Image processing method and device using parameterized Thiele continued fraction interpolation
Through the parameterized Thiele continuous fractional interpolation method, the color image is subject to region enhancement and noise reduction processing, which solves the problem of low color image quality, realizes high-resolution and high-quality image reconstruction, and protects the edges and textures of the image information.
Patent Information
- Application Number
- CN202011015740.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-09-24
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2040-09-24
AI Technical Summary
The prior art is difficult to effectively process color images, especially in situations where image quality is low, insufficient contrast, color distortion and unclear edge information, and commonly used image enhancement methods are sensitive to noise or cannot effectively process literary and scientific details of smooth areas.
The parameterized Thiele continuous fraction interpolation method is used to enhance the region, denoising and interpolation reconstruction of color images. Through frequency domain spatial conversion, filtering and parameterized Thiele continuous fraction interpolation formula, the high quality and high definition of the image are achieved. . The specific steps include converting the image into HSV color space, filtering and denoising the channel components, dividing the region based on image gradient changes and performing parameterized interpolation until the preset high resolution conditions are met.
The brightness and saturation uniformity of the image is enhanced, edge and texture information are protected, high-resolution and high-quality images are obtained, avoiding blurring caused by noise, and improving the overall quality of the image.
Smart Images

Figure CN112132749B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to an image processing method and device using parameterized Thiele continued fraction interpolation. Background Art
[0002] During the acquisition process, images may be affected by factors such as the dynamic range of the imaging device and the intensity of ambient light, resulting in low contrast, unclear image information, color distortion, and insufficient clarity of the target's outline or boundary information, which makes it difficult for human visual observation and machine analysis and processing. Therefore, image enhancement is required. Image enhancement refers to a processing method that highlights certain information of an image according to specific needs, while weakening or removing certain unnecessary information. It is the most basic means of image processing and is often a preprocessing process for various image analysis and processing.
[0003] More than 90% of the information humans obtain comes from vision, and the real natural scene is colorful and varied. Therefore, the information contained in color images is far greater than that of other types of images. However, since there are many problems to be solved in true color image processing, its processing difficulty is far greater than that of other types of images. At present, more than 90% of the literature and books on image processing are about grayscale image processing. At the same time, the commonly used image enhancement methods can be mainly divided into sharpening enhancement, fuzzy set enhancement, multi-scale geometric enhancement, and enhancement methods based on differential operators. According to different image imaging characteristics, various methods have both advantages and disadvantages. For example, sharpening enhancement can enhance the small details and edge parts of the image, but it is sensitive to noise. Multi-scale geometric enhancement can enhance the image in a targeted manner from multiple scales, but it is easy to produce a "ringing" effect; enhancement based on differential operators can effectively process corner points, tubular structures, etc., but it cannot solve the problem of texture details inside the smooth area of the detection image, resulting in low image quality. Summary of the invention
[0004] In view of the problems existing in the above-mentioned prior art, the present invention provides an image processing method using parameterized Thiele continued fraction interpolation, which performs image enhancement, noise reduction and interpolation reconstruction on color images in different regions, thereby achieving high quality and high definition of the processed images.
[0005] The image processing method using parameterized Thiele continued fraction interpolation of the present invention comprises the following steps:
[0006] Convert the image to be processed into a color space and extract a corresponding channel image, wherein the channel image at least includes a channel representing the brightness of the image;
[0007] Perform frequency domain space conversion on different channels, and use matched filtering method based on the characteristics of different channels to suppress the low-frequency components in the channel spectrum and improve the corresponding channel components;
[0008] Based on different channels with component enhancement, the smooth and non-smooth areas of the channel image are filtered to different degrees through filters to perform a denoising process based on edge protection;
[0009] Based on images from different channels, high-resolution images are obtained by performing image super-resolution reconstruction using the parameterized Thiele continued fraction interpolation method.
[0010] Perform inverse color space conversion to restore the original color space to obtain high-resolution and high-quality images.
[0011] As a further optimization of the above solution, the formula of the parameterized Thiele continued fraction interpolation adopts a binary Thiele continued fraction interpolation with a single parameter.
