Image noise reduction method, terminal device, and computer-readable storage medium

Abstract

This application discloses an image denoising method, a terminal device, and a computer-readable storage medium. The image denoising method includes: acquiring an original image; constructing a first denoising energy function, wherein the denoising energy function includes data terms and rule terms; initializing the first denoising energy function using the original image to obtain first initialization parameters; and iteratively calculating the first denoising energy function based on the first initialization parameters to obtain an image whose output of the first denoising energy function satisfies preset conditions, which is used as the final denoised image. The image denoising method of this application constructs the denoising energy function through rule terms, which effectively improves the denoising effect by preserving the edge information of the denoised image while reducing edge jaggedness.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to an image noise reduction method, a terminal device, and a computer-readable storage medium. Background Technology

[0002] Various interferences exist in the process of image acquisition, transmission, and reception. For example, the non-uniformity of sensitivity of sensitive elements during photoelectric conversion, quantization noise during digitization, errors during transmission, and human factors can all be affected by noise, which in turn affects the input and output links. Noise degrades image quality, making the image blurry or even obscuring image features, thus leading to a decrease in image quality. Not only does it reduce the resolution of the image, but it also destroys the original fine structure of the image, which brings difficulties to subsequent image region segmentation, analysis, and judgment.

[0003] The purpose of image denoising is to process images to obtain images that are more suitable for subsequent processing. Noise is extremely detrimental to subsequent image processing. Effectively suppressing noise, removing noise, and restoring the original image in the image preprocessing stage is an important aspect of image processing and plays a crucial role in image applications. Summary of the Invention

[0004] This application proposes an image noise reduction method, a terminal device, and a computer-readable storage medium.

[0005] One technical solution adopted in this application is to provide an image denoising method, the image denoising method comprising:

[0006] Obtain the original image;

[0007] Construct a first noise reduction energy function, wherein the noise reduction energy function includes a data term and a rule term;

[0008] The first noise reduction energy function is initialized using the original image to obtain the first initialization parameters;

[0009] Based on the first initialization parameters, the first noise reduction energy function is iteratively calculated to obtain the image when the output of the first noise reduction energy function meets the preset conditions, which is used as the final noise reduction image.

[0010] The construction of the first noise reduction energy function includes:

[0011] Construct rule terms, and combine rule terms and data terms according to the global noise reduction intensity to form the initial first noise reduction energy function.

[0012] The rule items include image pixel values, image edge direction noise reduction intensity control parameters, and weight values;

[0013] The weight value is determined by the ratio of the pixel value of the current image block to the pixel value of other image blocks in the original image.

[0014] The construction of the rule term, after combining the rule term and the data term according to the global noise reduction intensity to form the first noise reduction energy function, also includes:

[0015] Perform at least one Taylor expansion on the regular terms in the first noise reduction energy function to obtain the second noise reduction energy function.

[0016] Before initializing the first noise reduction energy function using the original image to obtain the first initialization parameters, the method further includes:

[0017] Map the first noise reduction energy function into a matrix function;

[0018] The matrix function includes data items in matrix form and rule items in matrix form.

[0019] The step of initializing the first noise reduction energy function using the original image to obtain the first initialization parameters includes:

[0020] The original image is input into the matrix function for initialization to obtain the first initialization parameters;

[0021] The initialization process, which involves inputting the original image into the matrix function, includes:

[0022] The data terms of the matrix function are initialized with the numerical values ​​of the original image, and the first initialization parameter of the regular terms in the matrix function is calculated.

[0023] The process of iteratively calculating the first denoising energy function based on the first initialization parameters to obtain an image where the output of the first denoising energy function satisfies a preset condition includes:

[0024] Differentiating the matrix function yields the denoised image function;

[0025] The first initialization parameter of the rule item is input into the denoised image function to obtain the first pixel parameter of the denoised image;

[0026] The first pixel parameter is input into the matrix function to calculate the second initialization parameter of the rule term;

[0027] The second initialization parameter of the rule item is input into the denoised image function to obtain the second pixel parameter of the denoised image;

[0028] Determine whether the difference between the first pixel parameter and the second pixel parameter meets a preset condition;

[0029] If so, the final denoised image is obtained based on the second pixel parameter;

[0030] If not, iteratively calculate the initialization parameters and pixel parameters of the rule item until the difference between adjacent pixel parameters satisfies the preset condition.

