CN116977207A - Image noise reduction method based on selective smoothing filtering - Google Patents
Image noise reduction method based on selective smoothing filtering Download PDFInfo
- Publication number
- CN116977207A CN116977207A CN202310809660.3A CN202310809660A CN116977207A CN 116977207 A CN116977207 A CN 116977207A CN 202310809660 A CN202310809660 A CN 202310809660A CN 116977207 A CN116977207 A CN 116977207A
- Authority
- CN
- China
- Prior art keywords
- rectangular frame
- image
- pixel
- value
- noise reduction
- 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.)
- Pending
Links
- 238000009499 grossing Methods 0.000 title claims abstract description 24
- 238000001914 filtration Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 title claims abstract description 17
- 238000010408 sweeping Methods 0.000 claims abstract description 5
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 238000012935 Averaging Methods 0.000 claims description 4
- 238000003708 edge detection Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 3
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- 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
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using 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
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- 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
- G06T2207/20021—Dividing image into blocks, subimages or windows
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The invention aims to provide an image noise reduction method based on selective smoothing filtering, which comprises the following steps of: A. inputting a gray image of an image to be denoised, establishing a 3*3 rectangular frame, sweeping the whole image to be denoised by using the rectangular frame, and calculating the gradient t of a pixel point in the center of the rectangular frame and each pixel point on the edge of the rectangular frame after each movement; b, setting a noise threshold value x, and inputting the rectangular frame into the step B if the noise threshold value x is not more than the t; otherwise, inputting the rectangular frame into the step C; B. set edge detectionA threshold value n, wherein a threshold value K is calculated based on each rectangular frame input in the step A; if K is larger than or equal to n, the square frame is used for obtaining the root mean square value h of the gray value E Replacing the gray value of the center point E of the rectangular frame; if K<n, inputting each pixel point of the rectangular frame into the step C; C. carrying out Gaussian smoothing processing on each input rectangular frame; D. and B, after the calculation of the steps B and C is finished, obtaining a final noise reduction image. The invention can improve the noise reduction effect and ensure the image quality.
Description
Technical Field
The invention relates to the field of image processing, in particular to application in images with different visual angles, and in particular relates to an image noise reduction method based on selective smoothing filtering.
Background
The image of the digital signal is subject to noise interference from the imaging device and the external environment during the digital processing and transmission. The size of noise is a very important factor in measuring image quality, so measuring noise and quickly and accurately filtering noise without affecting the overall performance of the system is an important method to improve image quality. The noise of an image can be classified from a source into system noise, which is noise from the system itself, and environmental noise, which is noise from the external environment. In order to obtain a good quality image, it is often necessary to process and optimize the image by means of an algorithm for image denoising.
One important difficulty in image noise reduction is in the control of smoothness; too low a smoothness can lead to insufficient noise reduction, too high a smoothness can lead to loss of image details, and the smoothness is also unnatural in vision; because the balance between insufficient noise reduction and excessive smoothing is not well mastered, and the judgment of the noise reduction quality varies from person to person, most noise reduction filtering often selectively corrects the noise reduction result by allowing a user to adjust parameters. Therefore, how to control the smoothness is an important issue for selective image noise reduction.
In the related art, when noise reduction processing is performed on an image, the same filtering is generally used for all the images, that is, a portion of the image where there is no noise is processed, so that the image where there is no noise is also disturbed by the filtering, and image quality is affected.
Therefore, how to selectively process noise contained in an image, select corresponding fast and accurate smooth filtering for different noises, and improve the efficiency of noise reduction processing of the image is a technical problem to be solved.
Disclosure of Invention
The invention aims to provide an image noise reduction method based on selective smooth filtering, which can pointedly select the optimal area to process image noise and avoid the phenomenon of image blurring caused by comprehensively filtering the whole image by using a filtering algorithm; meanwhile, the smooth filtering degree can be accurately controlled by selecting the smooth filtering, the noise reduction effect is effectively improved, and the quality of the image is ensured.
