Image denoising processing method and device, computer device and storage medium

Abstract

The application relates to the technical field of image processing, in particular to an image denoising processing method and device, computer equipment and a storage medium, which can solve the problem of excessive loss of image details and edges caused by guided filter denoising under complex image environment. The image denoising processing method first determines a to-be-processed image, which is a YUV format image with noise; the initial brightness component of the to-be-processed image is denoised through a first initial guide image, so that the determination of a target brightness component after denoising is realized; the initial chroma component of the to-be-processed image is denoised through a second initial guide image or mean filter, so that the determination of a target chroma component after denoising is realized; and the first target image is determined through the target brightness component and the target chroma component, so that the image quality is improved through denoising, and the details and edges of the image are fully retained.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and more specifically, to an image noise reduction processing method, apparatus, computer device, and storage medium. Background Technology

[0002] As imaging devices are widely used in fields such as security, intelligent transportation, and intelligent manufacturing, the captured images contain a variety of complex noise information, requiring noise reduction processing to improve image quality.

[0003] In related technologies, guided filtering is used to denoise images acquired by a camera. Guided filtering was proposed in the paper "Guided Image Filtering" by He Kaiming et al. Guided filtering overcomes the high computational complexity of bilateral filtering and improves edge protection during denoising. However, in complex image environments, guided filtering can lead to excessive loss of detail and edge features. Summary of the Invention

[0004] To address the problem of excessive loss of image detail edges caused by guided filtering noise reduction in complex image environments, this application provides an image noise reduction method, apparatus, computer device, and storage medium.

[0005] The embodiments of this application are implemented as follows:

[0006] The first aspect of this application provides an image noise reduction processing method, including the following steps:

[0007] The image to be processed is a noisy YUV format image;

[0008] Based on the first initial guiding image, the initial luminance component of the image to be processed is denoised, and the target luminance component after denoising is determined. The first initial guiding image is the mean filtered image of the initial luminance component of the image to be processed under the first preset window.

[0009] The initial chromaticity component of the image to be processed is denoised based on the second initial guiding image or mean filtering, and the target chromaticity component after denoising is determined. The second initial guiding image is the mean filtering image of the initial chromaticity component of the image to be processed under the first preset window.

[0010] The first target image is determined based on the target luminance component and the target chrominance component.

[0011] A second aspect of this application provides an image noise reduction processing apparatus, including an acquisition module, an execution module, and an output module;

[0012] The acquisition module is used to determine the image to be processed, which is a noisy YUV format image;

[0013] The execution module is used to denoise the initial luminance component of the image to be processed based on the first initial guiding image, and determine the target luminance component after denoising; it is also used to denoise the initial chrominance component of the image to be processed based on the first initial guiding image or mean filtering, and determine the target chrominance component after denoising; wherein, the first initial guiding image is the mean-filtered image of the initial luminance component of the image to be processed under the first preset window, and the second initial guiding image is the mean-filtered image of the initial chrominance component of the image to be processed under the first preset window.

[0014] The output module is used to determine the first target image based on the target luminance component and the target chrominance component.

[0015] A third aspect of this application provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the image noise reduction processing method of the first aspect.

[0016] A fourth aspect of this application provides a computer storage medium on which a computer program is stored. When the computer program is executed by a processor, the processor performs the steps of the image noise reduction processing method of the first aspect.

[0017] The beneficial effects of this application: This application provides an image denoising processing method, apparatus, computer device, and storage medium. The image denoising processing method first determines the image to be processed, which is a noisy YUV format image. The initial luminance component of the image to be processed is filtered by mean under a first preset window to determine a first initial guiding image. By denoising the initial luminance component of the image to be processed using the first initial guiding image, the target luminance component after denoising can be determined. By denoising the initial chrominance component of the image to be processed using the first initial guiding image or mean filtering, the target chrominance component after denoising can be determined. The first target image is determined using the target luminance component and the target chrominance component. Through the above denoising processing, noise can be reduced and image quality improved while ensuring that the detailed edges of the image are fully preserved. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 The diagram illustrates a flowchart of an image noise reduction method according to some embodiments of this application;

[0020] Figure 2 The diagram illustrates a flowchart of the process for determining the target luminance component after noise reduction, provided in some embodiments of this application.

[0021] Figure 3 The present application provides a flowchart illustrating the determination of the first variable according to some embodiments;

[0022] Figure 4 The illustration shows a flowchart of the process for determining target chromaticity components according to some embodiments of this application;

[0023] Figure 5 The illustration shows a flowchart of the process for determining a first target chromaticity component according to some embodiments of this application;

[0024] Figure 6 A schematic diagram of the process for determining a second target chromaticity component is shown in some embodiments of this application;

[0025] Figure 7 The diagram illustrates a flowchart of yet another image noise reduction method provided in some embodiments of this application;

[0026] Figure 8 The diagram illustrates a flowchart of another image noise reduction method provided by some embodiments of this application;

[0027] Figure 9 A schematic diagram of the structure of an image noise reduction processing device provided in an embodiment of this application is shown. Detailed Implementation

[0028] To make the objectives, implementation methods and advantages of this application clearer, the exemplary implementation methods of this application will be clearly and completely described below with reference to the accompanying drawings of the exemplary embodiments of this application. Obviously, the described exemplary embodiments are only some embodiments of this application, and not all embodiments.

