CN108460733B - Gradually refined image denoising method and system - Google Patents
Gradually refined image denoising method and system Download PDFInfo
- Publication number
- CN108460733B CN108460733B CN201810097180.8A CN201810097180A CN108460733B CN 108460733 B CN108460733 B CN 108460733B CN 201810097180 A CN201810097180 A CN 201810097180A CN 108460733 B CN108460733 B CN 108460733B
- Authority
- CN
- China
- Prior art keywords
- pixel point
- pixel
- gradient
- region
- fitted
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims description 33
- 238000001914 filtration Methods 0.000 claims abstract description 34
- 239000013598 vector Substances 0.000 claims description 59
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 230000000750 progressive effect Effects 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 6
- 230000006870 function Effects 0.000 description 6
- 230000002146 bilateral effect Effects 0.000 description 5
- 238000004590 computer program Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000000717 retained effect Effects 0.000 description 2
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration by the use of local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
Abstract
A gradient of each pixel point is calculated, then based on gradient similarity, an area with similar gradient of the pixel points is searched, then based on the curved surface, plane fitting is carried out, and based on the fitted plane, a normal line of the pixel point is calculated; and then, searching a region similar to the pixel point based on the normal of the pixel point, performing surface fitting (such as quadratic surface fitting) of the pixel value based on the region, and filtering the pixel point based on the fitted surface, thereby completing refined filtering and achieving the purpose of refined denoising.
Description
Technical Field
The invention relates to the field of image processing, in particular to a gradually refined image denoising method and system.
Background
Although an ancient problem is caused by image denoising, the image denoising is always paid attention by academia and industry, and various methods are continuously proposed to solve the problem of noise elimination in various scenes.
The image denoising methods which are concerned and used more frequently include bilateral filtering, guided filtering, L0 norm smoothing (L0-smooth) filtering, and the like. Although these filtering methods have made great progress in dealing with the image denoising problem, it is undeniable that they still have great limitations, such as not performing denoising with much refinement.
Disclosure of Invention
In view of the above, the present application provides a method and a system for denoising an image gradually refined.
According to a first aspect, an embodiment provides a method for denoising an image with progressive refinement, including:
calculating the gradient of each pixel point in the image;
for each pixel point in the image, based on gradient similarity, searching a region with similar gradient of the pixel point, performing plane fitting of pixel values according to all pixel points in the region, and calculating a normal vector of the pixel point based on the fitted plane;
for each pixel point in the image, based on the similarity of normal vectors, a region with similar normal vectors of the pixel point is searched, curved surface fitting of pixel values is carried out according to all pixel points in the region, and filtering is carried out on the pixel point based on the fitted curved surface.
In an embodiment, for any pixel point in the image, when a region with a similar gradient of the pixel point cannot be found based on the gradient similarity, the pixel point is filtered by using the pixel point of the neighborhood of the pixel point to obtain the pixel value of the pixel point after filtering, the gradient of the pixel point is recalculated, and a region with a similar gradient of the pixel point is found based on the gradient similarity.
In an embodiment, the calculating a normal vector of the pixel point based on the fitted plane includes:
calculating a normal vector of the fitted plane as a normal vector of the pixel point; alternatively, the first and second electrodes may be,
according to the fitted plane, verifying whether the pixel points in the region with similar gradient of the pixel point belong to the fitted plane, if so, retaining, otherwise, rejecting; and performing plane fitting of the pixel values again according to the reserved pixel points, and calculating a normal vector of the plane fitted again to serve as the normal vector of the pixel points.
In an embodiment, the filtering the pixel point based on the fitted curved surface includes:
calculating the pixel value of the pixel point according to the fitted curved surface and the coordinate of the pixel point; alternatively, the first and second electrodes may be,
according to the fitted curved surface, verifying whether pixel points in the area with the similar normal vector of the pixel point belong to the fitted curved surface, if so, retaining, otherwise, rejecting; and performing surface fitting of the pixel values again according to the reserved pixel points, and calculating the pixel values of the pixel points according to the surface fitted again and the coordinates of the pixel points.