[0012] As a further optimization of the above scheme, the binary Thiele continued fraction interpolation with a single parameter is:
[0013]
[0014] in,
[0015]
[0016] For j = 0, 1, ..., n; p = 1, 2, ..., m; i = p, p + 1, ..., m,
[0017]
[0018] a k,0 ,a k,1 ,...,a k,l , a k,l+1 ,a k,l+2 ,...,a k,n For this formula The calculated value after the introduction of parameter λ.
[0019] For p = 0, 1, ..., k-1, k+1, ..., m; q = 1, 2, ..., n; j = q, q+1, ..., n,
[0020]
[0021] As a further optimization of the above scheme, a high-resolution image is obtained by performing image super-resolution reconstruction using a parameterized Thiele continued fraction interpolation method, and different regions are divided based on image gradient changes. Interpolation is performed using a parameterized Thiele continued fraction interpolation formula, and the interpolation formula for each region optimizes the interpolated image based on the adjustment of variable parameters.
[0022] As a further optimization of the above scheme, the image super-resolution reconstruction using the parameterized Thiele continued fraction interpolation method to obtain a high-resolution image also includes:
[0023] Performing image quality calculation on the image obtained by image super-resolution reconstruction based on preset evaluation indicators;
[0024] If the image quality data does not meet the preset high-resolution image condition, the image super-resolution reconstruction process is iterated by changing the variable parameters in the parameterized Thiele continued fraction interpolation until an image meeting the preset high-resolution image condition is obtained.
[0025] As a further optimization of the above solution, the color space of the image to be processed is converted and the corresponding channel image is extracted, specifically, the image is converted from RGB to HSV color space.
[0026] As a further optimization of the above scheme, the frequency domain space conversion of different channels is performed, and a matched filtering method is used based on the characteristics of different channels to suppress the low-frequency components in the channel spectrum and improve the corresponding channel components, including:
[0027] Sliding the S channel and V channel images based on a preset sliding window, and performing homomorphic filtering on the image area within each sliding window, wherein the homomorphic filtering cutoff frequency of each image area matches the image area spectrum;
[0028] The images after homomorphic filtering in all sliding window image regions are merged to obtain an S channel image with enhanced saturation and a V channel image with enhanced brightness and contrast.
[0029] As a further optimization of the above scheme, the smooth area and the non-smooth area of the channel image are filtered to different degrees by the filter to perform a denoising process based on edge protection, including:
[0030] Filtering is performed based on a filter model, which is:
[0031]
[0032] The parameter a of the model p′ ,b p′ Satisfy the objective optimization function E:
[0033]
[0034]
[0035]
[0036] Among them, G is the guide map, X is the image before filtering, G and X are allowed to be the same, is a window centered at p' and with a radius of ζ1, is the filtered image, a p′ and b p′ yes Two constants in the window, is the edge perception weight, which is the local variance ratio of the pixels in the 3x3 window and all pixels in the (2ζ1+1)x(2ζ1+1) window, χ(p′)=σ G,1 (p′)σ G,ζ1 (p′),σ G,1 (p′) and σ G,ζ1 (p′) respectively represents the image G in the window Ω1(p′) and Standard deviation, ε=(0.001*L) 2 , L is the dynamic range of image X; k is a constant; var(x(p))-x(p') represents the difference between the variance of x(p) and x(p'); var(x(p))-min(x(p)) represents the difference between the variance of x(p) and the minimum value of x(p).
[0037] The present invention provides an image processing device using parameterized Thiele continued fraction interpolation, comprising:
[0038] An image channel image extraction unit, used for converting the image to be processed into a color space and extracting a corresponding channel image, wherein the channel image at least includes a channel representing the brightness of the image;
[0039] The channel image enhancement unit is used to perform frequency domain space conversion on different channels, and adopts a matched filtering method based on the characteristics of different channels to suppress the low-frequency components in the channel spectrum and enhance the corresponding channel components;
[0040] A denoising and edge protection unit is used to filter the smooth area and the non-smooth area of the channel image to different degrees based on different channels with component enhancement, and perform a denoising process based on edge protection;
[0041] A high-resolution channel image acquisition unit is used to acquire a high-resolution image by performing image super-resolution reconstruction based on different channel images through a parameterized Thiele continued fraction interpolation method;
[0042] The channel image fusion unit is used to perform inverse conversion of the color space to restore it to the original color space to obtain high-resolution and high-quality images.
[0043] Based on the above-mentioned image processing method using parameterized Thiele continued fraction interpolation, the present invention further provides an electronic device, characterized in that the electronic device comprises:
[0044] A memory for storing executable instructions;
[0045] A processor, configured to implement the image processing method using parameterized Thiele continued fraction interpolation as described in any one of claims 1 to 8 when running the executable instructions stored in the memory.
[0046] The image processing method using parameterized Thiele continued fraction interpolation of the present invention has the following beneficial effects:
[0047] 1. The present invention enhances each channel image in the HSV channel of the color image separately to achieve brightness enhancement and saturation enhancement, thereby improving the quality of the image after processing. At the same time, in the enhancement process, sliding is performed based on a preset sliding window, and homomorphic filtering is performed on the image area in each sliding window. The homomorphic filtering cutoff frequency of each image area matches the image area spectrum, avoiding the use of the same filter parameters to filter all areas, resulting in the original dark area and the bright area in the image being enhanced to the same degree of brightness, resulting in unreasonable brightness enhancement of the bright area, and achieving uniform brightness of the image after brightness enhancement.
[0048] 2. In the process of denoising after image enhancement, the present invention adopts different degrees of filtering on the smooth area and non-smooth area of the channel image, and performs a denoising process based on edge protection, thereby realizing the protection of the edge area and texture area of the image while denoising, and avoiding blurring of these non-smooth areas caused by filter denoising.
[0049] 3. The present invention adopts binary single-parameter Thiele continued fraction interpolation in the process of image interpolation and reconstruction. Based on the characteristics of the constructed interpolation function, multiple interpolation effects can be achieved based on multiple values of variable parameters, and different interpolations in different regions can be achieved while interpolation debugging and correction within the region to achieve the best interpolation and reconstruction effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] Figure 1 It is an overall flow chart of the image processing method using parameterized Thiele continued fraction interpolation of the present invention;
[0051] Figure 2 yes Figure 1 A flowchart of a method for improving channel components of each channel image by filtering;
[0052] Figure 3 for Figure 1 A flowchart of a method for obtaining a high-resolution image by interpolating images of each channel;
[0053] Figure 4 It is a structural block diagram of an image processing device applying parameterized Thiele continued fraction interpolation according to the present invention. DETAILED DESCRIPTION
[0054] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present invention. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present invention.
[0055] In life, we often encounter low-quality images. Due to poor illumination intensity when collecting images, the image brightness is weak, and most of the content in the image is difficult to recognize, resulting in the inability to obtain important content in the image, affecting the recognition and content extraction of the image. The image processing method using parameterized Thiele continued fraction interpolation provided in the embodiment of the present invention is applied to the preprocessing or enhancement processing of color images after collection, so as to reduce the impact of low-quality images on us, improve the image brightness and color saturation, reduce the interference of external factors on the image content, and flexibly amplify the image.
[0056] The image processing method using parameterized Thiele continued fraction interpolation according to the embodiment of the present invention comprises the following steps:
[0057] Convert the image to be processed into a color space and extract the corresponding channel image, wherein the channel image at least includes a channel representing the brightness of the image; in this embodiment, specifically, convert the image from RGB to HSV color space, and extract the H channel, S channel and V channel images therein; after processing the three channel images respectively, perform color space conversion and restore them to the original color space, that is, RGB color space;
[0058] Perform frequency domain space conversion on different channels, and use matched filtering method based on the characteristics of different channels to suppress the low-frequency components in the channel spectrum and improve the corresponding channel components;
[0059] Specifically, considering that the purpose is to improve the brightness contrast and image saturation of the image, the S channel and V channel images are slid based on a preset sliding window, and homomorphic filtering is performed on the image area within each sliding window, and the homomorphic filtering cutoff frequency of each image area matches the image area spectrum; wherein, in the homomorphic filtering process, the image is divided into multiple areas by sliding the preset sliding window, and the homomorphic filters are used to filter them respectively, and the cutoff frequency of each filter is set differently based on different areas, which can avoid the same degree of brightness improvement of dark areas and bright areas in the original image caused by filtering all areas with the same filter parameters, resulting in unreasonable brightness improvement of the bright areas, and the image after brightness improvement is uniformly bright and dark, and at the same time, the entire area is divided in the form of a sliding window, and there is no restriction on the original size of the image, thereby ensuring the flexibility of area division;
[0060] The homomorphically filtered images of all sliding window image regions are merged to obtain an S channel image with enhanced saturation and a V channel image with enhanced brightness and contrast;
[0061] In order to remove the noise brought by the above image enhancement process, based on the different channels with enhanced components, the smooth area and non-smooth area of the channel image are filtered to different degrees through filters, and a denoising process based on edge protection is performed;
[0062] Specifically, filtering is performed based on a filter model, and the model is:
[0063]
[0064] in, is the filtered image, G is the guide graph, and the model is There is a local linear relationship between G and G in the window centered on pixel p'. When there is an edge in the guidance graph, there is an edge in the output image. Specifically, if pixel p' is on the edge, then a p′ The expected value of is 1. If the pixel p' is in the flat area, then a p′ The expected value of is 0, making the smooth region smooth.
[0065] In order to solve the parameter a p′ ,b p′ , assuming X is the image before filtering, The goal is to minimize the difference between and X, then the linear parameter a of the model p′ ,b p′ Satisfy the objective optimization function E:
[0066]
[0067]
[0068]
[0069] Among them, in the guide map, G is the guide map, X is the image before filtering, G and X are allowed to be the same. In this embodiment, G and X are the same, so as to protect the edge information during the filtering process. is a window with p' as the center and ζ1 as the radius, is the filtered image, a p′ and b p′ yes Two constants in the window, is the edge perception weight, which is the local variance ratio of the pixels in the 3x3 window and all pixels in the (2ζ1+1)x(2ζ1+1) window, χ(p′)=σ G,1 (p′)σ G,ζ1 (p′),σ G,1 (p′) and σ G,ζ1 (p′) respectively represents the image G in the window Ω1(p′) and The standard deviation of , ε is a very small constant, its value is (0.001*L) 2 , L is the dynamic range of image X; k is a constant, usually 4; var(x(p))-x(p') represents the difference between the variance of x(p) and x(p'); var(x(p))-min(x(p)) represents the difference between the variance of x(p) and the minimum value of x(p).
[0070] In the above filter model, the parameter δ in the objective optimization function E is p′ The size of is related to the difference between the variance of the regional pixels and all the pixels, so that the filtering effect of the entire filter model is related to the difference between the variance of the window area pixels and all the pixels in the image. Different filtering denoising processes are implemented for smooth areas and non-smooth areas (edge or texture areas), effectively avoiding blurring of non-smooth areas of the image after denoising.
[0071] Based on images of different channels, a parameterized Thiele continued fraction interpolation method is used to perform image super-resolution reconstruction to obtain a high-resolution image. Specifically, the formula of the parameterized Thiele continued fraction interpolation adopts a binary Thiele continued fraction interpolation with a single parameter, and the method for constructing the binary Thiele continued fraction interpolation with a single parameter is:
[0072] For a given two-dimensional distinct point set Π m,n {(x i ,y j )|i=0,1,K,m;j=0,1,K,n}, assuming Let f(x,y) be a real function defined on D, and
[0073] f(x i ,y j ) = f i,j ,i=0,1,K,m; j=0,1,K,n, (1)
[0074] It is easy to construct a binary Thiele continued fraction interpolation:
[0075]
[0076] Where i = 0, 1, ..., m; j = 0, 1, ..., n, And the denominator in the formula is the binary contrast quotient.