[0029] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0030] The terms "first," "second," "third," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar or related objects or entities, and do not necessarily imply a specific order or sequence, unless otherwise specified. It should be understood that such terms are interchangeable where appropriate.

[0031] The terms “include” and “have”, and any variations thereof, are intended to cover but not exclusively include, for example, a product or device that includes a range of components is not necessarily limited to all of the components that are clearly listed, but may include other components that are not clearly listed or that are inherent to such product or device.

[0032] Guided filtering was proposed by Dr. Kaiming He et al. in their paper "Guided Image Filtering." Guided filtering is a type of edge-preserving filtering that, compared to bilateral filtering, overcomes the high computational complexity and preserves image edges during noise reduction. In practical applications of image filtering, excessive loss of image edge details can occur in complex image environments.

[0033] In complex image environments, to mitigate excessive loss of image edge details during image denoising, this application provides an image denoising method, apparatus, computer device, and storage medium. The image denoising method first determines the image to be processed, which is a noisy YUV format image. The initial luminance component of the image to be processed is filtered by mean under a first preset window to determine a first initial guiding image. Noise reduction of the initial luminance component of the image to be processed using the first initial guiding image allows for the determination of the denoised target luminance component. Noise reduction of the initial chrominance component of the image to be processed using either the first initial guiding image or mean filtering allows for the determination of the denoised target chrominance component. The first target image is determined using the target luminance component and the target chrominance component. Through the above denoising process, noise can be reduced and image quality improved while ensuring that the detailed edges of the image are fully preserved.

[0034] The image noise reduction processing method, apparatus, computer equipment, and storage medium of the present application embodiments are described in detail below with reference to the accompanying drawings.

[0035] Figure 1 The following is a flowchart illustrating an image noise reduction method according to some embodiments of this application, such as... Figure 1 As shown in the figure, this application provides an image noise reduction processing method.

[0036] The image noise reduction method includes the following steps:

[0037] S110. Determine the image to be processed. The image to be processed is a noisy YUV format image.

[0038] YUV encoding uses luminance and chrominance to represent the color of each pixel. Y represents luminance (Luminance, Luma), which is the grayscale value. U and V represent chrominance (Chrominance or Chroma). For the image to be processed, there are initial luminance components (Y component) and initial chrominance components (U component and V component).

[0039] It should be understood that the image to be processed is a noisy color image.

[0040] S120. A first initial guiding image is determined based on the mean-filtered image of the initial luminance component of the image to be processed under a first preset window; and a second initial guiding image is determined based on the mean-filtered image of the initial chrominance component of the image to be processed under a first preset window.

[0041] The first initial guide image is a mean-filtered image of the initial luminance component (Y component) of the image to be processed within a first preset window. This mean-filtered image improves the filtering effect on noise.

[0042] The second initial guide image is a mean-filtered image of the initial chromaticity components (U component or V component) of the image to be processed under the first preset window. This mean-filtered image improves the filtering effect on noise.

[0043] It should be understood that the second initial guide image corresponding to the initial chroma component being the U component is different from the second initial guide image corresponding to the initial chroma component being the V component.

[0044] The window size for mean filtering can be determined by a first preset window or by scene characteristics; that is, the window size can be adaptively determined based on the changes in the neighboring pixel values ​​around the pixel. For example, a larger window can be used in flat image areas, and a smaller window can be used in image areas with rich details.

[0045] For example, the window size for mean filtering can be set to 3*3, meaning the window size is 3 rows and 3 columns.

[0046] Taking the first initial guide image as an example, the first initial guide image can be calculated according to the following formula:

[0047] I CD (x, y) = Mean(Y(x, y))

[0048] In the formula, I CD Let (x, y) be the first initial guiding image, (x, y) be the row and column values ​​of the image, and Mean represent the mean filtering process applied to the Y component.

[0049] In some embodiments, the first initial guide image may be an image related to the image to be processed, or it may be other images, that is, it may be adjusted according to the scenario.

[0050] The principle of the second initial guide image is similar to that of the first initial guide image, and will not be repeated here.

[0051] When denoising the image, the luminance component (Y component) and chrominance components (U component and V component) are processed separately. The specific processing steps are as follows:

[0052] S130: Based on the first initial guiding image, the initial luminance component of the image to be processed is denoised, and the target luminance component after denoising is determined.

[0053] Figure 2 The following is a schematic diagram illustrating the process of determining the target luminance component after noise reduction according to some embodiments of this application, such as... Figure 2 As shown, in step 130, based on the first initial guiding image, the initial luminance component of the image to be processed is denoised, and the denoised target luminance component is determined, including the following steps:

[0054] S301. Based on the intermediate brightness component and the first intermediate guide image, determine the first image standard deviation parameter and the second image standard deviation parameter, wherein the intermediate brightness component is determined by normalizing the initial brightness component, and the first intermediate guide image is determined by normalizing the first initial guide image.

[0055] The initial luminance component and the first initial guide image are normalized to change the range of pixel values ​​from [0-255] to [0-1]. By normalizing the initial luminance component and the first initial guide image, the data can still remain within the range of image pixel values ​​[0-255] after the following process.