In one embodiment, the bilateral filtering is performed on the image before the gradient of each pixel point in the image is calculated.
In one embodiment, the pixel value is a gray value or a color value.
In one embodiment:
based on the gradient similarity, finding a region with similar gradient of the pixel point comprises the following steps: based on the gradient similarity, a region with similar gradient of the pixel point is segmented by a region segmentation algorithm;
based on the similarity of the normal vectors, finding a region with a similar normal vector of the pixel point comprises the following steps: based on the similarity of the normal vectors, a region with similar normal vectors of the pixel points is segmented by a region segmentation algorithm.
In one embodiment, the curved surface is a quadratic surface.
According to a second aspect, an embodiment provides a progressive refinement image denoising system, including:
a memory for storing a program;
a processor for implementing the method of any of the above embodiments by executing the program stored in the memory.
According to a third aspect, an embodiment provides a computer-readable storage medium comprising a program executable by a processor to implement the method of any of the above embodiments.
According to the gradually refined image denoising method, the gradually refined image denoising system and the computer readable storage medium of the embodiments, the gradient of each pixel point is calculated, then an area with similar gradient of the pixel points is found based on the gradient similarity, then plane fitting is carried out based on the curved surface, and the normal line of the pixel point is calculated based on the fitted plane; and then, searching a region similar to the pixel point based on the normal of the pixel point, performing surface fitting (such as quadratic surface fitting) of the pixel value based on the region, and filtering the pixel point based on the fitted surface, thereby completing refined filtering and achieving the purpose of refined denoising.
Drawings
FIG. 1 is a flowchart of an embodiment of a progressive refinement image denoising method;
FIG. 2 illustrates an embodiment of a progressive refinement image denoising system.
Detailed Description
The present invention will be described in further detail with reference to the following detailed description and accompanying drawings. Wherein like elements in different embodiments are numbered with like associated elements. In the following description, numerous details are set forth in order to provide a better understanding of the present application. However, those skilled in the art will readily recognize that some of the features may be omitted or replaced with other elements, materials, methods in different instances. In some instances, certain operations related to the present application have not been shown or described in detail in order to avoid obscuring the core of the present application from excessive description, and it is not necessary for those skilled in the art to describe these operations in detail, so that they may be fully understood from the description in the specification and the general knowledge in the art.
Furthermore, the features, operations, or characteristics described in the specification may be combined in any suitable manner to form various embodiments. Also, the various steps or actions in the method descriptions may be transposed or transposed in order, as will be apparent to one of ordinary skill in the art. Thus, the various sequences in the specification and drawings are for the purpose of describing certain embodiments only and are not intended to imply a required sequence unless otherwise indicated where such sequence must be followed.
The numbering of the components as such, e.g., "first", "second", etc., is used herein only to distinguish the objects as described, and does not have any sequential or technical meaning. The term "connected" and "coupled" when used in this application, unless otherwise indicated, includes both direct and indirect connections (couplings).
When the inventor researches the problem of image denoising, the inventor finds that the existing denoising methods, such as bilateral filtering, guided filtering, L0 norm smoothing filtering, and the like, do not perform denoising in a finer manner, which mainly lacks of full utilization of the geometric characteristics of the periphery of the image where the pixel points to be filtered are located. Therefore, on the basis of this, the inventors conceived that:
each local area of the image can be approximately regarded as a curved surface, the neighborhood of each pixel point in the curved surface can be approximately regarded as a plane, and therefore, a normal line exists, the normal lines have continuity, and the normal line of each pixel point can be calculated by finding the plane where the pixel point is located; and then, searching an area similar to the pixel point based on the normal of the pixel point, namely, the normals of the pixel point in the area are all similar, then performing surface fitting (for example, quadratic surface fitting) of the pixel value based on the area, and filtering the pixel point based on the fitted surface. In particular, the filtering for refinement may be performed in three stages, as described in detail below.
The first stage is as follows: and carrying out strong bilateral filtering and the like on the image to obtain an image which greatly eliminates the noise influence, and calculating the gradient of each pixel point based on the filtered image. It should be noted that the first stage is not necessary, and for example, the gradient of each pixel point in the original image may be directly calculated.