[0077] The parameter λ is introduced into the above formula to construct a binary Thiele continued fraction interpolation with a single parameter. The specific steps include:
[0078] Step 1: Initialize, let
[0079] Step 2: For j = 0, 1, ..., n; p = 1, 2, ..., m; i = p, p + 1, ..., m,
[0080] make
[0081]
[0082] Step 3: For p = 0, 1, ..., k-1, k+1, ..., m; q = 1, 2, ..., n; j = q, q+1, ..., n,
[0083] make
[0084]
[0085] Step 4: For formula (4), Introduce parameter λ and calculate according to formula (4-1)-(4-4):
[0086] right
[0087] When j=1,…,k+1, for q=j,j+1,…,n;
[0088]
[0089] For j=k+1,k+2,…,n,
[0090]
[0091] When j=k+2,k+3,…,n, for q=j,j+1,…,n;
[0092]
[0093] The above calculated The coefficients are uniformly recorded as:
[0094]
[0095] Step 5: Using the elements in equation (5) and the elements in equation (6), construct a single-parameter Thiele continued fraction interpolation about y:
[0096]
[0097] Step 6: Order
[0098]
[0099] but That is, it is a binary Thiele continued fraction interpolation function with a single parameter on a given set of points, A1(y), A2(y), ..., A m (y) is the binary contrast quotient of the interpolation function.
[0100] It can be proved that the continued fraction interpolation function constructed by this algorithm satisfies the theorem:
[0101]
[0102] Proof: For any node (x i ,y j ,f i,j ), when i=0,1,…,k-1,k+1,…,m,
[0103] Easy to see
[0104]
[0105] When i=k,
[0106]
[0107] Regardless of 0≤j<l,l=j,n≥j>l,
[0108]
[0109] From Theorem 4, it is easy to see that
[0110] The proof is complete.
[0111] Theorem 4: is a mutually different rectangular grid on the rectangular domain D, f(x,y) is a real function defined on the rectangular domain D, then the binary parameterized Thiele continued fraction interpolation satisfies
[0112]
[0113] In the process of interpolation based on the above-mentioned continued fraction interpolation, different regions are divided based on the image gradient change, and the pixel regions in the gradient change gentle region and the non-gentle region are interpolated using the parameterized Thiele continued fraction interpolation formula respectively. The interpolation formula of each region optimizes the interpolated image based on the adjustment of variable parameters. Different interpolation effects are achieved in different regions, so that the pixel values of the interpolation points are more consistent with the real pixel values, and better interpolation effects are achieved.
[0114] In addition, based on the interpolation formula proposed in the embodiment of the present invention, based on the variable parameter λ therein, different interpolation effects can be achieved by alignment and adjustment. Therefore, in the embodiment of the present application, if the interpolation data of the point to be interpolated does not meet the requirements, iterative correction can be performed until the best interpolation reconstructed image is obtained. Specifically, the high-resolution image is obtained by performing image super-resolution reconstruction through the parameterized Thiele continued fraction interpolation method, which also includes:
[0115] Performing image quality calculation on the image obtained by image super-resolution reconstruction based on preset evaluation indicators;
[0116] If the image quality data does not meet the preset high-resolution image condition, the image super-resolution reconstruction process is iterated by changing the variable parameters in the parameterized Thiele continued fraction interpolation until an image meeting the preset high-resolution image condition is obtained.
[0117] Perform inverse color space conversion to restore the original color space to obtain high-resolution and high-quality images.
[0118] Based on the above-mentioned image processing method using parameterized Thiele continued fraction interpolation, an embodiment of the present invention further provides an image processing device using parameterized Thiele continued fraction interpolation, including:
[0119] An image channel image extraction unit, used for converting the image to be processed into a color space and extracting a corresponding channel image, wherein the channel image at least includes a channel representing the brightness of the image;
[0120] The channel image enhancement unit is used to perform frequency domain space conversion on different channels, and adopts a matched filtering method based on the characteristics of different channels to suppress the low-frequency components in the channel spectrum and enhance the corresponding channel components;
[0121] A denoising and edge protection unit is used to filter the smooth area and the non-smooth area of the channel image to different degrees based on different channels with component enhancement, and perform a denoising process based on edge protection;
[0122] A high-resolution channel image acquisition unit is used to acquire a high-resolution image by performing image super-resolution reconstruction based on different channel images through a parameterized Thiele continued fraction interpolation method;
[0123] The channel image fusion unit is used to perform inverse conversion of the color space to restore it to the original color space to obtain high-resolution and high-quality images.
[0124] An embodiment of the present invention further provides an electronic device, which can be implemented in various forms, such as a dedicated terminal with an image processing function, or an electronic device or a cloud server with an image processing function, etc. The electronic device includes:
[0125] A memory for storing executable instructions;
[0126] The processor is used to implement the above-mentioned image processing method using parameterized Thiele continued fraction interpolation when running the executable instructions stored in the memory.
[0127] It can be understood that the memory can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories. The memory in the embodiment of the present invention can store data to support the operation of the terminal, and examples of these data include: any computer program used to operate on the terminal, such as an operating system and an application program. Among them, the operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application program can include various applications. The image processing device provided in the embodiment of the present invention can be implemented in a combination of software and hardware. As an example of the image processing device provided in the embodiment of the present invention being implemented in a combination of software and hardware, the image processing device provided in the embodiment of the present invention can be directly embodied as a combination of software modules executed by a processor, and the software module can be located in a storage medium, and the storage medium is located in a memory. The processor reads the executable instructions included in the software module in the memory, and combines the necessary hardware to complete the image processing method for applying parameterized Thiele continued fraction interpolation provided in the embodiment of the present invention.
[0128] The present invention is not limited to the above-mentioned specific implementation modes. Various changes made by ordinary technicians in this field based on the above-mentioned concepts without creative work are all within the protection scope of the present invention.
Claims
1. An image processing method using parameterized Thiele continued fraction interpolation, characterized in that: The steps include: Converting the image to be processed into a color space and extracting a corresponding channel image, wherein the channel image at least includes a channel representing the brightness of the image, and converting the image to be processed into a color space and extracting the corresponding channel image specifically converts the image from RGB to HSV color space; Perform frequency domain space conversion on different channels, and use matched filtering method based on the characteristics of different channels to suppress the low-frequency components in the channel spectrum and improve the corresponding channel components; Based on different channels with component enhancement, the smooth and non-smooth areas of the channel image are filtered to different degrees through filters to perform a denoising process based on edge protection; Based on images from different channels, high-resolution images are obtained by performing image super-resolution reconstruction using the parameterized Thiele continued fraction interpolation method. Perform inverse color space conversion to restore the original color space to obtain high-resolution and high-quality images; The frequency domain space conversion is performed on different channels, and a matched filtering method is used based on the characteristics of different channels to suppress low-frequency components in the channel spectrum and improve the corresponding channel components, including: sliding the S channel and V channel images based on a preset sliding window, and performing homomorphic filtering on the image area in each sliding window, wherein the homomorphic filtering cutoff frequency of each image area is matched with the image area spectrum; merging the images after homomorphic filtering of all sliding window image areas to obtain an S channel image with enhanced saturation and a V channel image with enhanced brightness and contrast; The method of filtering the smooth area and the non-smooth area of the channel image to different degrees through the filter to perform a denoising process based on edge protection includes: filtering based on a filter model, wherein the model is: The parameter a of the model p′ ,b p′ Satisfy the objective optimization function E: Among them, G is the guide map, X is the image before filtering, G and X are allowed to be the same, is a window with p' as the center and ζ1 as the radius, is the filtered image, a p′ and b p′ yes Two constants in the window, is the edge perception weight, which is the local variance ratio of the pixels in the 3x3 window and all pixels in the (2ζ1+1)x(2ζ1+1) window, χ(p′)=σ G,1 (p′)σ G,ζ1 (p′),σ G,1 (p′) and σ G,ζ1 (p′) respectively represents the image G in the window Ω1(p′) and Standard deviation, ε=(0.001*L) 2 , L is the dynamic range of image X; k is a constant; var(x(p))-x(p') represents the difference between the variance of x(p) and x(p'); var(x(p))-min(x(p)) represents the difference between the variance of x(p) and the minimum value of x(p); The formula of the parameterized Thiele continued fraction interpolation adopts a binary Thiele continued fraction interpolation with a single parameter, and the binary Thiele continued fraction interpolation with a single parameter is: in, For j = 0, 1, ..., n; p = 1, 2, ..., m; i = p, p + 1, ..., m, a k,0 ,a k,1 ,...,a k,l , a k,l+1 ,a k,l+2 ,...,a k,n For this formula The calculated value after the introduction of parameter λ; For p = 0, 1, ..., k-1, k+1, ..., m; q = 1, 2, ..., n; j = q, q+1, ..., n, 2. The image processing method using parameterized Thiele continued fraction interpolation according to claim 1, characterized in that: The high-resolution image is obtained by performing image super-resolution reconstruction using a parameterized Thiele continued fraction interpolation method, different regions are divided based on image gradient changes, and parameterized Thiele continued fraction interpolation formulas are used for interpolation respectively. The interpolation formula of each region optimizes the interpolated image based on the adjustment of variable parameters.