[0056] The initial luminance component and the first initial guiding image, after normalization, can be calculated using the following formula:

[0057] Y 1 (x, y) = Y(x, y) / 255

[0058]

[0059] In the formula, Y is the initial luminance component, and I... CD Let (x, y) be the first initial guiding image, and (x, y) be the row and column values ​​corresponding to the image. 1 For the intermediate brightness component, This is the first intermediate guiding image.

[0060] The image of the intermediate brightness component, the first intermediate guide image, the image after multiplying the intermediate brightness component and the first intermediate guide image, and the image after multiplying the first intermediate guide image are subjected to mean filtering under a second preset window to determine the standard deviation parameters of the first image and the second image.

[0061] Based on the intermediate brightness component and the first intermediate guide image, the standard deviation parameters of the first image and the second image can be determined by the following formula:

[0062] mean Y (x, y) = Mean(Y) 1 (x, y))

[0063]

[0064]

[0065]

[0066] var Yb (x, y) = corr Yb (x, y) - mean Yb (x, y).*mean Yb (x, y)

[0067] cov Y (x, y) = corr Y (x, y) - mean Yb (x, y).*mean Y (x, y)

[0068] In the formula, .* represents the element-wise multiplication operation of matrices, and Y 1 For the intermediate brightness component, This is the first intermediate guiding image. The image is the result of multiplying the intermediate brightness component and the first intermediate guide image. The image is the result of multiplying the first intermediate guiding image and the second intermediate guiding image. Mean represents the mean filtering applied to the Y component, where (x, y) are the corresponding row and column values ​​of the image. Y The image after mean filtering of the intermediate brightness component within the second preset window. Yb The first intermediate guiding image is the image after mean filtering under the second preset window, corr Y The image obtained by multiplying the intermediate brightness component and the first intermediate guide image is the image after mean filtering under the second preset window, corr YbThe image is the result of multiplying the first intermediate guide image and the image after mean filtering within a second preset window, var Yb First image standard deviation parameter, corr Yb This is the standard deviation parameter for the second image.

[0069] For image Y 1 , and Perform mean filtering within the second preset window, and simultaneously determine the standard deviation parameter var of the first image. Yb The second image standard deviation parameter cov Y .

[0070] In some embodiments, the window size for mean filtering in step 301 can be determined based on a second preset window or based on scene features; that is, the window size can be adaptively determined based on the changes in the neighboring pixel values ​​around the pixel at this time. For example, a larger window can be used in a flat image area, and a smaller window can be used in an image area with rich details.

[0071] For example, the window size for mean filtering can be set to 7*7, meaning the window size is 7 rows and 7 columns.

[0072] S302. Based on the first image standard deviation parameter and the control edge preservation degree parameter, determine the first variable, wherein the control edge preservation degree parameter is a non-negative constant less than 1.

[0073] Figure 3 The following is a schematic diagram illustrating the process of determining the first variable provided in some embodiments of this application, such as... Figure 3 As shown, the first variable is determined based on the first image standard deviation parameter and the parameter controlling the edge preservation degree, including the following steps:

[0074] S3021. Determine the first sum value based on the control edge preservation degree parameter and the first image standard deviation parameter.

[0075] The first intermediate value of the points in the image is determined by controlling the edge preservation parameter and the first image standard deviation parameter. The first intermediate values ​​of each point in the image are added together to determine the first sum.

[0076] The first sum is obtained by the following formula:

[0077]

[0078]

[0079] In the formula, denomin(x, y) is the first intermediate quantity, and var Yb(x, y) is the first image standard deviation parameter, ∈ is the parameter controlling the degree of edge preservation, and denomin sum is the first sum; m and n are the length and width of the image to be processed.

[0080] It should be understood that during the noise reduction process of the initial luminance component of the image to be processed, m and n can also be the length and width of the initial luminance component.

[0081] In some embodiments, the edge protection parameter is a preset threshold and is a non-negative constant less than 1.

[0082] S3022. Determine the first variable based on the first sum, the control edge preservation parameter, and the first image standard deviation parameter.

[0083] The first variable is obtained by calculating the following formula:

[0084] sigma(x, y) = (var Yb (x, y)+∈)*denomin sum / (m*n)

[0085] In the formula, sigma(x, y) is the first variable.

[0086] like Figure 2 As shown, it also includes: S303, determining the first coefficient and the second coefficient based on the first image standard deviation parameter, the second image standard deviation parameter, the first variable and the preset correction parameter.

[0087] The first coefficient is obtained by the following formula:

[0088] a(x, y) = cov Y (x, y). / (var Yb (x, y) + (eps. / sigma(x, y)))

[0089] In the formula, a is the first coefficient, eps. is the preset correction parameter, and . / is the element-wise division operation of the matrix.

[0090] The second coefficient is obtained by calculating the following formula:

[0091] b(x, y) = mean Y (x, y) - a(x, y).*mean Yb (x, y)

[0092] In the formula, b is the first coefficient.

[0093] It should be understood that the first and second coefficients can be adjusted by modifying the preset correction parameters. The strength of the noise reduction effect can be controlled by these preset correction parameters.

[0094] In some embodiments, the preset calibration parameters can also be determined by obtaining them.

[0095] In some embodiments, mean filtering can also be applied to the first coefficient and the second system to update the first coefficient and the second system, thereby improving the noise reduction effect. This can be calculated using the following formula:

[0096] mean a (x, y) = Mean(a(x, y))

[0097] mean b (x, y) = Mean(b(x, y))

[0098] In the formula, mean a The first coefficient after updating by mean filtering, mean b This is the second coefficient updated after mean filtering.