And a second stage: for each pixel point, the pixel points with similar gradients near the pixel point are searched, the pixel points are used for fitting a plane of a pixel value, and the normal of the pixel point is based on the plane of the fitted pixel value.
And a third stage: for each pixel point, a pixel point with similar normal lines near the pixel point is searched, a curved surface (such as a quadric surface) with a pixel value is fitted by utilizing the pixel points, and the pixel point is filtered based on the fitted curved surface, so that a refined filtering result is finally obtained.
The present invention is described above with reference to the accompanying drawings.
Referring to fig. 1, an embodiment of the present invention discloses a method for denoising an image with progressive refinement, including steps S10 to S50, which are described in detail below.
Step S10: and calculating the gradient of each pixel point in the image. Sometimes, the image is greatly interfered by noise, so that the noise has a very large influence on the calculation of the gradient, and therefore, the image can be considered to be filtered before the gradient is calculated, for example, the image is filtered by using the existing bilateral filtering algorithm, and then the gradient of each pixel point in the filtered image is calculated.
Step S30: for each pixel point in the image, based on gradient similarity, searching a region with similar gradient of the pixel point, performing plane fitting of pixel values according to all pixel points in the region, and calculating a normal vector of the pixel point based on the fitted plane.
Sometimes, when a certain pixel point in the image cannot find an area with a similar gradient of the pixel point based on the gradient similarity, it indicates that the pixel point itself is an outlier (outlier) caused by noise, and for this case, in an embodiment, the step S30 further includes: for any pixel point in the image, when a region with a similar gradient of the pixel point cannot be found based on the gradient similarity, filtering (for example, performing median filtering) the pixel point by using the pixel points of the neighborhood of the pixel point to obtain the pixel value of the pixel point after filtering, recalculating the gradient of the pixel point, and finding a region with a similar gradient of the pixel point based on the gradient similarity. Filtering the pixel point by using the pixel point of the neighborhood of the pixel point, for example, median filtering the pixel point by using the neighborhood of the pixel point, and calculating to obtain a new pixel value of the pixel point.
In the step S30, there are many ways to find a region with similar gradient of a pixel point based on the gradient similarity, for example, in an embodiment, finding a region with similar gradient of a pixel point based on the gradient similarity may include: based on the gradient similarity, a region with similar gradient of the pixel point is segmented by a region segmentation algorithm (such as a Graph-Cut algorithm).
By the method, a region with similar gradient can be found for each pixel point in the image. For any pixel point in the image, after a region with similar gradient of the pixel point is found, plane fitting of pixel values is carried out according to all pixel points in the region, and a normal vector of the pixel point is calculated based on the fitted plane, so that the normal vector of each pixel point in the image can be calculated. In an embodiment, the calculating a normal vector of the pixel point based on the fitted plane in step S30 may include: and calculating the normal vector of the fitted plane as the normal vector of the pixel point. In an embodiment, the calculating a normal vector of the pixel point based on the fitted plane in step S30 may include: according to the fitted plane, verifying whether the pixel points in the region with similar gradient of the pixel point belong to the fitted plane, if so, retaining, otherwise, indicating that the pixel points belong to outliers, and removing the outliers; and performing plane fitting of the pixel values again according to the reserved pixel points, and calculating a normal vector of the plane fitted again to serve as the normal vector of the pixel points.
The normal vector of each pixel point in the image can be calculated through step S30.
Step S50: for each pixel point in the image, based on the similarity of normal vectors, a region with similar normal vectors of the pixel point is searched, curved surface fitting of pixel values is carried out according to all pixel points in the region, and filtering is carried out on the pixel point based on the fitted curved surface. Similarly, in the step S50, there are many ways to find a region with similar normal vectors of the pixel points based on the similarity of the normal vectors, for example, in an embodiment, finding a region with similar normal vectors of the pixel points based on the similarity of the normal vectors includes: based on the similarity of the normal vectors, a region with similar normal vectors of the pixel points is segmented through a region segmentation algorithm (such as a Graph-Cut algorithm).