The technical scheme of the invention is as follows:
an image denoising method based on selective smoothing filtering, comprising the steps of:
A. inputting an image to be noise reduced, converting the image to be noise reduced into a gray image, and establishing a 3*3 rectangular frame as follows:
| A | B | C |
| D | E | F |
| G | H | I |
starting from any position of an image to be denoised, moving and sweeping the whole image to be denoised by a rectangular frame, wherein the moving distance is one pixel position each time, recording the gray value of each pixel in the rectangular frame after each moving, and respectively making differences between the gray value of the pixel in the center of the rectangular frame and the gray value of each pixel on the edge of the rectangular frame to obtain the gradient t of each edge pixel of the rectangular frame at the position; b, setting a noise threshold value x, and for each rectangular frame, if |t| is not more than x, representing that the pixel points in the rectangular frame range are all normal points, and inputting the pixel points in the rectangular frame into the step B; otherwise, representing that the pixel points in the rectangular frame range have noise points, and inputting the pixel points in the rectangular frame into the step C;
B. setting an edge detection threshold n, and inputting gray values of all pixel points of all rectangular frames input in the step A into the following formula to calculate a threshold K;
if K is larger than or equal to n, the gray value of each edge pixel point of the rectangular frame is brought into the following formula (2) to obtain a root mean square value h E The root mean square value h E Replacing the gray value of the center point E of the rectangular frame; if K<n, inputting each pixel point of the rectangular frame into the step C;
C. carrying out Gaussian smoothing treatment on each rectangular frame input in the step A or the step B;
D. and B, after the calculation of the steps B and C is finished, obtaining a final noise reduction image.
The setting of the threshold value x comprises the following steps:
based on the rectangular frame established in the step A, starting from any position of the image to be denoised, moving the rectangular frame to sweep the whole image to be denoised, wherein each moving distance is a pixel position, recording the gray value of each pixel in the rectangular frame after each moving, performing Gaussian smoothing on each rectangular frame, and respectively making difference between the gray value of each pixel in the center of each rectangular frame after Gaussian smoothing and the gray value of each pixel on the edge of the rectangular frame to obtain the gradient t of each edge pixel in each rectangular frame 0 Selecting |t 0 Maximum value of #, and the maximum value is set as a threshold value x.
The formula of Gaussian smoothing is as follows:
wherein σ represents the weight per unit distance; x and y represent template coordinates of the center pixel point E.
The setting of the threshold value n comprises the following steps:
averaging gray values of all pixels of an original imageThen bringing the value and the threshold value x into the following formula (4) together for solving, wherein the calculated result is the threshold value n;
according to the invention, through carrying out rectangular frame sweeping segmentation on the image to be noise-reduced, carrying out specific selection on the segmented pixel point areas and carrying out specific filtering treatment, the filtering treatment is carried out on different types of noise in a targeted manner, the image blurring caused by filtering all the images is avoided, the edge is highlighted by the root mean square noise reduction smoothness of the edge contour, and the quality of the image is ensured while the noise reduction effect is ensured.
Drawings
FIG. 1 is an original picture of a plastic part before noise reduction;
fig. 2 is a photograph of a plastic part after noise reduction.
Detailed Description
The invention is described in detail below with reference to the drawings and examples.
Example 1
An image denoising method based on selective smoothing filtering, comprising the steps of:
A. inputting an image to be noise reduced, converting the image to be noise reduced into a gray image, and establishing a 3*3 rectangular frame as follows:
| A | B | C |
| D | E | F |
| G | H | I |
starting from any position of an image to be denoised, moving and sweeping the whole image to be denoised by a rectangular frame, wherein the moving distance is one pixel position each time, recording the gray value of each pixel in the rectangular frame after each moving, and respectively making differences between the gray value of the pixel in the center of the rectangular frame and the gray value of each pixel on the edge of the rectangular frame to obtain the gradient t of each edge pixel of the rectangular frame at the position;
the noise threshold value x is set, and the specific process is as follows: based on the rectangular frame established in the step A, starting from any position of the image to be denoised, moving the rectangular frame to sweep the whole image to be denoised, wherein each moving distance is a pixel position, recording the gray value of each pixel in the rectangular frame after each moving, performing Gaussian smoothing on each rectangular frame, and respectively making difference between the gray value of each pixel in the center of each rectangular frame after Gaussian smoothing and the gray value of each pixel on the edge of the rectangular frame to obtain the gradient t of each edge pixel in each rectangular frame 0 Selecting |t 0 Maximum value of #, and the maximum value is set as a threshold value x.
B, if the t is less than or equal to x, representing that the pixel points in the rectangular frame range are normal points, and inputting the pixel points in the rectangular frame into the step B; otherwise, representing that the pixel points in the rectangular frame range have noise points, and inputting the pixel points in the rectangular frame into the step D;
B. setting an edge detection threshold n, and inputting gray values of all pixel points of all rectangular frames input in the step A into the following formula to calculate a threshold K;
the setting of the threshold value n comprises the following steps:
averaging gray values of all pixels of an original imageThen bringing the value and the threshold value x into the following formula (4) together for solving, wherein the calculated result is the threshold value n;
if K is more than or equal to n,the gray value of each edge pixel point of the rectangular frame is brought into the following formula (2) to obtain a root mean square value h E The root mean square value h E Replacing the gray value of the center point E of the rectangular frame; if K<n, inputting each pixel point of the rectangular frame into the step C;
C. carrying out Gaussian smoothing treatment on each rectangular frame input in the step A or the step B;
the formula of gaussian smoothing is:
wherein σ represents the weight per unit distance; x and y represent template coordinates of the center pixel point E.
D. And B, after the calculation of the steps B and C is finished, obtaining a final noise reduction image.
The setting of the threshold value n comprises the following steps:
averaging gray values of all pixels of an original imageAnd then bringing the value and the threshold value x into the following formula (4) to solve, wherein the calculated result is the threshold value n.
Example 2
The method of example 1 was used to make noise reduction on a plastic part picture, and specific results are shown in fig. 1 and 2, where fig. 1 is an original picture before noise reduction of the part, and fig. 2 is a picture after noise reduction.
As can be seen from comparison between fig. 1 and fig. 2, after the noise reduction treatment in fig. 1, the noise point is basically disappeared, and the image is clearer.
Claims (4)
1. An image noise reduction method based on selective smoothing filtering, which is characterized by comprising the following steps:
A. inputting an image to be noise reduced, converting the image to be noise reduced into a gray image, and establishing a 3*3 rectangular frame as follows:
starting from any position of an image to be denoised, moving and sweeping the whole image to be denoised by a rectangular frame, wherein the moving distance is one pixel position each time, recording the gray value of each pixel in the rectangular frame after each moving, and respectively making differences between the gray value of the pixel in the center of the rectangular frame and the gray value of each pixel on the edge of the rectangular frame to obtain the gradient t of each edge pixel of the rectangular frame at the position; b, setting a noise threshold value x, and for each rectangular frame, if |t| is not more than x, representing that the pixel points in the rectangular frame range are all normal points, and inputting the pixel points in the rectangular frame into the step B; otherwise, representing that the pixel points in the rectangular frame range have noise points, and inputting the pixel points in the rectangular frame into the step C;
B. setting an edge detection threshold n, and inputting gray values of all pixel points of all rectangular frames input in the step A into the following formula to calculate a threshold K;
if K is larger than or equal to n, the gray value of each edge pixel point of the rectangular frame is brought into the following formula (2) to obtain a root mean square value h E The root mean square value h E Replacing the gray value of the center point E of the rectangular frame; if K<n, inputting each pixel point of the rectangular frame into the step C;
C. carrying out Gaussian smoothing treatment on each rectangular frame input in the step A or the step B;
D. and B, after the calculation of the steps B and C is finished, obtaining a final noise reduction image.
2. The method for noise reduction of an image based on selective smoothing filtering as claimed in claim 1, wherein said setting of the threshold x comprises the steps of:
based on the rectangular frame established in the step A, starting from any position of the image to be denoised, moving the rectangular frame to sweep the whole image to be denoised, wherein each moving distance is a pixel position, recording the gray value of each pixel in the rectangular frame after each moving, performing Gaussian smoothing on each rectangular frame, and respectively making difference between the gray value of each pixel in the center of each rectangular frame after Gaussian smoothing and the gray value of each pixel on the edge of the rectangular frame to obtain the gradient t of each edge pixel in each rectangular frame 0 Selecting |t 0 Maximum value of #, and the maximum value is set as a threshold value x.