[0099] like Figure 2 As shown, it also includes: S304, determining the target brightness component based on the first coefficient, the second coefficient and the first initial guiding image, and inverse normalization.

[0100] It should be noted that the first and second coefficients described here can be the first and second coefficients before mean filtering, or the first and second coefficients after mean filtering.

[0101] When the first and second coefficients are the same as those obtained after mean filtering, the target brightness component can be calculated using the following formula:

[0102]

[0103] In the formula, Y new For the target brightness component.

[0104] In step 130, the brightness component processing of image detail edges is improved by determining the first variable and preset correction parameters during the noise reduction process.

[0105] like Figure 1 As shown, it also includes: S140, denoising the initial chromaticity component of the image to be processed based on the second initial guiding image or mean filtering, and determining the denoised target chromaticity component.

[0106] The noise reduction process for the initial chromaticity components (U and V components) can be the same as the process for the initial luminance components, determining the target chromaticity component (U) after noise reduction. new V new ).

[0107] The initial chromaticity components include the first initial chromaticity component (U component) and the second initial chromaticity component (V component). Figure 4 The following are schematic diagrams illustrating the process for determining target chromaticity components provided in some embodiments of this application, such as... Figure 4 As shown, in step 140, the initial chromaticity component of the image to be processed is denoised based on the second initial guiding image, and the target chromaticity component after denoising is determined. This may include the following steps:

[0108] S401. Based on the second initial guiding image, denoise the first initial chromaticity component of the image to be processed, and determine the first target chromaticity component after denoising.

[0109] The second initial guiding image at this point corresponds to the U component.

[0110] S402. Based on the second initial guiding image, denoise the second initial chromaticity component of the image to be processed, and determine the denoised second target chromaticity component.

[0111] The second initial guide image at this point corresponds to the V component.

[0112] Figure 5 The following is a schematic diagram illustrating the process of determining the first target chromaticity component according to some embodiments of this application, such as... Figure 5 As shown, the method for denoising the first initial chromaticity component of the image to be processed based on the second initial guiding image (U component) and determining the first target chromaticity component after denoising includes the following steps:

[0113] S4011. Based on the first intermediate chromaticity component and the second intermediate guide image, determine the standard deviation parameters of the third image and the fourth image, wherein the first intermediate chromaticity component is determined by normalizing the first initial chromaticity component, and the second intermediate guide image is determined by normalizing the second initial guide image.

[0114] S4012. Based on the third image standard deviation parameter and the control edge preservation degree parameter, determine the second variable, wherein the control edge preservation degree parameter is a non-negative constant less than 1.

[0115] The determination of the second variable includes: determining the second sum based on the control edge preservation parameter and the third image standard deviation parameter; wherein, the second intermediate value of the corresponding point in the image is determined by controlling the edge preservation parameter and the third image standard deviation parameter, and the second intermediate value corresponding to each point in the image is added together to determine the second sum. The second variable is then determined based on the second sum, the control edge preservation parameter, and the third image standard deviation parameter.

[0116] S4013. Determine the third coefficient and the fourth coefficient based on the third image standard deviation parameter, the fourth image standard deviation parameter, the second variable, and the preset correction parameter.

[0117] In some embodiments, mean filtering can be applied to the third coefficient and the fourth system to update the third coefficient and the fourth system and improve the noise reduction effect.

[0118] S4014. Based on the third coefficient, the fourth coefficient, and the second initial guiding image, as well as inverse normalization, determine the first target chromaticity component.

[0119] It should be noted that the third and fourth coefficients described here can be the third and fourth coefficients before mean filtering, or the third and fourth coefficients after mean filtering.

[0120] Figure 6 The following is a schematic diagram illustrating the process for determining the second target chromaticity component according to some embodiments of this application, such as... Figure 6 As shown, the method for denoising the second initial chromaticity component of the image to be processed based on the second initial guiding image (V component) and determining the denoised second target chromaticity component includes the following steps:

[0121] S4021. Based on the second intermediate chromaticity component and the second intermediate guide image, determine the standard deviation parameters of the fifth image and the sixth image, wherein the second intermediate chromaticity component is determined by normalizing the second initial chromaticity component, and the second intermediate guide image is determined by normalizing the second initial guide image.

[0122] S4022. Based on the standard deviation parameter of the fifth image and the control edge preservation parameter, determine the third variable, wherein the control edge preservation parameter is a non-negative constant less than 1.

[0123] The determination of the third variable includes: determining the third sum based on the control edge preservation parameter and the fifth image standard deviation parameter; specifically, determining the third intermediate value for each point in the image by using the control edge preservation parameter and the fifth image standard deviation parameter, and summing the third intermediate values ​​for each point in the image to determine the third sum; and determining the third variable based on the third sum, the control edge preservation parameter, and the fifth image standard deviation parameter.

[0124] S4023. Determine the fifth coefficient and the sixth coefficient based on the fifth image standard deviation parameter, the sixth image standard deviation parameter, the third variable, and the preset correction parameter.

[0125] In some embodiments, mean filtering can be applied to the fifth and sixth coefficients to update them and improve the noise reduction effect.