In an embodiment, the filtering the pixel point based on the fitted curved surface in step S50 may include: and calculating the pixel value of the pixel point according to the fitted curved surface and the coordinate of the pixel point. In an embodiment, the filtering the pixel point based on the fitted curved surface in step S50 may also include: according to the fitted curved surface, verifying whether pixel points in the area with the similar normal vector of the pixel point belong to the fitted curved surface, if so, retaining, otherwise, indicating that the pixel points are outliers, and removing the outliers; and performing surface fitting of the pixel values again according to the reserved pixel points, and calculating the pixel values of the pixel points according to the surface fitted again and the coordinates of the pixel points. In an embodiment, the curved surface involved in step S50 may be a quadratic surface.
It should be noted that the pixel values mentioned in the present invention may be gray values or color values. For example, the pixel values mentioned in the present invention may be color values of RGB space. As will be understood by those skilled in the art, when the pixel value is a color value of RGB space, the pixel value is composed of an R color component, a G color component, and a B color component; correspondingly, the plane fitting of the pixel values of all the pixel points in any one region can mean that the plane fitting of the R color component, the plane fitting of the G color component and the plane fitting of the B color component are respectively carried out on all the pixel points in the region, similarly, the surface fitting of the pixel values is carried out, and the surface fitting of the R color component, the surface fitting of the G color component and the surface fitting of the B color component are respectively carried out on all the pixel points in the region.
The following does not take any pixel point (x) in the image0,y0) The above steps are specifically described as an example.
Any pixel point in the image, e.g. pixel point (x)0,y0) The gradient is calculated in step S10 and is not recorded as gradI (x)0,y0). There are many ways to calculate the gradient similarity of two pixels, or to calculate the similarity of two vectors, for example, for any two pixels (x)1,y1)、(x2,y2) The gradient similarity between them can be calculated by:
gradI(x1,y1) Representing a pixel (x)1,y1) Gradient of (a), gradI (x)2,y2) Representing a pixel (x)2,y2) A gradient of (a); the larger the cos theta value obtained by calculation is, the larger the included angle theta is, namely, the smaller the similarity is.
After the gradient similarity is defined, the pixel point (x) can be found based on the gradient similarity0,y0) A region with a similar gradient. When the pixel point (x) cannot be found based on the gradient similarity0,y0) A region with a gradient, then a pixel point (x) is indicated0,y0) Is an outlier caused by noise and thus can pass through a pixel (x)0,y0) For example 8 neighborhoods, to pixel point (x)0,y0) Performing median filtering to obtain pixel points (x)0,y0) The filtered pixel values are then based on pixel points (x)0,y0) Recalculating pixel point (x) from the filtered pixel value0,y0) Gradient of (a) gradI (x)0,y0) Then based on the gradient similarity, searching pixel point (x)0,y0) A region with a similar gradient. By the method, for any pixel point, a region with similar gradient of the pixel point can be found finally. Pixel point (x) is not changed0,y0) Is referred to as P.
And then, carrying out plane fitting on the pixel values according to all the pixel points in the region P. There are many methods for plane fitting, such as least squares, etc., and will not be described herein. To pairIn all pixel points in the region P, the plane finally fitted is not made to be I (x, y) ═ a1x+b1y+c1Wherein I (x, y) represents the pixel value of the pixel point (x, y).
At a pixel point (x)0,y0) Is fitted out of a plane I (x, y) ═ a1x+b1y+c1Then, based on the fitted plane I (x, y) ═ a1x+b1y+c1Calculating a pixel (x)0,y0) The normal vector of (a) may be obtained by directly calculating the plane I (x, y) ═ a1x+b1y+c1E.g. to obtain a normal vector [ a ]1,b1,-1]Then, the plane I (x, y) is a1x+b1y+c1As a pixel point (x) is the normal vector of0,y0) The normal vector of (2). Of course, based on the fitted plane I (x, y) ═ a1x+b1y+c1Calculating a pixel (x)0,y0) The normal vector of (c) can also be calculated as follows: according to the fitted plane I (x, y) ═ a1x+b1y+c1Respectively verifying whether each pixel point in the region P belongs to the fitted plane, if so, retaining, if not, rejecting, and then performing plane fitting of the pixel values again according to the pixel points retained in the region P, for example, obtaining a plane I (x, y) ═ a2x+b2y+c2Then calculate the plane I (x, y) ═ a fitted again2x+b2y+c2E.g. to obtain a normal vector [ a ]2,b2,-1]As a pixel point (x)0,y0) The normal vector of (2).