3. The image processing method using parameterized Thiele continued fraction interpolation according to claim 2, characterized in that: The method of performing image super-resolution reconstruction by parameterized Thiele continued fraction interpolation method to obtain a high-resolution image also includes: Performing image quality calculation on the image obtained by image super-resolution reconstruction based on preset evaluation indicators; If the image quality data does not meet the preset high-resolution image condition, the image super-resolution reconstruction process is iterated by changing the variable parameters in the parameterized Thiele continued fraction interpolation until an image meeting the preset high-resolution image condition is obtained.
4. An image processing device using parameterized Thiele continued fraction interpolation, characterized in that: include: An image channel image extraction unit is used to convert the image to be processed into a color space and extract a corresponding channel image, wherein the channel image at least includes a channel representing the brightness of the image, and the image to be processed is converted into a color space and the corresponding channel image is extracted, specifically, the image is converted from RGB to HSV color space; The channel image enhancement unit is used to perform frequency domain space conversion on different channels, and adopts a matched filtering method based on the characteristics of different channels to suppress the low-frequency components in the channel spectrum and enhance the corresponding channel components; A denoising and edge protection unit is used to filter the smooth area and the non-smooth area of the channel image to different degrees based on different channels with component enhancement, and perform a denoising process based on edge protection; A high-resolution channel image acquisition unit is used to acquire a high-resolution image by performing image super-resolution reconstruction based on different channel images through a parameterized Thiele continued fraction interpolation method; Channel image fusion unit, used to perform inverse color space conversion to restore the original color space to obtain high-resolution and high-quality images; The channel image enhancement unit performs frequency domain space conversion on different channels, adopts a matched filtering method based on the characteristics of different channels, suppresses low-frequency components in the channel spectrum, and improves the corresponding channel components, including: sliding the S channel and V channel images based on a preset sliding window, and performing homomorphic filtering on the image area in each sliding window, wherein the homomorphic filtering cutoff frequency of each image area matches the image area spectrum; merging the images after homomorphic filtering of all sliding window image areas to obtain an S channel image with enhanced saturation and a V channel image with enhanced brightness and contrast; The noise reduction and edge protection unit filters the smooth area and the non-smooth area of the channel image to different degrees through a filter, including: filtering based on a filter model, the model is: The parameter a of the model p′ ,b p′ Satisfy the objective optimization function E: Among them, G is the guide map, X is the image before filtering, G and X are allowed to be the same, is a window with p' as the center and ζ1 as the radius, is the filtered image, a p′ and b p′ yes Two constants in the window, is the edge perception weight, which is the local variance ratio of the pixels in the 3x3 window and all pixels in the (2ζ1+1)x(2ζ1+1) window, χ(p′)=σ G,1 (p′)σ G,ζ1 (p′),σ G,1 (p′) and σ G,ζ1 (p′) respectively represents the image G in the window Ω1(p′) and Standard deviation, ε=(0.001*L) 2 , L is the dynamic range of image X; k is a constant; var(x(p))-x(p') represents the difference between the variance of x(p) and x(p'); var(x(p))-min(x(p)) represents the difference between the variance of x(p) and the minimum value of x(p); The high-resolution channel image acquisition unit uses a parameterized Thiele continued fraction interpolation formula using a binary Thiele continued fraction interpolation formula containing a single parameter. The binary Thiele continued fraction interpolation formula containing a single parameter is: in, For j = 0, 1, ..., n; p = 1, 2, ..., m; i = p, p + 1, ..., m, For this formula The calculated value after the introduction of parameter λ; For p = 0, 1, ..., k-1, k+1, ..., m; q = 1, 2, ..., n; j = q, q+1, ..., n, 5. An electronic device, characterized in that: The electronic device comprises: A memory for storing executable instructions; A processor, configured to implement the image processing method using parameterized Thiele continued fraction interpolation as described in any one of claims 1 to 3 when running the executable instructions stored in the memory.