[0126] S4024. Based on the fifth coefficient, the sixth coefficient, and the second initial guiding image, as well as inverse normalization, determine the second target chromaticity component.

[0127] It should be noted that the third and fourth coefficients described here can be the third and fourth coefficients before mean filtering, or the third and fourth coefficients after mean filtering.

[0128] In some embodiments, in step 140, the target chromaticity component after denoising is determined based on the initial chromaticity component of the image to be processed by the second initial guiding image. Alternatively, the target chromaticity component after denoising can be determined by the guiding filter proposed by Dr. He Kaiming, based on the initial chromaticity component of the image to be processed by the second initial guiding image.

[0129] It should be understood that the guided filtering process proposed by Dr. He Kaiming is also a process of denoising the initial chromaticity components (U component and V component) separately, and determining the corresponding denoised target chromaticity component (U... new V new ).

[0130] In some embodiments, step 140, which involves denoising the initial chroma components of the image to be processed based on mean filtering and determining the denoised target chroma components, may further include:

[0131] By applying the mean-filtered image of the initial chromaticity components (U component, V component) within a third preset window, the corresponding denoised target chromaticity component (U component) is determined using this mean-filtered image. new V new ).

[0132] The window size for mean filtering can be determined by a third preset window or by scene characteristics; that is, the window size can be adaptively determined based on the changes in the values ​​of neighboring pixels around the pixel. For example, a larger window can be used in flat image areas, and a smaller window can be used in image areas with rich details.

[0133] For example, the window size for mean filtering can be set to 7*7, meaning the window size is 7 rows and 7 columns.

[0134] In some embodiments, before the noise reduction processing of the initial chromaticity components (U component, V component) in step 140, the step may be: performing downsampling processing on the image to be processed based on a preset downsampling module.

[0135] Downsampling simplifies the initial chromaticity components of the image to be processed. Downsampling scales the size of the image to be processed to reduce the number of pixels involved in subsequent noise reduction.

[0136] The preset downsampling module can determine the image to be processed. For example, if the preset downsampling module determines that it is based on 2*2 downsampling, it can take the average value of the 2*2 pixel values ​​in the image to be processed as a new pixel, so that the amount of data involved in the calculation is one-quarter of the amount of data in the image to be processed.

[0137] like Figure 1 As shown, it also includes: S150, determining the first target image based on the target luminance component and the target chrominance component.

[0138] Based on the target brightness component (Y) new ) and target chromaticity component (U new V new ), to determine the first target image.

[0139] In some embodiments, the color image acquired by the imaging device is an RGB format image, and the image requiring noise reduction is also an RGB format image. Figure 7 The following is a schematic flowchart illustrating another image noise reduction method provided in some embodiments of this application, such as... Figure 7 As shown, this application provides another image noise reduction method compared to... Figure 1 The image noise reduction processing method shown includes, before determining the image to be processed, acquiring a captured image, wherein the captured image is an RGB format image with noise; converting the captured image into the image to be processed; and after determining the first target image, converting the first target image into a second target image, wherein the second target image is an RGB format image.

[0140] The image noise reduction method includes the following steps:

[0141] S210. Acquire a captured image, wherein the captured image is an RGB format image with noise.

[0142] S220: Convert the captured image into an image to be processed, and determine the image to be processed. The image to be processed is a noisy YUV format image.

[0143] The captured image can be converted into an image to be processed using the following formula:

[0144] Y(x,y)=0.299R(x,y)+0.587G(x,y)+0.114B(x,y)

[0145] V(x,y)=0.713(R(x,y)-Y(x,y))+128

[0146] U(x,y)=0.564(B(x,y)-Y(x,y))+128

[0147] In the formula, R, G, B are the RGB component values ​​of the captured image, Y, U, V are the YUV component values ​​of the image to be processed, and (x, y) are the row and column values ​​of the image.

[0148] S230. Based on the mean-filtered image of the initial luminance component of the image to be processed under the first preset window, determine the first initial guide image; based on the mean-filtered image of the initial chrominance component of the image to be processed under the first preset window, determine the second initial guide image.

[0149] S240. Based on the first initial guiding image, the initial luminance component of the image to be processed is denoised, and the target luminance component after denoising is determined.

[0150] S250: Based on the second initial guiding image or mean filtering, the initial chromaticity component of the image to be processed is denoised, and the target chromaticity component after denoising is determined.

[0151] S260. Determine the first target image based on the target luminance component and the target chrominance component.

[0152] The process from step 230 to step 260 is the same as the process from step 120 to step 150, and will not be repeated here.

[0153] S270. Convert the first target image into a second target image, wherein the second target image is an RGB format image.

[0154] R out (x, y) = Y new (x, y) + 1.403(V) new (x, y) - 128)

[0155] G out (x, y) = Y new (x, y) - 0.344(V) new (x, y) - 128) - 0.714(U new (x, y) - 128)

[0156] B out (x, y) = Y new (x, y) + 1.773(U new (x, y) - 128)

[0157] In the formula, R out G out B out The RGB component values ​​of the second target image, Y new U new V new Let (x, y) be the YUV component values ​​of the first target image, and (x, y) be the row and column values ​​of the corresponding image.

[0158] In some embodiments, the method further includes outputting a denoised second target image.

[0159] Figure 8 The following is a flowchart illustrating another image noise reduction method provided by some embodiments of this application, such as... Figure 8 As shown, the image noise reduction method includes the following steps:

[0160] S310. Obtain preset correction parameters and captured image; S320. Convert captured image into image to be processed (RGB domain to YUV domain); S330. Determine guide image; S340. Perform noise reduction processing on the luminance component of the image to be processed; S350. Perform noise reduction processing on the chrominance component of the image to be processed; S360. Obtain the first target image after noise reduction; S370. Convert the first target image into a second target image (YUV domain to RGB domain conversion); S380. Output the second target image after noise reduction.

[0161] This application provides an image denoising method. First, an image to be processed is determined, which is a noisy YUV format image. The initial luminance component of the image to be processed is filtered by mean under a first preset window to determine a first initial guiding image. The initial chrominance component of the image to be processed is filtered by mean under a first preset window to determine a second initial guiding image. By denoising the initial luminance component of the image to be processed using the first initial guiding image, the target luminance component after denoising can be determined. By denoising the initial chrominance component of the image to be processed using the second initial guiding image or mean filtering, the target chrominance component after denoising can be determined. The first target image is determined by the target luminance component and the target chrominance component. Through the above denoising process, noise can be reduced and image quality can be improved while ensuring that the details and edges of the image are fully preserved.

[0162] Figure 9 This invention provides a schematic diagram of the structure of an image noise reduction processing apparatus according to an embodiment of the present application. Figure 9 As shown, the image noise reduction processing device 900 includes an acquisition module 910, an execution module 920, and an output module 930.

[0163] The acquisition module is used to determine the image to be processed, which is a noisy YUV format image;

[0164] The execution module is used to denoise the initial luminance component of the image to be processed based on the first initial guiding image, and determine the target luminance component after denoising; it is also used to denoise the initial chrominance component of the image to be processed based on the first initial guiding image or mean filtering, and determine the target chrominance component after denoising; wherein, the first initial guiding image is the mean-filtered image of the initial luminance component of the image to be processed under the first preset window, and the second initial guiding image is the mean-filtered image of the initial chrominance component of the image to be processed under the first preset window.

[0165] The output module is used to determine the first target image based on the target luminance component and the target chrominance component.

[0166] In some embodiments, the execution module further includes a luminance component execution unit, configured to determine the denoised target luminance component based on the initial luminance component denoising of the image to be processed using a first initial guiding image, specifically including:

[0167] The system is used to determine a first image standard deviation parameter and a second image standard deviation parameter based on an intermediate luminance component and a first intermediate guide image, wherein the intermediate luminance component is determined by normalizing an initial luminance component, and the first intermediate guide image is determined by normalizing a first initial guide image; it is used to determine a first variable based on the first image standard deviation parameter and a control edge preservation parameter, wherein the control edge preservation parameter is a non-negative constant less than 1; it is used to determine a first coefficient and a second coefficient based on the first image standard deviation parameter, the second image standard deviation parameter, the first variable, and a preset correction parameter; and it is used to determine a target luminance component based on the first coefficient, the second coefficient, the first initial guide image, and inverse normalization.

[0168] In some embodiments, the luminance component execution unit is further configured to determine a first sum based on a control edge preservation parameter and a first image standard deviation parameter; and to determine a first variable based on the first sum, the control edge preservation parameter, and the first image standard deviation parameter.

[0169] In some embodiments, the initial chromaticity components include a first initial chromaticity component and a second initial chromaticity component, and the execution module further includes a first chromaticity component execution unit and a second chromaticity execution unit.

[0170] The first chroma component execution unit is used to denoise the first initial chroma component of the image to be processed based on the second initial guide image, and to determine the denoised first target chroma component. Specifically, it includes: determining a third image standard deviation parameter and a fourth image standard deviation parameter based on the first intermediate chroma component and the second intermediate guide image, wherein the first intermediate chroma component is determined by normalizing the first initial chroma component, and the second intermediate guide image is determined by normalizing the second initial guide image; determining a second variable based on the third image standard deviation parameter and a control edge preservation parameter, wherein the control edge preservation parameter is a non-negative constant less than 1; determining a third coefficient and a fourth coefficient based on the third image standard deviation parameter, the fourth image standard deviation parameter, the second variable, and a preset correction parameter; and determining the first target chroma component based on the third coefficient, the fourth coefficient, the second initial guide image, and inverse normalization.

[0171] The second chroma component execution unit is used to denoise the second initial chroma component of the image to be processed based on the second initial guiding image, and to determine the denoised second target chroma component. Specifically, it includes: determining the fifth image standard deviation parameter and the sixth image standard deviation parameter based on the second intermediate chroma component and the second intermediate guiding image, wherein the second intermediate chroma component is determined by normalizing the second initial chroma component; determining the third variable based on the fifth image standard deviation parameter and the edge preservation parameter; determining the fifth coefficient and the sixth coefficient based on the fifth image standard deviation parameter, the sixth image standard deviation parameter, the third variable, and the preset correction parameter; and determining the second target chroma component based on the fifth coefficient, the sixth coefficient, the second initial guiding image, and inverse normalization.

[0172] In some embodiments, the execution module further includes a simplification unit, which may be disposed before the first chroma component execution unit and the second chroma execution unit, for downsampling the image to be processed based on a preset downsampling module.