Any pixel point in the image, e.g. pixel point (x)0,y0) After step S30, the normal vector can be finally calculated and recorded as nor (x)0,y0). Similarly, there are many ways to calculate the similarity of the normal vectors of the two pixels, for example, calculating the included angle between the two normal vectors, and the like, which are not described herein again. Then based on the similarity of normal vectors, searching pixel points (x)0,y0) A region with similar normal vectors. Pixel point (x) is not changed0,y0) The region with similar normal vectors of (a) is called Q.
And then, performing surface fitting of the pixel values according to all the pixel points in the region Q. There are many methods for surface fitting, for example, one of the existing methods may be adopted, and the details are not described herein. For all pixel points in the region Q, the curved surface I (x, y) finally fitted is not made to be a3x2+b3y2+c3xy+d3x+e3y+f3Wherein I (x, y) represents the pixel value of the pixel point (x, y).
At a pixel point (x)0,y0) The region Q of (a) is fitted to a curved surface I (x, y) ═ a3x2+b3y2+c3xy+d3x+e3y+f3Then, based on the fitted curved surface I (x, y) ═ a3x2+b3y2+c3xy+d3x+e3y+f3For the pixel point (x)0,y0) And (6) filtering. For example, according to the fitted curved surface I (x, y) ═ a3x2+b3y2+c3xy+d3x+e3y+f3And the pixel point (x)0,y0) Calculating the pixel point (x)0,y0) Pixel value of (a) I (x)0,y0)＝a3x0 2+b3y0 2+c3x0y0+d3x0+e3y0+f3. Of course, the curved surface I (x, y) ═ a based on the fitting3x2+b3y2+c3xy+d3x+e3y+f3For the pixel point (x)0,y0) The filtering can also be performed as follows: according to the fitted curved surface I (x, y) ═ a3x2+b3y2+c3xy+d3x+e3y+f3Respectively verifying whether each pixel point in the region Q belongs to the fitted curved surface, if so, retaining, if not, rejecting, and then performing curved surface fitting of the pixel values again according to the pixel points retained in the region Q, for exampleObtain a curved surface I (x, y) ═ a4x2+b4y2+c4xy+d4x+e4y+f4Then, the curve I (x, y) is fitted again according to the curve a4x2+b4y2+c4xy+d4x+e4y+f4And a pixel point (x)0,y0) Calculating the pixel point (x)0,y0) Pixel value of (a) I (x)0,y0)＝a4x0 2+b4y0 2+c4x0y0+d4x0+e4y0+f4Completing the pixel point (x)0,y0) Filtering of (2).
It should be noted that, when a pixel value is a color value, for example, taking RGB space as an example, plane fitting of R color component, plane fitting of G color component, and plane fitting of B color component are performed respectively, and accordingly, whether a certain pixel point in a region belongs to a fitted plane is verified, and if a color value of any component of a pixel point does not belong to the plane fitting of its corresponding color component, the pixel point can be eliminated; the curved surface is similar, and the description is omitted here. Of course, the pixel value of the pixel point is calculated according to the fitted curved surface and the space coordinates of the pixel point, and if the pixel value is a color value, the value of the color component corresponding to the pixel point is also calculated according to the fitted curved surface of the color component and the space coordinates of the pixel point.
Accordingly, referring to fig. 2, an embodiment of the present invention further discloses a gradually refined image denoising system, which includes a memory 10 and a processor 30, wherein the memory 10 is used for storing a program, and the processor 30 is used for implementing the gradually refined image denoising method according to any embodiment of the present invention by executing the program stored in the memory 10.
