CN100454338C - Dynamic image processing method - Google Patents
Dynamic image processing method Download PDFInfo
- Publication number
- CN100454338C CN100454338C CNB200610007850XA CN200610007850A CN100454338C CN 100454338 C CN100454338 C CN 100454338C CN B200610007850X A CNB200610007850X A CN B200610007850XA CN 200610007850 A CN200610007850 A CN 200610007850A CN 100454338 C CN100454338 C CN 100454338C
- Authority
- CN
- China
- Prior art keywords
- block
- sad
- computing
- picture
- macro zone
- 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.)
- Expired - Fee Related
Links
Images
Abstract
A dynamic image processing method includes receiving all the pictures in dynamic image by the order, and dividing each picture into many macro blocks. Then, one micro block of the present screen is defined as a ratio block, in which, the pixels and that of block in the former picture search window are processed SAD operation, to access a SAD critical value. Then the invention takes SAD operation for pixels of the ratio block and that without SAD operation in micro block, and acquires the accumulated value of many SAD values. When the accumulated value is higher than the critical value, SAD operation is immediately halted.
Description
Technical field
The invention relates to a kind of disposal route of dynamic image, and particularly relevant for a kind of disposal route of utilizing the dynamic image of absolute difference sum (hereinafter to be referred as SAD).
Background technology
Dynamic image is combined by the picture of many static state (Frame).When many still images switch in regular turn fast, because the principle of the persistence of vision of human eye just looks like that image is in motion.
From the obtained digital video stream of digital camera, usually comprised the motion of in continuous pictures, not expecting (Unwanted Motion).The so-called motion of not expecting may be because camera rocking and produce when the acquisition video signal.In order to eliminate this motion of not expecting, therefore will use the solution of optics, machinery or other electronics, detect and proofread and correct the motion that these are not expected.Yet these methods are all very expensive and complicated.In addition, these methods can not solve because the motion of not expecting in the dynamic image that special-effect the produced stream.In order to stablize dynamic image stream, therefore just developed the disposal route of many numerals.
Wherein a kind of disposal route utilizes the calculating of absolute value difference to handle the method for dynamic image exactly.At first the summation that takes absolute value of two image subtractions that acquisition is come in is exactly a sad value, find out the mutual connection of two images then according to minimum sad value, just find out the movement locus of two images, and utilize this movement locus to come these two images are carried out Flame Image Process.
Yet known SAD operational method need be carried out computing with each pixel in two images, and therefore can expend a large amount of time is carrying out computing.
Summary of the invention
Therefore, the invention provides a kind of disposal route of dynamic image, can handle dynamic image fast, and cost that needn't consumes expensive.
For addressing the above problem, the invention provides a kind of disposal route of dynamic image, and this dynamic image is made up of several continuous pictures, and each those picture comprises several pixels, it is characterized in that this disposal route comprises:
Receive those pictures in regular turn;
Each those picture is divided into most macro zone blocks, and each macro zone block has the pixel of one first preset number;
One of them is defined as a comparison block with those macro zone blocks in the present picture, and will compare the interior pixel of block, carry out the computing of absolute difference sum with the pixel in the macro zone block in the search window in the previous picture, and obtain a SAD critical value, the macro zone block that wherein comprises one second preset number in this search window, and this comparison block meeting is carried out the computing of absolute difference sum to the macro zone block of the image border in this search window earlier;
With the pixel in this comparison block, carry out the computing of absolute difference sum with the pixel in the macro zone block that did not also carry out the computing of absolute difference sum in this search window, and obtain an operation values, the gained that adds up of a sad value that is produced when wherein this operation values is carried out the computing of absolute difference sum for each those pixel in the macro zone block that did not also carry out the computing of absolute difference sum in this search window;
When the sad value that adds up during, then interrupt the computing of absolute difference sum immediately more than or equal to this SAD critical value; And
When the accumulated value of the sad value of all pixels in the macro zone block that carries out the computing of absolute difference sum during, then this SAD critical value is updated to and carries out the operation values that the computing of absolute difference sum is obtained at present less than this SAD critical value.