[0173] In some embodiments, the acquisition module is further configured to acquire a captured image, wherein the captured image is an RGB format image with noise; and convert the captured image into an image to be processed.

[0174] In some embodiments, the output module is further configured to convert the first target image into a second target image after determining the first target image, wherein the second target image is an image in RGB format.

[0175] This application provides an image denoising processing apparatus, including an acquisition module, an execution module, and an output module. First, an image to be processed is determined, which is a noisy YUV format image. The initial luminance component of the image to be processed is filtered by mean under a first preset window to determine a first initial guiding image. The initial chrominance component of the image to be processed is filtered by mean under a first preset window to determine a second initial guiding image. Noise reduction of the initial luminance component of the image to be processed using the first initial guiding image allows for the determination of the denoised target luminance component. Noise reduction of the initial chrominance component of the image to be processed using the second initial guiding image or mean filtering allows for the determination of the denoised target chrominance component. A first target image is determined using the target luminance component and the target chrominance component. Through the above denoising processing, noise can be reduced and image quality improved while ensuring that the detailed edges of the image are fully preserved.

[0176] The computer device provided in this application embodiment includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program. The computer program is used to implement the above-described image noise reduction processing method. The implementation principle and technical effects are similar to those in the above-described method embodiments, and will not be repeated here.

[0177] This application also provides a computer storage medium on which a computer program is stored. The computer program is executed by a processor using the above-described image noise reduction method. Its implementation principle and technical effects are similar to those of the above-described method embodiments, and will not be repeated here.

[0178] The following paragraphs will compare and list the Chinese terms used in this application specification and their corresponding English terms to facilitate reading and understanding.

[0179] For ease of explanation, the above description has been provided in conjunction with specific embodiments. However, the discussion in some embodiments above is not intended to be exhaustive or to limit the embodiments to the specific forms disclosed above. Various modifications and variations can be obtained based on the above teachings. The selection and description of the above embodiments are for the purpose of better explaining the principles and practical applications, thereby enabling those skilled in the art to better utilize the embodiments and various different variations of the embodiments suitable for specific application considerations.

Claims

1. An image noise reduction processing method, characterized in that, include: The image to be processed is determined, wherein the image to be processed is a noisy YUV format image; Based on the first initial guiding image, the initial luminance component of the image to be processed is denoised, and the target luminance component after denoising is determined. The first initial guiding image is the mean filtered image of the initial luminance component of the image to be processed under the first preset window. The step of denoising the initial luminance component of the image to be processed based on the first initial guiding image, and determining the denoised target luminance component, includes: Based on the intermediate brightness component and the first intermediate guide image, the standard deviation parameters of the first image and the second image are determined, wherein the intermediate brightness component is determined by normalizing the initial brightness component, and the first intermediate guide image is determined by normalizing the first initial guide image. The initial luminance component and the first initial guiding image are normalized and calculated using the following formula: In the formula, For the initial luminance component, This is the first initial guide image. The row and column values ​​corresponding to the image. For the intermediate brightness component, This is the first intermediate guiding image. Indicates to The components are subjected to mean filtering. The image of the intermediate brightness component, the first intermediate guide image, the image after multiplying the intermediate brightness component and the first intermediate guide image, and the image after multiplying the first intermediate guide image and the first intermediate guide image are subjected to mean filtering under a second preset window to determine the standard deviation parameter of the first image and the standard deviation parameter of the second image. Based on the intermediate brightness component and the first intermediate guide image, the standard deviation parameters of the first image and the second image are determined and obtained by the following formula: In the formula, This is the element-wise multiplication operation of matrices. For the intermediate brightness component, This is the first intermediate guiding image. The image is the result of multiplying the intermediate brightness component and the first intermediate guide image. The image is the result of multiplying the first intermediate guiding image and the second intermediate guiding image. This indicates that the Y component is subjected to mean filtering. The row and column values ​​corresponding to the image. This is the image after mean filtering of the intermediate brightness component within the second preset window. The image is the first intermediate guiding image after mean filtering within the second preset window. The image obtained by multiplying the intermediate brightness component and the first intermediate guide image is the image after mean filtering within a second preset window. The image obtained by multiplying the first intermediate guiding image and the image obtained by multiplying the first intermediate guiding image, is the image after mean filtering under the second preset window. First image standard deviation parameter, The standard deviation parameter of the second image; Based on the first image standard deviation parameter and the control edge preservation degree parameter, a first variable is determined, wherein the control edge preservation degree parameter is a non-negative constant less than 1; The determination of the first variable based on the first image standard deviation parameter and the edge preservation control parameter includes: Based on the control edge preservation parameter and the first image standard deviation parameter, determine the first sum value; The first intermediate value of the points in the image is determined by controlling the edge preservation parameter and the first image standard deviation parameter. The first intermediate values ​​of each point in the image are added together to determine the first sum value. The first sum is obtained by the following formula: In the formula, As the first intermediate quantity, The standard deviation parameter of the first image. To control the edge preservation parameters, Let m and n be the first sum; m and n are the length and width of the image to be processed. This represents the first intermediate value corresponding to each point in the image; In the noise reduction process of the initial luminance component of the image to be processed, m and n are the length and width of the initial luminance component; The first variable is determined based on the first sum, the control edge preservation parameter, and the first image standard deviation parameter; The first variable is obtained by calculating the following formula: In the formula, As the first variable; Based on the first image standard deviation parameter, the second image standard deviation parameter, the first variable, and the preset correction parameter, the first coefficient and the second coefficient are determined; The first coefficient is obtained by the following formula: In the formula, As the first coefficient, For preset calibration parameters, This is an element-wise division operation between corresponding elements of a matrix. The second coefficient is obtained by calculating the following formula: In the formula, The second coefficient; Based on the first coefficient, the second coefficient, and the first initial guiding image, as well as inverse normalization, the target brightness component is determined; Based on the second initial guiding image or mean filtering, the initial chroma component of the image to be processed is denoised, and the denoised target chroma component is determined. The second initial guiding image is the mean filtering image of the initial chroma component of the image to be processed under the first preset window. A first target image is determined based on the target luminance component and the target chrominance component.

2. The image noise reduction processing method according to claim 1, characterized in that, The initial chromaticity component includes a first initial chromaticity component. Based on the second initial guiding image, the initial chromaticity component of the image to be processed is denoised to determine the denoised target chromaticity component, which includes based on the second initial guiding image, the first initial chromaticity component of the image to be processed is denoised to determine the denoised first target chromaticity component. The step of denoising the first initial chromaticity component of the image to be processed based on the second initial guiding image, and determining the denoised first target chromaticity component, includes: Based on the first intermediate chromaticity component and the second intermediate guide image, the standard deviation parameters of the third image and the fourth image are determined, wherein the first intermediate chromaticity component is determined by normalizing the first initial chromaticity component, and the second intermediate guide image is determined by normalizing the second initial guide image. Based on the third image standard deviation parameter and the control edge preservation parameter, a second variable is determined, wherein the control edge preservation parameter is a non-negative constant less than 1, including: determining a second intermediate value for the points corresponding to the image by using the control edge preservation parameter and the third image standard deviation parameter; summing the second intermediate values ​​corresponding to each point in the image to determine a second sum; and determining the second variable based on the second sum, the control edge preservation parameter, and the third image standard deviation parameter. Based on the third image standard deviation parameter, the fourth image standard deviation parameter, the second variable, and the preset correction parameter, the third coefficient and the fourth coefficient are determined; Based on the third coefficient, the fourth coefficient, and the second initial guiding image, as well as inverse normalization, the first target chromaticity component is determined.

3. The image noise reduction method according to claim 2, characterized in that, The initial chromaticity component further includes a second initial chromaticity component. Based on the second initial guiding image, the initial chromaticity component of the image to be processed is denoised to determine the denoised target chromaticity component. The method also includes based on the second initial guiding image, the second initial chromaticity component of the image to be processed is denoised to determine the denoised second target chromaticity component. The step of denoising the second initial chromaticity component of the image to be processed based on the second initial guiding image, and determining the denoised second target chromaticity component, includes: Based on the second intermediate chromaticity component and the second intermediate guide image, the standard deviation parameters of the fifth image and the sixth image are determined, wherein the second intermediate chromaticity component is determined by normalizing the second initial chromaticity component; Based on the fifth image standard deviation parameter and the edge preservation control parameter, a third variable is determined, including: The third intermediate value of the corresponding point in the image is determined by controlling the edge preservation parameter and the fifth image standard deviation parameter. The third intermediate value of each point in the image is added together to determine the third sum. The third variable is determined based on the third sum, the edge preservation parameter, and the fifth image standard deviation parameter. The fifth and sixth coefficients are determined based on the fifth image standard deviation parameter, the sixth image standard deviation parameter, the third variable, and the preset correction parameter; Based on the fifth coefficient, the sixth coefficient, and the second initial guiding image, as well as inverse normalization, the second target chromaticity component is determined.

4. The image noise reduction processing method according to claim 1, characterized in that, Before performing initial chroma component noise reduction processing on the image to be processed based on the second initial guiding image or mean filtering, the method further includes: The image to be processed is downsampled based on a preset downsampling module.

5. The image noise reduction method according to claim 1, characterized in that, Before determining the image to be processed, the process also includes: Acquire a captured image, wherein the captured image is an RGB format image with noise; Convert the captured image into the image to be processed; After determining the first target image, the process further includes: The first target image is converted into a second target image, wherein the second target image is an RGB format image.

6. An image noise reduction processing apparatus, applied to the image noise reduction processing method of claim 1, characterized in that, include: An acquisition module is used to determine the image to be processed, wherein the image to be processed is a noisy YUV format image; An execution module is used to denoise the initial luminance component of the image to be processed based on a first initial guiding image, and determine the denoised target luminance component; wherein, the first initial guiding image is the mean-filtered image of the initial luminance component of the image to be processed under a first preset window. The execution module is further configured to denoise the initial chroma component of the image to be processed based on the second initial guiding image or mean filtering, and determine the denoised target chroma component, wherein the second initial guiding image is the mean filtered image of the initial chroma component of the image to be processed under the first preset window. The output module is used to determine the first target image based on the target luminance component and the target chrominance component.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the image noise reduction processing method according to any one of claims 1 to 5.

8. A computer storage medium, characterized in that, The computer storage medium stores a computer program, which, when executed by a processor, causes the processor to perform the steps of the image noise reduction processing method according to any one of claims 1 to 5.

Images