[0031] The preset condition is that the norm of the difference between adjacent pixel parameters is less than a preset threshold.

[0032] Another technical solution adopted in this application is to provide a terminal device, the terminal device including a processor and a memory connected to the processor, wherein the memory stores program instructions; the processor is used to execute the program instructions stored in the memory to implement the image noise reduction method as described above.

[0033] Another technical solution adopted in this application is to provide a computer-readable storage medium storing program instructions, which, when executed, implement the above-described image noise reduction method.

[0034] The beneficial effects of this application are as follows: The original image is acquired; a first denoising energy function is constructed, wherein the denoising energy function includes data terms and rule terms; the first denoising energy function is initialized using the original image to obtain first initialization parameters; the first denoising energy function is iteratively calculated based on the first initialization parameters to obtain an image whose output satisfies preset conditions, which is then used as the final denoised image. The image denoising method of this application constructs a denoising energy function through rule terms, which, while preserving the edge information of the denoised image, also reduces edge jaggedness, effectively improving the denoising effect. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a flowchart illustrating an embodiment of the image noise reduction method provided in this application;

[0037] Figure 2 yes Figure 1 A flowchart illustrating the sub-steps of S4 in the middle;

[0038] Figure 3 This is a schematic flowchart of another embodiment of the image noise reduction method provided in this application;

[0039] Figure 4 This is a schematic diagram of the structure of an embodiment of the terminal device provided in this application;

[0040] Figure 5 This is a schematic diagram of the structure of an embodiment of the computer storage medium provided in this application. Detailed Implementation

[0041] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0042] The reference to "embodiment" in this application means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0043] The steps in the embodiments of this application are not necessarily processed in the order described. The steps can be rearranged, deleted, or added as needed. The step descriptions in the embodiments of this application are only optional combinations of sequences and do not represent all possible combinations of steps in the embodiments of this application. The order of steps in the embodiments should not be considered as a limitation of this application.

[0044] The terms "first," "second," etc., used in this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, or apparatuses.

[0045] Furthermore, although the terms "first," "second," etc., are used repeatedly in this application to describe various data (or various applications, or various instructions, or various operations), these data (or applications, instructions, or operations) should not be limited by these terms. These terms are only used to distinguish one type of data (or application, instruction, or operation) from another type of data (or application, instruction, or operation).

[0046] Currently, the mainstream image denoising methods include spatial domain image denoising methods, frequency domain image denoising methods, and learning-based image denoising methods.

[0047] Image denoising methods in the spatial domain mainly include mean filtering, Gaussian filtering, median filtering, and other algorithms. These algorithms tend to lose a lot of useful image information while denoising. Slightly more complex bilateral filtering and nonlocal mean algorithms overcome some of the above shortcomings, but still have the disadvantage of poor texture preservation.

[0048] Frequency domain image denoising methods mainly include wavelet denoising and discrete cosine transform (DCT), which transform the image to the frequency domain and then perform denoising there. Compared with spatial domain image denoising methods, frequency domain image denoising methods are less effective at suppressing large noise levels.

[0049] Learning-based image denoising methods mainly include deep learning-based methods and dictionary-based learning-based methods. Learning-based image denoising methods train denoising algorithms using pre-collected samples, which can reduce noise while preserving image information to the maximum extent. However, these methods are computationally intensive, making them difficult to apply in real-world scenarios.

[0050] This application proposes an image denoising method to reduce edge coarseness in the denoised image. For details, please refer to [link to details]. Figure 1 , Figure 1 This is a schematic flowchart of an embodiment of the image noise reduction method provided in this application.

[0051] like Figure 1 As shown, the image denoising method of this application embodiment may specifically include the following steps:

[0052] S1, Obtain the original image.

[0053] In one embodiment, the camera captures the original image and transmits it to an image noise reduction terminal for image noise reduction processing.

[0054] S2, construct the first noise reduction energy function.

[0055] The steps for constructing the first noise reduction energy function may include:

[0056] Construct rule terms, and combine rule terms and data terms according to the global noise reduction intensity to form the initial first noise reduction energy function.