The disposal route of dynamic image provided by the invention can be handled dynamic image.Wherein, dynamic image is made up of most continuous pictures, and each picture has comprised most pixels.Disposal route of the present invention comprises and receives all pictures in regular turn, and each picture is divided into most macro zone blocks, and each macro zone block has the pixel of a preset number.In addition, the present invention can be defined as one of them macro zone block in the present picture one comparison block, and with the pixel in this comparison block, carry out the SAD computing with the pixel in the macro zone block in the search window in the previous picture, and obtain a SAD critical value, wherein comprised the macro zone block of a preset number in the search window.In addition, the present invention will compare the pixel in the block, with the pixel SAD computing in the macro zone block that did not also carry out the SAD computing in the search window, and obtain an operation values.Wherein, this operation values is the gained that adds up of the sad value that produced when each pixel is carried out the SAD computing in the macro zone block.When the sad value that adds up during more than or equal to the SAD critical value, then interrupt the SAD computing immediately, and when the sad value of all pixels in the macro zone block that carries out the SAD computing gets accumulated value less than the SAD critical value, then the SAD critical value is updated to and carries out the operation values that the SAD computing is obtained at present.
In an embodiment of the present invention, the comparison block can carry out the SAD computing earlier to image edge pixels in the macro zone block in the search window earlier, and obtains a SAD critical value.
In addition, the present invention more comprises the motion vector of obtaining the comparison block in each picture according to the SAD critical value, and the motion vector of the comparison block in each picture and obtain the movement locus of these continuous pictures.In addition, the present invention can judge in each picture that whether the motion vector of comparison block and the angle of movement locus are greater than a predetermined angle.The motion vector of the comparison block in picture and the angle of movement locus are then adjusted the parameter of picture greater than this predetermined angle, so that the motion vector of comparison block levels off to movement locus, and make that movement locus can be level and smooth.
Under preferable situation, above-mentioned predetermined angle can be set voluntarily by a user.
Because the present invention is when carrying out the SAD computing to the macro zone block in the search window,, therefore can not expend the too many time as long as the accumulated value of sad value during greater than the SAD critical value, will interrupt computing immediately.In addition, the present invention also can set a predetermined angle by the user, is used as whether proofreading and correct the foundation of the motion vector of comparing block, so the present invention can flexibly handle because the motion vector that special-effect produced.
In sum, the present invention has following advantage at least:
1. because the present invention adopts the mode of digital processing to handle dynamic image, so the present invention need not expend a large amount of costs.
Since the present invention when carrying out the SAD computing, the comparison block can carry out the SAD computing earlier to image edge pixels in the macro zone block in the search window earlier, and when the comparison block with search window in macro zone block when carrying out the SAD computing as long as just stop the SAD computing at once greater than the SAD critical value, so the present invention can reduce operation time.
3. because meeting of the present invention is carried out the correction of dynamic image according to the motion vector of comparison block and the angle between the movement locus, therefore can make picture seem more natural.
4. the present invention also allows the user to set up above-mentioned predetermined angle on their own, so the present invention has more elasticity.
For above and other objects of the present invention, feature and advantage can be become apparent, preferred embodiment cited below particularly, and cooperate appended accompanying drawing, be described in detail below.
Description of drawings
Fig. 1 is the flow chart of steps according to the disposal route of a kind of dynamic image of one of the present invention preferred embodiment.
Fig. 2 carries out the synoptic diagram of absolute difference sum for previous picture and present picture.
Fig. 3 is the synoptic diagram of a kind of image and search window.
Fig. 4 is the flow chart of steps of the method for the movement locus that calculates picture according to one of the present invention preferred embodiment a kind of and track filtering.
Fig. 5 is the synoptic diagram of the motion vector of comparison block.
The main element symbol description:
201,501: present picture
203,503: previous picture
205,207: macro zone block
301: image
303: search window
505: motion vector
S101, S103, S105, S107, S109, S111, S113, S115: the steps flow chart of the disposal route of dynamic image
S401, S403, S405, S407: the steps flow chart of the movement locus of calculating picture and the method for track filtering
Embodiment
Fig. 1 is the flow chart of steps according to the disposal route of a kind of dynamic image of one of the present invention preferred embodiment.Please refer to Fig. 1, well known, a dynamic video signal is made up of many continuous pictures, and each picture all has most pixels, to constitute the image on each picture.And the disposal route of dynamic image proposed by the invention is to be used for handling in dynamic video signal, because noise or external factor similarly are the situation of hand shake, and has influenced the quality of dynamic image.
Please continue with reference to Fig. 1, at first, the present invention can receive most continuous pictures, and these continuous pictures can constitute a dynamic video signal as described in the step S101.Because each picture all has most pixels, it can be arranged in the array mode.Therefore, the present invention can be divided into each picture most macro zone blocks (Macro Block), as described in step S103.Wherein, the pixel that all comprises one first preset number in each macro zone block.
Then, as described in step S105, the present invention can be defined as the comparison block with in the macro zone block in the present picture (Current Frame) one of them, so that with the pixel in this comparison block, carry out the SAD computing with the pixel in interior first macro zone block of the search window (Searching Window) in the previous picture (Pre-Frame), and obtain a SAD critical value.Well known, comprised the macro zone block of one second preset number in the search window.
Fig. 2 is the synoptic diagram that present picture (Current Frame) and previous picture (Pre-Frame) carry out the absolute difference sum.Suppose 201 for present picture (Current Frame), and 203 be previous picture (Pre-Frame).In this definition, picture 203 is an i picture, and 201 is i-1 picture.Suppose in i-1 picture 201, select macro zone block 205 to be the comparison block.Therefore, the present invention can carry out SAD computing (being assumed to be 207) with a macro zone block of (promptly nine palace lattice in i picture 203 are represented) in the search window in comparison block 205 and i the picture 203, to obtain a SAD critical value.The comparison block can carry out the SAD computing earlier to image edge pixels in the macro zone block in the search window earlier.
Please get back to Fig. 1, when the present invention as described in the step S105, obtain next can carry out S107 after the SAD critical value, will compare the macro zone block that did not carry out the SAD computing in interior pixel of block and the search window exactly and carry out the SAD computing, and obtain an operation values.So-called SAD operation values is that the sad value that each pixel is obtained when carrying out the SAD computing in the macro zone block adds up back resultant.
Specifically, in the present invention, can carry out step S109, exactly when macro zone block carries out the SAD computing, whether the sad value that the meeting real-time judge adds up is greater than above-mentioned SAD critical value.Suppose, when a macro zone block when carrying out the SAD computing, when the present invention judges that the sad value add up is more than or equal to the SAD critical value, then, directly interrupt the SAD computing, and do not need remaining pixel is carried out the SAD computing as described in the step S111.Therefore, the present invention can reduce the time of SAD computing effectively.
Relatively, suppose after the pixel that a macro zone block is all is finished the SAD computing, find the accumulated value of all SAD, when just the SAD operation values of this macro zone block is less than the SAD critical value, meeting then of the present invention is as described in the step S113, and the numerical value of SAD critical value is updated to the operation values that this SAD computing is obtained.After finishing step S111 or S113, meeting of the present invention is as described in the step S115, judge the comparison block whether with search window in all macro zone block all carried out the SAD computing.When also having macro zone block not carry out SAD computing (just step S115 indicated "No"), then repeat steps flow charts such as step S107.
Flow process shown in Figure 1 is in order to seek minimum SAD critical value, so that find out macro zone block corresponding in two pictures, and to calculate the motion vector between corresponding macro zone block, and its step can have detailed explanation in Fig. 4.
In the present invention, when the comparison block with search window in macro zone block when carrying out the SAD computing, the present invention can select in the macro zone block significant image border to carry out the SAD computing earlier.For example, the present invention's pixel that can select to be positioned at macro zone block image border (Edge) is preferentially carried out the SAD computing.
Fig. 3 is the synoptic diagram of a kind of image and search window.Please refer to Fig. 3, suppose in picture, have image 301, search window 303 then is positioned at the lower right corner of image.In the present invention, the position of search window 303 can be by system's picked at random, the present invention and indefinite.
The present invention when carrying out the SAD computing, can select the pixel of image border in the macro zone block to carry out the SAD computing earlier earlier in present picture and previous picture, again remaining macro zone block is carried out the SAD computing.Whereby, the present invention just can find out minimum SAD critical value fast, and accelerates the speed that the present invention handles.
Fig. 4 is the flow chart of steps of the method for the movement locus that calculates picture according to one of the present invention preferred embodiment a kind of and track filtering.Please refer to Fig. 4, after the present invention obtains minimum SAD critical value between two pictures, can obtain the motion vector of comparison block in each picture according to the SAD critical value as described in the step S401.
Fig. 5 is the synoptic diagram of the motion vector of comparison block.Please refer to Fig. 5,501 is present picture, and 503 is previous picture.After the present invention finds minimum SAD critical value, just can find out the position of same macro zone block in two pictures, and can calculate its motion vector, for example arrow 505 is represented.
Referring again to Fig. 4, when the present invention as described in the step S401, obtained the motion vector of comparison block after, just can carry out step S403, be exactly according to comparing the motion vector of block in each picture, and obtain the movement locus of all pictures.In the present invention, obtaining the purpose of movement locus, is that detailed explanation can be with reference to following narration for the foundation as the correction dynamic image.
After obtaining movement locus, the present invention can be as in each picture of judgement as described in the step S405, and whether the motion vector of comparison block and the angle of movement locus be greater than a predetermined angle.Suppose that in k picture the motion vector of its comparison block and the angle of movement locus are greater than predetermined angle.Therefore, the present invention just judges the user when taking k picture, because the factor of hand shake etc. for example, and has influenced the picture of shooting.Therefore, meeting of the present invention is just revised the parameter of k picture, so that will compare the motion vector convergence movement locus of block as the parameter of the picture of correction generation problem as described in the step S407.Whereby, just can be so that movement locus can be level and smooth, more natural with what when the user watches, can show.
Yet in some cases, for example need special effect, and when taking dynamic image, the situation of video camera may take place to shake in a large number.Therefore, the present invention allows the user to set up above-mentioned predetermined angle on their own.Therefore, just can not filter out the motion vector that originally produces because of special-effect.
In sum, the present invention has following advantage at least:
1. because the present invention adopts the mode of digital processing to handle dynamic image, so the present invention need not expend a large amount of costs.
Since the present invention when carrying out the SAD computing, the comparison block can carry out the SAD computing earlier to image edge pixels in the macro zone block in the search window earlier, and when the comparison block with search window in macro zone block when carrying out the SAD computing as long as just stop the SAD computing at once greater than the SAD critical value, so the present invention can reduce operation time.
3. because meeting of the present invention is carried out the correction of dynamic image according to the motion vector of comparison block and the angle between the movement locus, therefore can make picture seem more natural.
4. the present invention also allows the user to set up above-mentioned predetermined angle on their own, so the present invention has more elasticity.
Though the present invention discloses as above with preferred embodiment; right its is not in order to limiting the present invention, anyly has the knack of this skill person, without departing from the spirit and scope of the present invention; when can doing a little change and retouching, so protection scope of the present invention is when looking being as the criterion that claim defines.
Claims (4)
1. the disposal route of a dynamic image, and this dynamic image is made up of several continuous pictures, and each those picture comprises and it is characterized in that several pixels that this disposal route comprises:
Receive those pictures in regular turn;
Each those picture is divided into most macro zone blocks, and each macro zone block has the pixel of one first preset number;
One of them is defined as a comparison block with those macro zone blocks in the present picture, and will compare the interior pixel of block, carry out the computing of absolute difference sum with the pixel in the macro zone block in the search window in the previous picture, and obtain a SAD critical value, the macro zone block that wherein comprises one second preset number in this search window, and this comparison block meeting is carried out the computing of absolute difference sum to the macro zone block of the image border in this search window earlier;
With the pixel in this comparison block, carry out the computing of absolute difference sum with the pixel in the macro zone block that did not also carry out the computing of absolute difference sum in this search window, and obtain an operation values, the gained that adds up of a sad value that is produced when wherein this operation values is carried out the computing of absolute difference sum for each those pixel in the macro zone block that did not also carry out the computing of absolute difference sum in this search window;
When the sad value that adds up during, then interrupt the computing of absolute difference sum immediately more than or equal to this SAD critical value; And
When the accumulated value of the sad value of all pixels in the macro zone block that carries out the computing of absolute difference sum during, then this SAD critical value is updated to and carries out the operation values that the computing of absolute difference sum is obtained at present less than this SAD critical value.
2. the disposal route of dynamic image as claimed in claim 1, it is characterized in that, after wherein this comparison block carries out the computing of absolute difference sum to the macro zone block of the image border in this search window, again remaining macro zone block in this search window is carried out the computing of absolute difference sum.
3. the disposal route of dynamic image as claimed in claim 1 is characterized in that, more comprises:
Obtain the motion vector of this comparison block in each those picture according to this SAD critical value;
Obtain a movement locus of those pictures according to the motion vector of this comparison block in each those picture;
Judge in each those picture that whether the motion vector of this comparison block and the angle of this movement locus are greater than a predetermined angle; And
In those pictures picture wherein, the motion vector of this comparison block and the angle of this movement locus are greater than this predetermined angle, then adjust the parameter of the angle of the motion vector of this comparison block and this movement locus greater than the picture of this predetermined angle, make the motion vector of this comparison block level off to this movement locus, and make that movement locus can be level and smooth.
4. the disposal route of dynamic image as claimed in claim 3 is characterized in that, wherein this predetermined angle is to be set up on their own by a user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB200610007850XA CN100454338C (en) | 2006-02-21 | 2006-02-21 | Dynamic image processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB200610007850XA CN100454338C (en) | 2006-02-21 | 2006-02-21 | Dynamic image processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101025827A CN101025827A (en) | 2007-08-29 |
| CN100454338C true CN100454338C (en) | 2009-01-21 |
Family
ID=38744096
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB200610007850XA Expired - Fee Related CN100454338C (en) | 2006-02-21 | 2006-02-21 | Dynamic image processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100454338C (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102932581B (en) * | 2011-08-08 | 2016-03-30 | 联咏科技股份有限公司 | Image processing circuit and image processing method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1436424A (en) * | 2000-06-07 | 2003-08-13 | 英特尔公司 | Adaptive early exit techniques for image correlation minimum distortion calculation |
| EP1383337A2 (en) * | 2002-07-18 | 2004-01-21 | Samsung Electronics Co., Ltd. | Hierarchical motion vector estimation |
| CN1556652A (en) * | 2003-12-30 | 2004-12-22 | 中国科学院计算技术研究所 | Video freguencying frame image fast coding method based on optimal prediction mode probability |
-
2006
- 2006-02-21 CN CNB200610007850XA patent/CN100454338C/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1436424A (en) * | 2000-06-07 | 2003-08-13 | 英特尔公司 | Adaptive early exit techniques for image correlation minimum distortion calculation |
| EP1383337A2 (en) * | 2002-07-18 | 2004-01-21 | Samsung Electronics Co., Ltd. | Hierarchical motion vector estimation |
| CN1556652A (en) * | 2003-12-30 | 2004-12-22 | 中国科学院计算技术研究所 | Video freguencying frame image fast coding method based on optimal prediction mode probability |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101025827A (en) | 2007-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100381348B1 (en) | How to convert 2D image to 3D image | |
| US8131014B2 (en) | Object-tracking computer program product, object-tracking device, and camera | |
| US8983175B2 (en) | Video processing method and device for depth extraction | |
| CN110322542B (en) | Reconstructing views of a real world 3D scene | |
| CN101366271A (en) | Automatic flicker correction in an image capture device | |
| KR0182058B1 (en) | Apparatus and method of multi-resolution circulating search for motion estimation | |
| US20100302438A1 (en) | Image processing apparatus and image processing method | |
| CN102883099A (en) | Anti-shake method and device for shooting | |
| CN101102504B (en) | A mixing motion detection method combining with video encoder | |
| US20030227972A1 (en) | Apparatus and method for coding moving image | |
| WO2023006009A1 (en) | Photographing parameter determination method and apparatus, and electronic device | |
| CN106713762B (en) | Image processing method and device | |
| US7957468B2 (en) | Method for processing motion image | |
| CN100454338C (en) | Dynamic image processing method | |
| CN112508820A (en) | Image processing method and device and electronic equipment | |
| CN104182940A (en) | Blurred image restoration method and system | |
| JP5451364B2 (en) | Subject tracking device and control method thereof | |
| JP2020046960A (en) | Image processing apparatus, control method for image processing apparatus, and program | |
| CN112261262B (en) | Image calibration method and device, electronic equipment and readable storage medium | |
| CN114286011A (en) | Focusing method and device | |
| CN112887629B (en) | Frequency detection method, frequency detection device, electronic equipment and storage medium | |
| JP2000350188A (en) | Background image update device and its method | |
| Xue et al. | Video retargeting using optimized crop-and-scale | |
| JP2022138536A (en) | Image processing device, image processing method, imaging apparatus, program, and recording medium | |
| JP2016158052A (en) | Image processing apparatus and image processing 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 | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090121 Termination date: 20190221 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |