CN102075669A - Method and system for adaptively recursively denoising digital video signal - Google Patents
Method and system for adaptively recursively denoising digital video signal Download PDFInfo
- Publication number
- CN102075669A CN102075669A CN2009101992662A CN200910199266A CN102075669A CN 102075669 A CN102075669 A CN 102075669A CN 2009101992662 A CN2009101992662 A CN 2009101992662A CN 200910199266 A CN200910199266 A CN 200910199266A CN 102075669 A CN102075669 A CN 102075669A
- Authority
- CN
- China
- Prior art keywords
- diff
- image
- adj
- bwd
- 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
Images
Abstract
The invention discloses a method and a system for adaptively recursively denoising a digital video signal. Key technology is that: a recursion coefficient k set by a user is corrected to obtain a corrected coefficient k_adj; if an absolute difference value diff_fwd between corresponding pixels in a current image t and a previous image t-1 is more than or equal to the absolute difference value diff_bwd between the corresponding pixels in the current image t and the next image t+1, k_adj is equal to k; if the absolute difference value diff_fwd is less than the absolute difference value diff_bwd, the corrected coefficient k_adj is less than k; and the pixel G (i, j, t) in the current image t is denoised by adopting the corrected coefficient k_adj. In the method and the system, the information of the image t+1 after a current time is introduced to correct the recursion coefficient k. The method and the system provided by the invention can well eliminate the trailing phenomenon, produced by recursive denoising, of a moving object.
Description
Technical field
The invention belongs to the digital video signal processing field, specifically, relate to a kind of self adaptation recursive noise reduction method and system of digital video signal.
Background technology
No matter digital video signal still is the digital form transmission by analog form, all needs to carry out the denoising sound and handle when finally being presented on the screen, to reach level and smooth clean display effect.At present the most frequently used denoising sound method is that signal to input carries out the recursive operation on the time shaft in digital video signal processing, for example, if a pixel value on the t-1 moment inputted video image is G (x, y, t-1), wherein x and y represent the position of this pixel in image G, and the pixel on the t moment same position is G (x, y, t), so to G (x, y t) carry out following computing:
G (x, y, t)=(1-k) * G (x, y, t)+k*G (x, y, t-1) (1) 0≤k≤1.0 wherein, (t) numerical value replaces old numerical value for x, y to utilize the G that newly calculates.Such recursive operation constantly go on occur in G (x, y) noise on the location of pixels will present the process of a convergence, and k is fast more near 1.0 convergence processs more, the effect of denoising sound is also good more.
The denoising sound method of introducing has above obtained using widely in commercial at present Digital Video Processing chip.But this method has a problem, is exactly that it can only be applied to static video image in theory, i.e. pixel G (x, y is t-1) with pixel G (x, y, t) should be same pixel in the original image, that is to say that this pixel is not moved to moment t from moment t-1.If image changes, for example constantly pixel G of t-1 (x, y is t-1) on the object in image, arrive t this movement of objects of the moment and arrived other places, (x, y t) no longer are original G (x to pixel G, y, t-1), so-called problem of trailing will appear in the computing of still carrying out so in this case in (1), and promptly a track can appear in the back of mobile object.At this problem; common way is the static judgement classification someway with pixel; the big more k value of static more use is to reach strong more denoising audio fruit; otherwise not static more then uses more little k value to weaken the effect of denoising sound; but this method greatly depends on the accuracy to static judgement classification; and the corresponding k value of the static judged result of different stage choose usually also relevantly with different concrete picture materials, it is much difficult that these uncertainties produce the application of this method.
Summary of the invention
The object of the present invention is to provide a kind of self adaptation recursive noise reduction method and system of digital video signal, to solve the technical problem that existing recursive noise reduction method and system is too dependent on the actual noise reduction process application difficult that the accuracy of static judgement classification is produced.
In order to achieve the above object, technical scheme of the present invention is as follows:
A kind of self adaptation recursive noise reduction method of digital video signal comprises the steps: to store the step of the video flowing of input; Calculate the step that produces the absolute difference numerical value diff_bwd of respective pixel among present image t and its previous image t-1; Calculate to produce present image t and the step of the absolute difference numerical value diff_fwd of respective pixel among the image t+1 thereafter; The recursion coefficient k that the user is provided with revises to obtain the step of revised coefficient k _ adj, wherein as diff_fwd during more than or equal to diff_bwd, and k_adj=k; Otherwise diff_fwd compares more little with diff_bwd, and it is more little that then revised coefficient k _ adj compares k; (i, j t) carry out the step of noise reduction to the pixel G of present image t to adopt revised coefficient k _ adj.
Accordingly, a kind of self adaptation recursive noise reduction system of digital video signal comprises: the device of the video flowing of storage input; Calculate the device that produces the absolute difference numerical value diff_bwd of respective pixel among present image t and its previous image t-1; Calculate to produce present image t and the device of the absolute difference numerical value diff_fwd of respective pixel among the image t+1 thereafter; The recursion coefficient k that the user is provided with revises to obtain the device of revised coefficient k _ adj, wherein as diff_fwd during more than or equal to diff_bwd, and k_adj=k; Otherwise diff_fwd compares more little with diff_bwd, and it is more little that then revised coefficient k _ adj compares k; (i, j t) carry out the device of noise reduction to the pixel G of present image t to adopt revised coefficient k _ adj.
Accordingly, a kind of self adaptation recursive noise reduction system of digital video signal comprises: the image buffer of the video flowing of storage input; The absolute difference computation unit receives from respective pixel among the present image t in the described image buffer and its previous image t-1 backward, calculates the absolute difference numerical value diff_bwd that produces both; Absolute difference computation unit forward receives from present image t in the described image buffer and respective pixel among the image t+1 thereafter, calculates both absolute difference numerical value diff_bwd of generation; The recursion coefficient amending unit receives the recursion coefficient k and described absolute difference numerical value diff_bwd, the diff_bwd that are provided with from the user; As diff_fwd during more than or equal to diff_bwd, revised coefficient k _ adj=k; Otherwise diff_fwd compares more little with diff_bwd, and it is more little that then revised coefficient k _ adj compares k; The noise reduction arithmetic element, (i, j t) carry out noise reduction process to the pixel G of present image t to receive respective pixel among revised coefficient k _ adj and present image t and its previous image t-1.
Preferably, as diff_fwd during less than diff_bwd, k_adj=k*diff_fwd/diff_bwd.Correction algorithm herein gets final product as long as satisfy " diff_fwd compares more little with diff_bwd, it is more little that then revised coefficient k _ adj compares k ".
When described diff_bwd is less than or equal to the noise threshold that the user is provided with, make G (i, j, t)=(1-k_adj) * G (i, j, t)+k_adj*G (i, j, t-1); Otherwise maintenance G (i, j is t) constant.
In traditional recursive noise reduction method,, do not use the data among the image t+1 to only used information when the operation of pre-treatment image t from image t-1 and image t itself.In the present invention, we have introduced the piece image information after the current time, and promptly image t+1 carries out the correction of recursion coefficient k.Utilize method and system among the present invention can well eliminate the conditions of streaking of the moving object that produces in the recursive noise reduction.
Description of drawings
Fig. 1 is a self adaptation recursive noise reduction system block diagram of the present invention.
Embodiment
According to Fig. 1, provide preferred embodiment of the present invention, and described in detail below, enable to understand better function of the present invention, characteristics.
Fig. 1 is a self adaptation recursive noise reduction system block diagram of the present invention, has described the self adaptation recursive noise reduction flow process among the present invention.Here be noted that for the image among input video stream Fig. 1 line by line to be exactly the continuous progressive image of input, as image t-1, image t and image t+1 representative be exactly continuous three width of cloth progressive images.But, for the image among input video stream Fig. 1 of interlacing then corresponding to being all strange or being all the continuous interlaced video image of idol.Because strange field picture and even field picture are alternately to transmit in the input video stream of interlacing, so the data that will store in the image buffer can be than many one times shown in Fig. 1.For example, if visual t is a strange field picture in the interlacing input video stream, then image t-1 and image t+1 are just corresponding to its a previous strange field picture and a back strange field picture, also have width of cloth idol field picture between the continuous strange field picture of per two width of cloth, these even field picture being stored among Fig. 1 in image buffer do not drawn.Similarly, if visual t representative is an even field picture in the interlacing input video stream, image t-1 and image t+1 are just corresponding to its a previous even field picture and a back even field picture so, and also having the strange field picture of a width of cloth between the continuous even field picture of per two width of cloth, these strange field picture being stored among Fig. 1 in image buffer do not drawn yet.In a word, if image t is strange (idol) field picture in the interlacing input video stream, then image t-1 among Fig. 1 and image t+1 just represent two strange (idol) field picture that its front and back are adjacent.
In traditional recursive noise reduction method,, do not use the data among the image t+1 to only used information when the operation of pre-treatment image t from image t-1 and image t itself.That is to say that traditional recursive noise reduction method is only used the information before the current time, and do not use the information after the current time.And in the present invention, we have introduced the piece image information after the current time, and promptly image t+1 carries out the correction of recursion coefficient k.
At first, the absolute difference computation unit is in order to calculate the absolute difference numerical value that produces respective pixel among image t-1 and the image t, promptly backward
diff_bwd＝|G(i，j，t-1)-G(i，j，t)|
G (i, j, the t-1) pixel value of the capable j of i row among the presentation video t-1, G (i, j, t) pixel value of the capable j row of i among the presentation video t wherein.Similarly, the absolute difference computation unit is in order to calculate the absolute difference numerical value that produces respective pixel among image t+1 and the image t, promptly forward
diff_fwd＝|G(i，j，t+1)-G(i，j，t)|
In traditional recursive noise reduction method not to the use of diff_fwd.Then, diff_bwd and diff_fwd are used in the recursion coefficient amending unit recursion coefficient k that the user is provided with is revised.The principle of revising be " diff_fwd compare with diff_bwd more little then for G (t) operation of the noise reduction of pixel is just weak more for i, j ".Revise the principle user based on this and can design different concrete modification methods voluntarily.The invention provides a kind of modification method as an example:
From this example as can be seen, as diff_fwd during more than or equal to diff_bwd, it is inoperative to revise operation, recursion coefficient k does not change, and as diff_fwd during less than diff_bwd, utilize the ratio of diff_fwd and diff_bwd that recursion coefficient k is diminished, meet top said correction principle.
In addition, when we will notice that diff_bwd is zero here, diff_fwd necessarily also is zero, and we can not do noise reduction process simply in this case, because diff_bwd is null representation G (i, j, t-1) equal G (i, j, t), noise reduction is also inoperative as can be seen from following noise reduction recursive operation.
Revised recursion coefficient k_adj be used in the noise reduction arithmetic element, carry out G (i, j, t) the noise reduction computing of pixel:
Wherein NRT is the noise threshold that the user is provided with, and noise magnitude is stopped greater than the situation noise reduction operation of this threshold value.Utilize method among the present invention can well eliminate the conditions of streaking of the moving object that produces in the recursive noise reduction.
At last, this denoising sound circuit all is suitable for for the luminance signal and the chrominance signal of video image.
The front provides the description to preferred embodiment, so that any technical staff in this area can use or utilize the present invention.To this preferred embodiment, those skilled in the art can make various modifications or conversion on the basis that does not break away from the principle of the invention.Should be appreciated that these modifications or conversion do not break away from protection scope of the present invention.
Claims (13)
1. the self adaptation recursive noise reduction method of a digital video signal comprises the steps:
The step of the video flowing of storage input;
Calculate the step that produces the absolute difference numerical value diff_bwd of respective pixel among present image t and its previous image t-1;
Calculate to produce present image t and the step of the absolute difference numerical value diff_fwd of respective pixel among the image t+1 thereafter;
The recursion coefficient k that the user is provided with revises to obtain the step of revised coefficient k _ adj, wherein as diff_fwd during more than or equal to diff_bwd, and k_adj=k; Otherwise diff_fwd compares more little with diff_bwd, and it is more little that then revised coefficient k _ adj compares k;
(i, j t) carry out the step of noise reduction to the pixel G of present image t to adopt revised coefficient k _ adj.
2. the self adaptation recursive noise reduction method of digital video signal as claimed in claim 1 is characterized in that, as diff_fwd during less than diff_bwd, and k_adj=k*diff_fwd/diff_bwd.
3. the self adaptation recursive noise reduction method of digital video signal as claimed in claim 1 is characterized in that, the step of described noise reduction is, when described diff_bwd is less than or equal to the noise threshold of user's setting, G (i, j, t)=(1-k_adj) * G (i, j, t)+and k_adj*G (i, j, t-1); Otherwise (i, j's G t) remain unchanged.
4. as the denoising sound method in the described digital video signal processing of arbitrary claim in the claim 1 to 3, it is characterized in that, for input video stream line by line, described present image t and its previous image t-1 and thereafter an image t+1 be the continuous progressive image of input.
5. as the denoising sound method in the described digital video signal processing of arbitrary claim in the claim 1 to 3, it is characterized in that, for the input video stream of interlacing, described present image t and its previous image t-1 and thereafter an image t+1 be corresponding to being all strange or be all the continuous interlaced video image of idol.
6. as the denoising sound method in the described digital video signal processing of arbitrary claim in the claim 1 to 3, it is characterized in that described pixel is a luminance signal.
7. as the denoising sound method in the described digital video signal processing of arbitrary claim in the claim 1 to 3, it is characterized in that described pixel is a chrominance signal.
8. the self adaptation recursive noise reduction system of a digital video signal comprises:
The device of the video flowing of storage input;
Calculate the device that produces the absolute difference numerical value diffb_wd of respective pixel among present image t and its previous image t-1;
Calculate to produce present image t and the device of the absolute difference numerical value diff_fwd of respective pixel among the image t+1 thereafter;
The recursion coefficient k that the user is provided with revises to obtain the device of revised coefficient k _ adj, wherein as diff_fwd during more than or equal to diff_bwd, and k_adj=k; Otherwise diff_fwd compares more little with diff_bwd, and it is more little that then revised coefficient k _ adj compares k;
(i, j t) carry out the device of noise reduction to the pixel G of present image t to adopt revised coefficient k _ adj.
9. the self adaptation recursive noise reduction system of digital video signal as claimed in claim 8 is characterized in that, as diff_fwd during less than diff_bwd, and k_adj=k*diff_fwd/diff_bwd.
10. the self adaptation recursive noise reduction system of digital video signal as claimed in claim 8 is characterized in that, when described diff_bwd was less than or equal to the noise threshold of user's setting, the device of described noise reduction made G (i, j, t)=(1-k_adj) * G (i, j, t)+and k_adj*G (i, j, t-1); Otherwise the device of described noise reduction maintenance G (i, j is t) constant.
11. the self adaptation recursive noise reduction system of a digital video signal comprises:
The image buffer of the video flowing of storage input;
The absolute difference computation unit receives from respective pixel among the present image t in the described image buffer and its previous image t-1 backward, calculates the absolute difference numerical value diff_bwd that produces both;
Absolute difference computation unit forward receives from present image t in the described image buffer and respective pixel among the image t+1 thereafter, calculates both absolute difference numerical value diff_bwd of generation;
The recursion coefficient amending unit receives the recursion coefficient k and described absolute difference numerical value diff_bwd, the diff_bwd that are provided with from the user; As diff_fwd during more than or equal to diff_bwd, revised coefficient k _ adj=k; Otherwise diff_fwd compares more little with diff_bwd, and it is more little that then revised coefficient k _ adj compares k;
The noise reduction arithmetic element, (i, j t) carry out noise reduction process to the pixel G of present image t to receive respective pixel among revised coefficient k _ adj and present image t and its previous image t-1.
12. the self adaptation recursive noise reduction system of digital video signal as claimed in claim 11 is characterized in that, as diff_fwd during less than diff_bwd, and k_adj=k*diff_fwd/diff_bwd.
13. the self adaptation recursive noise reduction system of digital video signal as claimed in claim 11, it is characterized in that, when described diff_bwd is less than or equal to the noise threshold of user's setting, described noise reduction noise reduction arithmetic element makes G (i, j, t)=(1-k_adj) * G (i, j, t)+and k_adj*G (i, j, t-1); Otherwise described noise reduction arithmetic element maintenance G (i, j is t) constant.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101992662A CN102075669A (en) | 2009-11-24 | 2009-11-24 | Method and system for adaptively recursively denoising digital video signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101992662A CN102075669A (en) | 2009-11-24 | 2009-11-24 | Method and system for adaptively recursively denoising digital video signal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102075669A true CN102075669A (en) | 2011-05-25 |
Family
ID=44033999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009101992662A Pending CN102075669A (en) | 2009-11-24 | 2009-11-24 | Method and system for adaptively recursively denoising digital video signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102075669A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104333716A (en) * | 2014-11-06 | 2015-02-04 | 南华大学 | Anti-radiation video image system and method |
| CN105872310A (en) * | 2016-04-20 | 2016-08-17 | 上海联影医疗科技有限公司 | Image motion detection method and image noise reduction method for movable imaging equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030161516A1 (en) * | 2002-02-22 | 2003-08-28 | Pieter Vuylsteke | Method for enhancing the contrast of an image |
| CN1595433A (en) * | 2004-06-25 | 2005-03-16 | 东软飞利浦医疗设备系统有限责任公司 | Recursion denoising method based on motion detecting image |
| CN1607834A (en) * | 2003-10-14 | 2005-04-20 | 法国耐思讯公司 | Device and method for estimating the noise of a video signal |
| US20080106642A1 (en) * | 2006-11-08 | 2008-05-08 | Sujith Srinivasan | Advanced deinterlacer for high-definition and standard-definition video |
| CN101448077A (en) * | 2008-12-26 | 2009-06-03 | 四川虹微技术有限公司 | Self-adapting video image 3D denoise method |
-
2009
- 2009-11-24 CN CN2009101992662A patent/CN102075669A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030161516A1 (en) * | 2002-02-22 | 2003-08-28 | Pieter Vuylsteke | Method for enhancing the contrast of an image |
| CN1607834A (en) * | 2003-10-14 | 2005-04-20 | 法国耐思讯公司 | Device and method for estimating the noise of a video signal |
| CN1595433A (en) * | 2004-06-25 | 2005-03-16 | 东软飞利浦医疗设备系统有限责任公司 | Recursion denoising method based on motion detecting image |
| US20080106642A1 (en) * | 2006-11-08 | 2008-05-08 | Sujith Srinivasan | Advanced deinterlacer for high-definition and standard-definition video |
| CN101448077A (en) * | 2008-12-26 | 2009-06-03 | 四川虹微技术有限公司 | Self-adapting video image 3D denoise method |
Non-Patent Citations (1)
| Title |
|---|
| 李岩等: "基于运动补偿的自适应时域视频降噪算法研究", 《半导体光电》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104333716A (en) * | 2014-11-06 | 2015-02-04 | 南华大学 | Anti-radiation video image system and method |
| CN104333716B (en) * | 2014-11-06 | 2017-07-28 | 南华大学 | Radiation-resistant video image system and method |
| CN105872310A (en) * | 2016-04-20 | 2016-08-17 | 上海联影医疗科技有限公司 | Image motion detection method and image noise reduction method for movable imaging equipment |
| CN105872310B (en) * | 2016-04-20 | 2020-03-17 | 上海联影医疗科技有限公司 | Image motion detection method and image noise reduction method for movable imaging equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103402098B (en) | A kind of video frame interpolation method based on image interpolation | |
| US8773595B2 (en) | Image processing | |
| CN102637293B (en) | Moving image processing device and moving image processing method | |
| JP2008507909A (en) | Method and apparatus for motion vector processing | |
| CN101207707A (en) | System and method for advancing frame frequency based on motion compensation | |
| KR101643026B1 (en) | Method and device for calculating blur in video images | |
| CN102082896B (en) | Method for treating video of liquid crystal display device | |
| US20100128181A1 (en) | Seam Based Scaling of Video Content | |
| CN104717509B (en) | A kind of video encoding/decoding method and device | |
| JP2008125112A (en) | Visual processing apparatus, display apparatus, visual processing method, program and integrated circuit | |
| CN104142730B (en) | A kind of method that gesture tracking result is mapped to mouse event | |
| CN102447870A (en) | Detection method for static objects and motion compensation device | |
| CN102075669A (en) | Method and system for adaptively recursively denoising digital video signal | |
| CN111556244B (en) | Video style migration method and device | |
| US8265419B2 (en) | Image processing apparatus and image processing method | |
| CN106101716B (en) | A kind of video frame rate upconversion method | |
| CN108830146A (en) | A kind of uncompressed domain lens boundary detection method based on sliding window | |
| CN104715456A (en) | Image defogging method | |
| US8090210B2 (en) | Recursive 3D super precision method for smoothly changing area | |
| CN111833262A (en) | Image noise reduction method and device and electronic equipment | |
| CN100594723C (en) | Image processor having frame speed conversion and its method | |
| JP2001245155A (en) | Method for edge correction of image data, and device thereof | |
| CN113727176A (en) | Video motion subtitle detection method | |
| CN108074248B (en) | OSD automatic detection method and device based on image content | |
| CN102158678A (en) | Motion vector field interpolation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110525 |