3. The selective smoothing filter-based image noise reduction method according to claim 1 or 2, characterized in that:
the formula of Gaussian smoothing is as follows:
wherein σ represents the weight per unit distance; x and y represent template coordinates of the center pixel point E.
4. The method for noise reduction of an image based on selective smoothing filtering as claimed in claim 2, wherein the setting of the threshold n comprises the steps of:
averaging gray values of all pixels of an original imageThen bringing the value and the threshold value x into the following formula (4) together for solving, wherein the calculated result is the threshold value n;
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310210618.XA CN116402705A (en) | 2023-03-07 | 2023-03-07 | Image noise reduction method based on selective smoothing filtering |
| CN202310210618X | 2023-03-07 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116977207A true CN116977207A (en) | 2023-10-31 |
Family
ID=87016894
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310210618.XA Withdrawn CN116402705A (en) | 2023-03-07 | 2023-03-07 | Image noise reduction method based on selective smoothing filtering |
| CN202310809660.3A Pending CN116977207A (en) | 2023-03-07 | 2023-07-04 | Image noise reduction method based on selective smoothing filtering |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310210618.XA Withdrawn CN116402705A (en) | 2023-03-07 | 2023-03-07 | Image noise reduction method based on selective smoothing filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN116402705A (en) |
-
2023
- 2023-03-07 CN CN202310210618.XA patent/CN116402705A/en not_active Withdrawn
- 2023-07-04 CN CN202310809660.3A patent/CN116977207A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| CN116402705A (en) | 2023-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112819772B (en) | High-precision rapid pattern detection and recognition method | |
| CN108921800B (en) | Non-local mean denoising method based on shape self-adaptive search window | |
| CN109785245B (en) | Light spot image trimming method | |
| CN109377450B (en) | Edge protection denoising method | |
| CN105046655B (en) | A kind of automatic sharpening method of video image and device | |
| CN105913396A (en) | Noise estimation-based image edge preservation mixed de-noising method | |
| CN103369209A (en) | Video noise reduction device and video noise reduction method | |
| CN109003233B (en) | Image denoising method based on self-adaptive weight total variation model | |
| CN107292834B (en) | Infrared image detail enhancement method | |
| CN102298774B (en) | Non-local mean denoising method based on joint similarity | |
| CN108280836B (en) | Image processing method and device | |
| CN116468641A (en) | Infrared blood vessel image enhancement processing method | |
| CN114418904A (en) | Infrared image enhancement method based on improved histogram equalization and enhanced high-pass filtering | |
| CN107025640A (en) | The Medical Image Denoising method of non-local mean based on optimization | |
| CN109767407B (en) | Secondary estimation method for atmospheric transmissivity image in defogging process | |
| CN115424107A (en) | Underwater pier apparent disease detection method based on image fusion and deep learning | |
| CN113432723B (en) | Image processing method, system and computer system for weakening stray radiation | |
| CN113610734B (en) | Image edge processing method based on guided filtering and application | |
| WO2017088391A1 (en) | Method and apparatus for video denoising and detail enhancement | |
| CN113450272B (en) | Image enhancement method based on sinusoidal variation and application thereof | |
| CN106023097A (en) | Iterative-method-based flow field image preprocessing algorithm | |
| CN116977207A (en) | Image noise reduction method based on selective smoothing filtering | |
| CN107945196B (en) | Image edge detection method for aerial photography correction image splicing | |
| CN113379631B (en) | Image defogging method and device | |
| CN114519675A (en) | Image processing method and device, electronic equipment and readable storage medium |
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 |