pixel is divided, and the size of the first pixel region is K*K.

[0068] Wherein, K=2r+1; r is a natural number, representing the radius of the first pixel region.

[0069] The first pixel region is reduced or enlarged to a second pixel region of a preset size.

[0070] The size of the second pixel region is not less than 5*5.

[0071] The second derivative of the current pixel point is calculated according to the pixels in the second pixel region.

[0072] It should be noted that the method of calculating the second derivative is not limited by the embodiment of the present application, and the derivative method can be selected according to the actual situation.

[0073] Different sizes of the first pixel region are selected, and the above steps are repeated to obtain a multi-scale second derivative.

[0074] The function relationship f(r) between the multi-scale second derivative and the radius of the first pixel region is fitted.

[0075] According to the function relationship f(r), one or more first pixel region radius sizes r_argmax corresponding to the maximum value of the second derivative of the current pixel are counted.

[0076] A threshold T is preset for the first pixel region radius, if the first pixel region radius r_argmax is less than the threshold T, the current pixel is a strong edge pixel, if the first pixel region radius r_argmax is greater than the threshold T, the current pixel is a weak edge pixel.

[0077] The burr region determination module is used for simulating the burr region based on the strong edge pixel.

[0078] The burr region based on the strong edge pixel, and the burr region is simulated, and the burr region determination module comprises the following steps:

[0079] The preset burr generating range does not exceed a square region of R*R centered on the current strong edge pixel, and one or more rectangular regions containing the current strong edge pixel are randomly selected in the square region.

[0080] It should be noted that the size of the square region is not limited in the embodiment of the application, and can be preset according to actual conditions.

[0081] The rectangular regions are superimposed to obtain a union region, and the union region is the burr region.

[0082] All the strong edge pixels are traversed to obtain the corresponding burr region.

[0083] The burr component construction module is used for constructing a burr component according to the original image.

[0084] The burr component is constructed according to the original image, and the burr component construction module comprises the following steps:

[0085] The pixels in the burr region are marked as 1, and the other pixels are marked as 0, and a burr region statistical component C of the current image is constructed.

[0086] A rectangular window is randomly constructed for each pixel.

[0087] The size of the rectangular window does not exceed a square region of R'*R' centered on the current pixel, and the rectangular window contains the current pixel.

[0088] It should be noted that the size of the square region is not limited in the embodiment of the application, and can be preset according to actual conditions.

[0089] The original image is filtered by using the rectangular window to obtain a second image.

[0090] Preferably, in order to reduce randomness, when the rectangular window is randomly constructed for each pixel, the selection of the filtering rectangular window of the adjacent pixel should be similar, otherwise, the final burr generation result will tend to be disordered noise characteristics.

[0091] It should be noted that the embodiment of the present application does not limit the selection of the filtering mode, and can be selected according to the actual situation. For example, frequency domain filtering, median filtering, average filtering, Gaussian filtering, etc.

[0092] According to the original image and the second image, a residual error is calculated to obtain a burr component.

[0093] Y=C * (Y2-Y1) .

[0094] Wherein, C is a burr area statistical component, Y1 represents the original image, and Y2 represents the second image.

[0095] An image fusion module is configured to fuse the original image and the burr component to obtain a burr edge simulation image.

[0096] The burr edge simulation image obtained by fusion is calculated according to the following formula:

[0097] result=Y1+α * Y.

[0098] Wherein, Y1 represents the original image, Y represents the burr component, and alpha represents the fusion coefficient, alpha >= 0.

[0099] It should be noted that the value of alpha as the fusion coefficient is not limited by the embodiment of the present application, and in actual application, the strength of the burr can be adjusted by adjusting the size of the fusion coefficient. The greater the fusion coefficient, the stronger the burr; the smaller the fusion coefficient, the weaker the burr.

[0100] Based on the same inventive concept, another embodiment of the present application provides an electronic device, as shown in the figure. Figure 5 The electronic device specifically includes the following contents: a processor 501, a memory 502, a communication interface 503 and a communication bus 504.

[0101] Wherein, the processor 501, the memory 502 and the communication interface 503 complete the communication among each other through the communication bus 504; the communication interface 503 is used to realize the information transmission between devices.

[0102] The processor 501 is used to call the computer program in the memory 502, and the processor executes the computer program to realize all steps of the burr edge simulation quantitative method.

[0103] Based on the same inventive concept, another embodiment of the present application provides a non-transitory computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize all steps of the burr edge simulation quantitative method.

[0104] In addition, in the embodiments of the present application, the relationship terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or sequence between these entities or operations. In the description of the present specification, the description of "in one embodiment", "in another embodiment" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment are contained in at least one embodiment of the present application. The illustrative expressions in the present specification do not necessarily refer to the same embodiment. Moreover, the specific features, structures, materials or characteristics described can be combined in any suitable manner in any one or more embodiments. In addition, different embodiments described in the present specification and the features of different embodiments can be combined and combined by those skilled in the art without contradiction.

[0105] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A burr edge simulation quantitative method, characterized in that, The method comprises the following steps: Edge detection is performed on all pixels in the original image, and then the region where the burr is located is simulated based on the strong edge pixel points; Next, the burr component is constructed according to the original image; Finally, the burr edge simulation image is obtained by fusing the original image and the burr component; The burr component constructed according to the original image comprises: Marking the pixels in the burr region as 1 and marking other pixels as 0 to construct the burr region statistical component C of the original image; A rectangular window is constructed for each pixel; The size of the rectangular window is not more than the square region of R'*R' centered on the current pixel, and the rectangular window contains the current pixel; The original image is filtered by using the rectangular window to obtain a second image; The burr component is obtained by calculating the residual error between the original image and the second image; Y=C * ( Y2-Y1) ; Wherein, C is the burr region statistical component, Y1 represents the original image, and Y2 represents the second image.

2. The burr edge simulation quantitative method according to claim 1, characterized in that, The edge detection on the original image comprises: A square first pixel region centered on the current pixel is divided, and the size of the first pixel region is K*K; Wherein, K=2r+1; r is a natural number, representing the radius of the first pixel region; The first pixel region is reduced or enlarged to a second pixel region of a preset size; The size of the second pixel region is not less than 5*5; The second derivative of the current pixel point is calculated according to the pixels in the second pixel region; Different sizes of the first pixel region are selected, and the above steps are repeated to obtain the second derivative of multiple scales; The function relationship f(r) between the second derivative of multiple scales and the radius of the first pixel region is fitted; According to the function relationship f(r), one or more first pixel region radii r_argmax corresponding to the maximum value of the second derivative of the current pixel are counted; A threshold T is preset for the first pixel region radius, if the first pixel region radius r_argmax is less than the threshold T, the current pixel is a strong edge pixel; if the first pixel region radius r_argmax is greater than the threshold T, the current pixel is a weak edge pixel.

3. The burr edge emulation quantification method of claim 1, wherein, The region where the burr is located is simulated based on the strong edge pixel points, comprising: The range of burr generation is preset to be not more than the square region of R*R centered on the current strong edge pixel, and one or more rectangular regions containing the current strong edge pixel are randomly selected in the square region; The rectangular regions are superimposed to obtain a union region, which is the region where the burr is located; All strong edge pixels are traversed to obtain the corresponding region where the burr is located.

4. The burr edge emulation quantification method of claim 1, wherein, The calculation formula for fusing the original image and the burr component to obtain the burr edge simulation image is as follows: result=Y1+α * Y; Wherein, Y1 represents the original image, Y represents the burr component, and a represents the fusion coefficient, a ≥0.

5. A burr edge simulation quantification device, characterized by, The method comprises the following steps: An edge detection module is used to traverse all pixels in the original image and perform edge detection on the original image; A burr region determination module is used to simulate the region where the burr is located based on the strong edge pixel points; A burr component construction module is used to construct the burr component according to the original image; The constructing burr component from the original image comprises: Marking pixels in a burr area as 1 and marking other pixels as 0 to construct a burr area statistic component C of the original image; Randomly constructing a rectangular window for each pixel; The size of the rectangular window is not more than a square area of R'*R' centered on the current pixel, and the rectangular window contains the current pixel; Filtering the original image with the rectangular window to obtain a second image; Calculating a residual from the original image and the second image to obtain a burr component; Y = C * (Y2-Y1); Wherein, C is a burr area statistic component, Y1 represents the original image, and Y2 represents the second image; An image fusion module is configured to fuse the original image and the burr component to obtain a burr edge simulation image.

6. The burr edge simulation quantification device of claim 5, wherein, The edge detection of the original image comprises: Dividing a square first pixel area centered on the current pixel, and the size of the first pixel area is K*K; Wherein, K = 2r + 1; r is a natural number, representing the radius of the first pixel area; Reducing or enlarging the first pixel area to a second pixel area of a preset size; The size of the second pixel area is not less than 5*5; Calculating the second derivative of the current pixel point according to the pixels in the second pixel area; Selecting first pixel areas of different sizes and repeating the above steps to obtain multi-scale second derivatives; Fitting a function relationship f(r) between the multi-scale second derivatives and the radius of the first pixel area; According to the function relationship f(r), one or more first pixel area radii r_argmax corresponding to the maximum value of the second derivative of the current pixel are calculated; A threshold T is preset for the radius of the first pixel area, if the radius r_argmax of the first pixel area is less than the threshold T, the current pixel is a strong edge pixel; if the radius r_argmax of the first pixel area is greater than the threshold T, the current pixel is a weak edge pixel.

7. The burr edge simulation quantification device of claim 5, wherein, The simulation of the burr area based on the strong edge pixel comprises: Presetting that the range of burr generation is not more than a square area of R*R centered on the current strong edge pixel, and randomly selecting one or more rectangular areas containing the current strong edge pixel in the square area; Superimposing the rectangular areas to obtain a union area, which is the burr area; Traversing all strong edge pixels to obtain corresponding burr areas.

8. The burr edge simulation quantification device of claim 5, wherein, The fusion of the original image and the burr component to obtain the burr edge simulation image is calculated according to the following formula: result = Y1 + a * Y; Wherein, Y1 represents the original image, Y represents the burr component, a represents the fusion coefficient, and 0 ≤ a.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the burr edge simulation quantitative method of any one of claims 1 to 4.

10. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the burr edge simulation quantitative method of any one of claims 1 to 4.

Images