[0057] In one embodiment, the first noise reduction energy function constructed by the image noise reduction terminal satisfies:

[0058] E(X) = g(X) + λf(X)

[0059] Where g(X) is the data term, λ is the global denoising intensity, f(X) is the rule term, and X is the denoised image.

[0060] The rule items include image pixel values, image edge direction noise reduction intensity control parameters, and weight values.

[0061] In one embodiment, the rule terms constructed by the image denoising terminal satisfy:

[0062]

[0063] Where i represents the position information of a pixel in the image, and x i S represents the pixel value at position i. i Let σ be a range centered at position i. i These are the noise reduction control parameters for the edge direction.

[0064] Optionally, S i It can be 3x3, 5x5, 7x7, etc.

[0065] Optionally, σ i Different values ​​can be set based on different pixel values.

[0066] The weight value is determined by the ratio of the pixel value of the current image block to the pixel value of other image blocks in the original image.

[0067] In one embodiment, the image noise reduction terminal calculates weight values ​​using image block matching and normalization. The weight values ​​can be calculated as follows:

[0068]

[0069] Among them, w j is x i Relative to x i The weight value.

[0070] This application constructs rule terms and uses the weight values ​​within these rule terms to constrain the direction of filtering, ensuring that filtering occurs along the edge tangent during the noise reduction process. This preserves edge information while also resolving jagged edges. Furthermore, exponential operations are introduced to reduce the weights of high-frequency noise reduction, further protecting information from large edges.

[0071] In one embodiment, the first noise reduction energy function constructed by the image noise reduction terminal can be:

[0072]

[0073] in, For data items, y i Let i be the pixel value of the original image at position i. λ represents the rule term, and λ represents the global noise reduction intensity.

[0074] The image denoising method proposed in this application optimizes the first denoising energy function, that is, it obtains the image that minimizes the first denoising energy function.

[0075] X = argmin X E(X)

[0076] After establishing the first noise reduction energy function, the regular terms in the first noise reduction energy function are subjected to at least one Taylor expansion to obtain the second noise reduction energy function.

[0077] In one embodiment, optimizing the established first denoising energy function by the image denoising terminal is a non-convex optimization problem. To address this problem, a first Taylor expansion is performed on the data terms in the first denoising energy function to obtain a preliminary second denoising energy function:

[0078]

[0079] in, f′ σ (x) is f σ The derivative of (x).

[0080] Solving the initial second denoising energy function obtained by the image denoising terminal is still a non-convex optimization problem. The image denoising terminal further performs a Taylor expansion on the initial second denoising energy function to obtain the second denoising energy function:

[0081]

[0082] in,

[0083] The image denoising terminal performs at least one Taylor expansion on the first denoising energy function to obtain a second denoising energy function that is an approximate convex function of the first denoising energy function, which facilitates the solution of the final denoised image.

[0084] S3. Initialize the first noise reduction energy function using the original image to obtain initial parameters.

[0085] Before S3, it also includes mapping the first noise reduction energy function into a matrix function, where the matrix function includes data terms in matrix form and rule terms in matrix form.

[0086] In one embodiment, the image denoising terminal can map the approximate expansion function of the first denoising energy function, i.e., the second denoising energy function, into a matrix function, i.e.:

[0087]

[0088] Where L and H are diagonal matrices, and the diagonal elements of L are... The diagonal elements of H are

[0089] Where D = FM, is the product of diagonal matrix F and diagonal matrix M, and the diagonal elements of diagonal matrix F are... The value of the i-th row and i-th column of the diagonal matrix M is 1, and in j∈S i The value in the i-th row and j-th column within the range is w j All other elements in the i-th row are 0.

[0090] The original image is input into a matrix function for initialization to obtain the first initialization parameters.

[0091] This step also includes: initializing the data terms of the original image's numerical values ​​into the matrix function, and calculating the first initialization parameter of the regular terms in the matrix function.

[0092] In one embodiment, the image noise reduction terminal will X 0 Initialize it as Y, substitute it into the matrix function E(X) and calculate the first initialization parameter L of the rule term. 0 D 0 and H 0 .

[0093] S4. Based on the first initialization parameters, iteratively calculate the first denoising energy function to obtain the image when the output of the first denoising energy function meets the preset conditions, which is used as the final denoised image.

[0094] For detailed steps on iteratively calculating the first noise reduction energy function, please refer to [link / reference needed]. Figure 2 , Figure 2 yes Figure 1 A flowchart illustrating the sub-steps of S4.

[0095] like Figure 2 As shown, the specific steps for iteratively calculating the first noise reduction energy function may include:

[0096] S41, differentiate the matrix function to obtain the denoised image function.

[0097] In one embodiment, the derivative of the matrix function E(X) is taken to obtain the denoised image function:

[0098] X k+1 =(I+2λL) k ) -1 (Y-λD t,k 1+λH t,k 1)

[0099] S42, input the first initialization parameter of the rule item into the denoised image function to obtain the first pixel parameter of the denoised image.

[0100] In one embodiment, the image noise reduction terminal initializes the first parameter L of the rule term. 0 D 0 and H 0 The input is fed into the image denoising function to calculate and solve for the first pixel parameter X. 1 .

[0101] In another embodiment, the image noise reduction terminal initializes the first parameter L of the rule term. k-1 D k-1 and H k-1 The input is fed into the image denoising function to calculate and solve for the first pixel parameter X. k .

[0102] Where k = 2, 3, 4...

[0103] S43, input the first pixel parameter into the matrix function to calculate the second initialization parameter of the rule term.

[0104] In one embodiment, the image noise reduction terminal will use the first pixel parameter X 1 The input is given to the matrix function E(X), and the second initialization parameter L of the rule term is calculated and solved. 1 D 1 and H 1 .

[0105] In another embodiment, the image noise reduction terminal will use the first pixel parameter X k The input is given to the matrix function E(X), and the second initialization parameter L of the rule term is calculated and solved. k D k and H k .

[0106] Where k = 1, 2, 3, ...

[0107] S44, input the second initialization parameter of the rule term into the denoised image function to obtain the second pixel parameter of the denoised image.

[0108] In one embodiment, the image noise reduction terminal initializes the second initialization parameter L of the rule term. 1 D 1 and H 1 The input is fed into the image denoising function to calculate and solve for the second pixel parameter X. 2 .

[0109] In another embodiment, the image noise reduction terminal initializes the second initialization parameter L of the rule term. k D k and Hk The input is fed into the image denoising function to calculate and solve for the second pixel parameter X. k+1 .

[0110] S45, determine whether the difference between the first pixel parameter and the second pixel parameter meets the preset condition. If yes, jump to S46; otherwise, jump to S47.

[0111] The preset condition is that the norm of the difference between adjacent pixel parameters is less than a preset threshold.

[0112] In one embodiment, if the first pixel parameter X 1 With the second pixel parameter X 2 If the norm of the difference is less than a preset threshold, proceed to S46, i.e., the following condition is met:

[0113]

[0114] Where ε is a preset threshold.

[0115] If the first pixel parameter X 1 With the second pixel parameter X 2 If the norm of the difference is greater than or equal to the preset threshold, proceed to S47.

[0116] In another embodiment, if the first pixel parameter X k With the second pixel parameter X k+1 If the norm is less than a preset threshold, proceed to S46, i.e., the following conditions are met:

[0117]

[0118] Where ε is a preset threshold, and k = 1, 2, 3, ...

[0119] If the first pixel parameter X k With the second pixel parameter X k+1 If the norm is greater than or equal to the preset threshold, proceed to S47.

[0120] S46, the final denoised image is obtained based on the second pixel parameter.

[0121] In one embodiment, the image noise reduction terminal determines the image noise reduction based on the second pixel parameter X. k+1 To obtain the final denoised image, i.e., X = X k+1 Where k = 1, 2, 3, ...

[0122] S47, iteratively calculate the initialization parameters and pixel parameters of the rule items until the difference between adjacent pixel parameters meets the preset conditions.

[0123] In one embodiment, the image denoising terminal calculates the initialization parameters of the rule terms, and sequentially calculates L. 2 D2 H 2 The above initialization parameter L 2 D 2 H 2 Substitute into the denoising image function to obtain X 3 .

[0124] If not satisfied:

[0125]

[0126] Then iteratively calculate the initialization parameters and pixel parameters of the rule terms until the following conditions are met:

[0127]

[0128] Where k = 1, 2, 3, ...

[0129] Please see Figure 3 , Figure 3 This is a flowchart illustrating another embodiment of the image noise reduction method provided in this application.

[0130] like Figure 3 As shown, another embodiment of the image denoising method provided in this application may include the following steps:

[0131] S100, propose rule items.

[0132] In one embodiment, the rule terms constructed by the image denoising terminal satisfy:

[0133]

[0134] Where i represents the position information of a pixel in the image, and x i S represents the pixel value at position i. i Let σ be a range centered at position i. i These are the noise reduction control parameters for the edge direction.

[0135] Optionally, S i It can be 3x3, 5x5, 7x7, etc.

[0136] Optionally, σ i Different values ​​can be set based on different pixel values.

[0137] The weight value is determined by the ratio of the pixel value of the current image block to the pixel value of other image blocks in the original image.

[0138] In one embodiment, the image noise reduction terminal calculates weight values ​​using image block matching and normalization. The weight values ​​can be calculated as follows:

[0139]

[0140] Among them, w j is x i Relative to x i The weight value.

[0141] This application constructs rule terms and uses the weight values ​​within these rule terms to constrain the direction of filtering, ensuring that filtering occurs along the edge tangent during the noise reduction process. This preserves edge information while also resolving jagged edges. Furthermore, exponential operations are introduced to reduce the weights of high-frequency noise reduction, further protecting information from large edges.

[0142] S200, establish a preliminary noise reduction and optimization energy function.

[0143] In one embodiment, the noise reduction optimization energy function constructed by the image noise reduction terminal can be:

[0144]

[0145] in, For data items, y i Let i be the pixel value of the original image at position i. λ represents the rule term, and λ represents the global noise reduction intensity.

[0146] The image denoising method proposed in this application optimizes the aforementioned denoising optimization energy function, that is, it obtains the image that minimizes the initial denoising optimization energy function.

[0147] X = argmin X E(X)

[0148] The optimization of the established denoising optimization energy function by the image denoising terminal is a non-convex optimization problem. To address this problem, a first Taylor expansion is performed on the data terms in the denoising optimization energy function, yielding the preliminary denoising optimization energy function as follows:

[0149]

[0150] in, f′ σ (x) is f σ The derivative of (x).

[0151] S300, establish the final noise reduction optimization energy function.

[0152] Solving the initial denoising optimization energy function obtained by the image denoising terminal is still a non-convex optimization problem. The image denoising terminal further performs a Taylor expansion on the initial denoising optimization energy function to obtain the final denoising optimization energy function:

[0153]

[0154] in,

[0155] The image denoising terminal maps the final denoising optimization energy function into a matrix function, namely:

[0156]

[0157] Where L and H are diagonal matrices, and the diagonal elements of L are... The diagonal elements of H are

[0158] Where D = FM, is the product of diagonal matrix F and diagonal matrix M, and the diagonal elements of diagonal matrix F are... The value of the i-th row and i-th column of the diagonal matrix M is 1, and in j∈S i The value in the i-th row and j-th column within the range is w j All other elements in the i-th row are 0.

[0159] S400, iterative optimization.

[0160] The image denoising terminal differentiates the matrix function to obtain the denoised image function:

[0161] X k+1 =(I+2λL) k ) -1 (Y-λD t,k 1+λH t,k 1)

[0162] Image noise reduction terminal will X 0 Initialize Y and substitute it into the matrix function E(X) to find L. 0 D 0 and H 0 Then L 0 D 0 and H 0 Substituting into the image denoising function, we obtain X. 1 ;

[0163] Image noise reduction terminal will X 1 Substitute the matrix function E(X) to find L. 1 D 1 and H 1 Then L 1 D 1 and H 1 Substituting into the image denoising function, we obtain X. 2 ;

[0164] Iterate continuously until the following conditions are met:

[0165]

[0166] Where ε is a preset threshold, and k = 1, 2, 3, ...

[0167] X = X k+1 This is the final denoised image.

[0168] Please continue reading Figure 4 , Figure 4 This is a schematic diagram of the structure of an embodiment of the terminal device provided in this application. The terminal device 500 of this application embodiment includes a processor 51 and a memory 52.

[0169] The processor 51 and the memory 52 are connected to the bus. The memory 52 stores program data, and the processor 51 is used to execute the program data to implement the image noise reduction method described in the above embodiments.

[0170] In this embodiment, processor 51 can also be referred to as a CPU (Central Processing Unit). Processor 51 may be an integrated circuit chip with signal processing capabilities. Processor 51 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor, or processor 51 can be any conventional processor.

[0171] This application also provides a computer storage medium; please refer to the following: Figure 5 , Figure 5 This is a schematic diagram of a computer storage medium according to an embodiment of the present application. The computer storage medium 600 stores program data 61, which is used to implement the image noise reduction method of the above embodiment when executed by the processor.

[0172] When the embodiments of this application are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0173] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. An image denoising method, characterized in that, The image noise reduction method includes: Obtain the original image; Constructing a first noise reduction energy function, wherein the first noise reduction energy function includes data terms and rule terms; constructing the first noise reduction energy function includes: Construct rule terms, and combine rule terms and data terms according to the global noise reduction intensity to form the initial first noise reduction energy function; The first noise reduction energy function satisfies the following relationship: in, For data items, y i x is the pixel value of the original image at position i. i Let be the pixel value of the denoised image at position i. Here, λ represents the rule term, λ represents the global noise reduction intensity, and S represents the standard term. i w is a range centered at position i. j σ is the weight value. i These are the noise reduction control parameters for the image edge direction; The first noise reduction energy function is initialized using the original image to obtain the first initialization parameters; Based on the first initialization parameters, the first noise reduction energy function is iteratively calculated to obtain the image when the output of the first noise reduction energy function meets the preset conditions, which is used as the final noise reduction image; The construction of the rule term, after combining the rule term and the data term according to the global noise reduction intensity to form the first noise reduction energy function, also includes: Perform at least one Taylor expansion on the regular terms in the first noise reduction energy function to obtain the second noise reduction energy function.

2. The image denoising method according to claim 1, characterized in that, The rule items include image pixel values, image edge direction noise reduction intensity control parameters, and weight values; The weight value is determined by the ratio of the pixel value of the current image block to the pixel value of other image blocks in the original image.

3. The image denoising method according to claim 1, characterized in that, Before initializing the first noise reduction energy function using the original image to obtain the first initialization parameters, the method further includes: Map the first noise reduction energy function into a matrix function; The matrix function includes data items in matrix form and rule items in matrix form.

4. The image denoising method according to claim 3, characterized in that, The initialization of the first noise reduction energy function using the original image to obtain the first initialization parameters includes: The original image is input into the matrix function for initialization to obtain the first initialization parameters; The initialization process, which involves inputting the original image into the matrix function, includes: The data terms of the matrix function are initialized with the numerical values ​​of the original image, and the first initialization parameter of the regular terms in the matrix function is calculated.

5. The image denoising method according to claim 4, characterized in that, Based on the first initialization parameters, the first noise reduction energy function is iteratively calculated to obtain an image when the output of the first noise reduction energy function satisfies a preset condition, including: Differentiating the matrix function yields the denoised image function; The first initialization parameter of the rule item is input into the denoised image function to obtain the first pixel parameter of the denoised image; The first pixel parameter is input into the matrix function to calculate the second initialization parameter of the rule term; The second initialization parameter of the rule item is input into the denoised image function to obtain the second pixel parameter of the denoised image; Determine whether the difference between the first pixel parameter and the second pixel parameter meets a preset condition; If so, the final denoised image is obtained based on the second pixel parameter; If not, iteratively calculate the initialization parameters and pixel parameters of the rule item until the difference between adjacent pixel parameters satisfies the preset condition.

6. The image denoising method according to claim 5, characterized in that, The preset condition is that the norm of the difference between adjacent pixel parameters is less than a preset threshold.

7. A terminal device, characterized in that, The terminal device includes a processor and a memory connected to the processor, wherein... The memory stores program instructions; The processor is used to execute program instructions stored in the memory to implement the image noise reduction method as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that, The storage medium stores program instructions, which, when executed, implement the image noise reduction method as described in any one of claims 1 to 6.

Images