Those skilled in the art will appreciate that all or part of the functions of the various methods in the above embodiments may be implemented by hardware, or may be implemented by computer programs. When all or part of the functions of the above embodiments are implemented by a computer program, the program may be stored in a computer-readable storage medium, and the storage medium may include: a read only memory, a random access memory, a magnetic disk, an optical disk, a hard disk, etc., and the program is executed by a computer to realize the above functions. For example, the program may be stored in a memory of the device, and when the program in the memory is executed by the processor, all or part of the functions described above may be implemented. In addition, when all or part of the functions in the above embodiments are implemented by a computer program, the program may be stored in a storage medium such as a server, another computer, a magnetic disk, an optical disk, a flash disk, or a removable hard disk, and may be downloaded or copied to a memory of a local device, or may be version-updated in a system of the local device, and when the program in the memory is executed by a processor, all or part of the functions in the above embodiments may be implemented.
The present invention has been described in terms of specific examples, which are provided to aid understanding of the invention and are not intended to be limiting. For a person skilled in the art to which the invention pertains, several simple deductions, modifications or substitutions may be made according to the idea of the invention.
Claims (10)
1. A method for denoising an image gradually refined is characterized by comprising the following steps:
calculating the gradient of each pixel point in the image;
for each pixel point in the image, based on gradient similarity, searching a region with similar gradient of the pixel point, performing plane fitting of pixel values according to all pixel points in the region, and calculating a normal vector of the pixel point based on the fitted plane;
for each pixel point in the image, based on the similarity of normal vectors, a region with similar normal vectors of the pixel point is searched, curved surface fitting of pixel values is carried out according to all pixel points in the region, and filtering is carried out on the pixel point based on the fitted curved surface.
2. The image denoising method of claim 1, wherein for any pixel point in the image, when a region with similar gradient of the pixel point is not found based on gradient similarity, the pixel point is filtered by using the pixel points of the neighborhood of the pixel point to obtain the filtered pixel value of the pixel point, and the gradient of the pixel point is recalculated, and a region with similar gradient of the pixel point is found based on gradient similarity.
3. The method of denoising image of claim 1, wherein the calculating the normal vector of the pixel point based on the fitted plane comprises:
calculating a normal vector of the fitted plane as a normal vector of the pixel point; alternatively, the first and second electrodes may be,
according to the fitted plane, verifying whether the pixel points in the region with similar gradient of the pixel point belong to the fitted plane, if so, retaining, otherwise, rejecting; and performing plane fitting of the pixel values again according to the reserved pixel points, and calculating a normal vector of the plane fitted again to serve as the normal vector of the pixel points.
4. The image denoising method of claim 1, wherein the filtering the pixel point based on the fitted surface comprises:
calculating the pixel value of the pixel point according to the fitted curved surface and the coordinate of the pixel point; alternatively, the first and second electrodes may be,
according to the fitted curved surface, verifying whether pixel points in the area with the similar normal vector of the pixel point belong to the fitted curved surface, if so, retaining, otherwise, rejecting; and performing surface fitting of the pixel values again according to the reserved pixel points, and calculating the pixel values of the pixel points according to the surface fitted again and the coordinates of the pixel points.
5. An image denoising method as claimed in any one of claims 1 to 4, wherein the image is bilaterally filtered before the gradient of each pixel point in the image is calculated.
6. The image denoising method of any one of claims 1 to 4, wherein the pixel value is a gray value or a color value.
7. The image denoising method of claim 1, wherein:
based on the gradient similarity, finding a region with similar gradient of the pixel point comprises the following steps: based on the gradient similarity, a region with similar gradient of the pixel point is segmented by a region segmentation algorithm;
based on the similarity of the normal vectors, finding a region with a similar normal vector of the pixel point comprises the following steps: based on the similarity of the normal vectors, a region with similar normal vectors of the pixel points is segmented by a region segmentation algorithm.
8. The image denoising method of claim 1, wherein the curved surface is a quadratic surface.
9. A progressive refinement image denoising system, comprising:
a memory for storing a program;
a processor for implementing the method of any one of claims 1 to 8 by executing a program stored by the memory.
10. A computer-readable storage medium, characterized by comprising a program executable by a processor to implement the method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810097180.8A CN108460733B (en) | 2018-01-31 | 2018-01-31 | Gradually refined image denoising method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810097180.8A CN108460733B (en) | 2018-01-31 | 2018-01-31 | Gradually refined image denoising method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108460733A CN108460733A (en) | 2018-08-28 |
| CN108460733B true CN108460733B (en) | 2022-04-22 |
Family
ID=63238411
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810097180.8A Active CN108460733B (en) | 2018-01-31 | 2018-01-31 | Gradually refined image denoising method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108460733B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106228609A (en) * | 2016-07-09 | 2016-12-14 | 武汉广图科技有限公司 | A kind of oblique photograph three-dimensional modeling method based on spatial signature information |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7747106B2 (en) * | 2005-06-13 | 2010-06-29 | Sarnoff Corporation | Method and system for filtering, registering, and matching 2.5D normal maps |
| JP5197414B2 (en) * | 2009-02-02 | 2013-05-15 | オリンパス株式会社 | Image processing apparatus and image processing method |
| US9576346B2 (en) * | 2015-03-24 | 2017-02-21 | Intel Corporation | Non-local means image denoising with an adaptive directional spatial filter |
| CN104778669B (en) * | 2015-04-16 | 2017-12-26 | 北京邮电大学 | rapid image denoising method and device |
| CN105046710A (en) * | 2015-07-23 | 2015-11-11 | 北京林业大学 | Depth image partitioning and agent geometry based virtual and real collision interaction method and apparatus |
| CN107223268B (en) * | 2015-12-30 | 2020-08-07 | 中国科学院深圳先进技术研究院 | Three-dimensional point cloud model reconstruction method and device |
| CN107256557B (en) * | 2017-05-03 | 2020-05-22 | 华南理工大学 | Error-controllable subdivision surface image vectorization method |
-
2018
- 2018-01-31 CN CN201810097180.8A patent/CN108460733B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106228609A (en) * | 2016-07-09 | 2016-12-14 | 武汉广图科技有限公司 | A kind of oblique photograph three-dimensional modeling method based on spatial signature information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108460733A (en) | 2018-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9292928B2 (en) | Depth constrained superpixel-based depth map refinement | |
| Chiu et al. | Fast SIFT design for real-time visual feature extraction | |
| WO2018082185A1 (en) | Image processing method and device | |
| JP5487106B2 (en) | Image processing apparatus and method, and image display apparatus | |
| Xie et al. | Single depth image super resolution and denoising via coupled dictionary learning with local constraints and shock filtering | |
| Krig et al. | Image pre-processing | |
| Lo et al. | Joint trilateral filtering for depth map super-resolution | |
| JP5908844B2 (en) | Image processing apparatus and image processing method | |
| WO2016183743A1 (en) | System and method for supporting image denoising based on neighborhood block dimensionality reduction | |
| Tan et al. | Multipoint filtering with local polynomial approximation and range guidance | |
| WO2016062259A1 (en) | Transparency-based matting method and device | |
| CN113744142B (en) | Image restoration method, electronic device and storage medium | |
| CN108492326B (en) | Stereo matching method and system with gradually refined resolution ratio from low to high | |
| Schäfer et al. | Depth and intensity based edge detection in time-of-flight images | |
| Yu et al. | Single image interpolation exploiting semi-local similarity | |
| CN111476750A (en) | Method, device and system for carrying out stain detection on imaging module and storage medium | |
| CN108460733B (en) | Gradually refined image denoising method and system | |
| Zhao et al. | Single depth image super-resolution with multiple residual dictionary learning and refinement | |
| JP5617426B2 (en) | Jaggy mitigation processing apparatus and jaggy mitigation processing method | |
| GB2545649B (en) | Artefact detection | |
| Wu et al. | An effective content-aware image inpainting method | |
| Dung et al. | Implementation of shock filter for digital X-ray image processing | |
| Zuo et al. | Rotation invariant similarity measure for non-local self-similarity based image denoising | |
| CN108337498B (en) | Parallax calculation method and system for curved surface fitting | |
| Sintunata et al. | Skeleton extraction in cluttered image based on Delaunay triangulation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |