CN101483747A - Movement detection method suitable for deinterlacing technique - Google Patents
Movement detection method suitable for deinterlacing technique Download PDFInfo
- Publication number
- CN101483747A CN101483747A CNA2008103064508A CN200810306450A CN101483747A CN 101483747 A CN101483747 A CN 101483747A CN A2008103064508 A CNA2008103064508 A CN A2008103064508A CN 200810306450 A CN200810306450 A CN 200810306450A CN 101483747 A CN101483747 A CN 101483747A
- Authority
- CN
- China
- Prior art keywords
- motion
- interpolation point
- motion detection
- detection result
- absolute difference
- 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.)
- Granted
Links
Landscapes
- Television Systems (AREA)
Abstract
The invention relates to a method of motion detection section of de-interlacing technology. Technical problem to be solved of the invention is to provide a motion detection method with more accurate adaptive algorithm. A method applied for motion detection section of de-interlacing technology comprises steps of: a, setting a motion threshold T1 and a stillness threshold T2; b, calculating a first field difference, a second field difference and a frame difference; c, using a maximal value of the first field difference, second field difference and frame difference as a motion parameter; d, judging whether the motion parameter is more than the T1,if so, motion detection result of interpolation point to be detected is moving state; if not, judging whether the motion parameter is less than the T2, if so, motion detection result of interpolation point to be detected is still state, if not, motion detection result of interpolation point to be detected is between motion and still states. The invention takes adequate consideration of motion detection condition, which avoids misjudge and emissive judgement in the motion detection process to improve definition of picture.
Description
Technical field
The invention belongs to Digital Image Processing and video display technology field, be specifically related to a kind of method of motion detection section of de-interlacing technology.
Background technology
The most frequently used radio and television standard comprises that the pal mode (being mainly used in Europe and China) of NTSC 525 row 60Hz field frequency standards (being mainly used in North America and Japan) and 625 row 50Hz field frequencies is because interlaced picture is lost vertical resolution, and there are shortcomings such as line flicker and line crawl, so will be interlaced to conversion line by line inevitably, i.e. deinterlacing technique.The video image deinterlacing technique is an image processing at field of television application present stage one of indispensable a kind of important technology.
Go the interlacing algorithm to be divided into linear algorithm, nonlinear algorithm, Motion Adaptive algorithm and movement compensating algorithm according to the kind difference of its used filter.Wherein, the Motion Adaptive algorithm is selected interpolation method according to the kinematic parameter that motion detection obtains, at present most common way is for judging that parameter is complete when static, adopt between the field and duplicate or average interpolation, pure field interpolation is then adopted in motion fully, then above-mentioned two kinds of interpolation results is done ranking operation between motion and standstill.Duplicate wherein, promptly adjacent last one or next respective coordinates location point pixel number are composed to current point to be inserted.Average interpolation is promptly asked average value as current point to be inserted to adjacent two the corresponding same coordinate position pixel numbers in front and back.Pure field interpolation is and only utilizes current to be inserted some institute's presence data to calculate the value of trying to achieve point to be inserted.Employing can be eliminated the motion sawtooth based on the interlacing algorithm that goes of motion detection, improves the vertical definition of image.
About the general detection method that adopts based on the field difference of motion detection, promptly utilize adjacent two pixel value and threshold ratio, then be judged as motion greater than this value, otherwise just think rest point.Perhaps being set at two threshold values, that motion conditions is divided into motion is static and between three kinds of situations between the two.Because the information field and the pixel that utilize are less, the motion erroneous judgement that may cause multiple situation or fail to judge, cause carrying out wrong interpolation, have influence on the interlacing effect at last, quality problems such as ghost image or details distortion occur based on the de-interlacing interpolation of motion detection.
Summary of the invention
Technical problem to be solved by this invention is, provides a kind of by many Pixel Information, makes Motion Adaptive algorithm method for testing motion more accurately.
The present invention solves the problems of the technologies described above the technical scheme that is adopted to be, a kind of method for testing motion that is applicable to deinterlacing technique may further comprise the steps:
A, the motion thresholding T1 that motion determination is set and static thresholding T2;
B, vertical 2 the adjacent pixel value average X of calculating, i.e. X=(f with interpolation point
n(i-1, j)+f
n(i+1, j))/2, with the absolute difference of interpolation point pixel value of the point of correspondence position in previous field, its absolute difference result is first poor f
L, i.e. f
L=| f
N-1(i, j)-X|; Calculate vertical 2 adjacent pixel value average X with interpolation point, with the absolute difference of interpolation point pixel value of the point of correspondence position in Hou Yichang, its absolute difference result is second poor f
R, f
R=| f
N+1(i, j)-X|; Calculate the absolute difference with the pixel value of adjacent two corresponding interpolation point positions, its absolute difference result is frame difference f
LR, f
LR=| f
N-1(i, j)-f
N+1(i, j) |;
C, get maximum between first difference, second difference, the frame difference as kinematic parameter;
D, whether judge kinematic parameter greater than motion thresholding T1, in this way, the motion detection result of interpolation point is a motion state; As not entering step e;
E, whether judge kinematic parameter less than static thresholding T2, in this way, the motion detection result of interpolation point is an inactive state; As not, the motion detection result of interpolation point is the state between motion and standstill.
The present invention by many Pixel Information try to achieve poor, frame is poor, makes that the result of motion detection is accurate.
Further, for the result who makes motion detection meets the custom that human eye is watched more, selected human eye sensitivity threshold TM (the TM value is between 12-24) as the foundation of determining motion thresholding T1 and static thresholding T2 to the sensitivity of brightness according to human eye.Concrete, motion thresholding T1=TM+TM/8, described static thresholding T1=TM-TM/8.
For the influence of determining that vertical orientation data causes field difference calculating, the present invention before the step b, has introduced the step x that calculates vertical direction weight coefficient б further after step a:
X-1, the 3rd threshold value T3, the 4th threshold value T4 are set;
X-2, calculating interpolation point be the absolute difference V of 2 adjacent pixel value, i.e. V=|f vertically
n(i-1, j)-f
n(i+1, j) |;
X-3, judge vertical adjacent 2 the pixel value of interpolation point absolute difference V whether greater than the 4th threshold value T4, in this way, vertical direction weight coefficient б is 0; As not, enter step x-4;
X-4, judge vertical adjacent 2 the pixel value of interpolation point absolute difference V whether less than the 3rd threshold value T3, in this way, vertical direction weight coefficient б is 1; As not, vertical direction weight coefficient б is (T4-V)/(T4-T3).
Among the described step b, calculate vertical 2 adjacent pixel value average with interpolation point, with the absolute difference of interpolation point pixel value of the point of correspondence position in previous field, its absolute difference result and vertical direction weight coefficient are multiplied each other, with multiplied result as first poor f
L, i.e. f
L=| f
N-1(i, j)-X| * б; Calculate vertical 2 adjacent pixel value average,, its absolute difference result and vertical direction weight coefficient multiplied each other with the absolute difference of interpolation point pixel value of the point of correspondence position in Hou Yichang with interpolation point, with multiplied result as second poor f
R, f
R=| f
N+1(i, j)-X| * б.
For avoiding the single-point noise effect, and be convenient to hardware and realize, in the frame difference is calculated, add any absolute difference of pixel value of the adjacent left side, adjacent two corresponding interpolation point positions, any absolute difference of pixel value of adjacent the right, adjacent two corresponding interpolation point positions; Average together with the absolute difference of the pixel value of adjacent two corresponding interpolation point positions, with the mean value result as frame difference f
LR, f
LR=(2 * | f
N-1(i, j)-f
N+1(i, j) |+| f
N-1(i, j-1)-f
N+1(i, j-1) |+| f
N-1(i, j+1)-f
N+1(i, j+1) |)/4.
Further, consider that there are inaccurate factor in adjacent two Pixel Information in front and back to be inserted and the field difference calculating of working as the front court, and can not accurately judge all motion conditions with two Pixel Information in front and back separately, so introduce the judgement that auxiliary parameter participates in motion detection result, through test, can accurately judge multiple motion conditions: i.e. the present invention is in step b, also need calculate when interpolation between front court and one the field of being separated by and put the absolute difference of the pixel value of vertical consecutive points, with this absolute difference result as auxiliary parameter f
L2U, f
L2Df
L2U=| f
n(i-1, j)-f
N-2(i-1, j) |; f
L2D=| f
n(i+1, j)-f
N-2(i+1, j) |;
Steps d, whether judge kinematic parameter greater than motion thresholding T1, and whether auxiliary parameter is greater than the 5th threshold value T5, in this way, the motion detection result of interpolation point is a motion state; As not entering step e;
Step e, whether judge kinematic parameter less than static thresholding T2, and whether auxiliary parameter is greater than the 5th threshold value T5, in this way, the motion detection result of interpolation point is an inactive state; As not, the motion detection result of interpolation point is the state between motion and standstill.
The effect of the 5th threshold value T5 is further to judge whether interpolation point remains static, and therefore the value of the 5th threshold value T5 is relevant with static thresholding T2, but the 5th threshold value T5 value is T2/2.
Further, last in motion detection, the motion conditions of point is judged the motion conditions of point to be inserted again around utilizing, if point to be inserted is judged as rest point, and most of on every side point is judged as the motor point, a motion conditions then to be inserted is modified to motion, does interlacing according to the method for frame interpolate value and handles.When point to be inserted is judged as motion, and when putting major part on every side and being judged as rest point, it is static then revising its testing result, according to duplicating between the field or the method for average interpolation is gone interlacing, be in the steps d, after the motion detection result of interpolation point is judged as motion state, judge again whether the consecutive points of this interpolation point are judged as inactive state, in this way, the motion detection result of then revising this interpolation point is an inactive state; As not, then keep motion detection result; Among the step e, after the motion detection result of interpolation point is judged as inactive state, judge whether the consecutive points of this interpolation point are judged as motion state, in this way, the motion detection result of then revising this interpolation point is a motion state again; As not, then keep motion detection result.
The invention has the beneficial effects as follows, the motion detection condition has been done sufficient consideration, avoid the erroneous judgement in the motion detection process or fail to judge the raising image sharpness.
Embodiment
1, determines motion thresholding T1, static thresholding T2: establish according to the corresponding relation curve of vertical adjacent 2 average X=250 and human eye sensitivity and choose human eye sensitivity threshold TM, TM value 24; T1=TM+TM/8=27 then; T2=TM-TM/8=21;
2, ask vertical direction weight coefficient б: determine that vertical orientation data to the influence that field difference calculating causes, influences the calculating of field difference in the mode of weight.If the 3rd threshold value T3=38; The 4th threshold value T4=130;
V is the absolute difference of vertical adjacent 2 pixel values of point to be inserted, that is:
V＝|f
n(i-1，j)-f
n(i+1，j)|；
Obtain the vertical direction weight coefficient according to Rule of judgment:
if(V<38)
б＝1.0；
else?if(V>130)
б＝0.0；
else
б＝(130-V)/(130-38)
3, ask frame difference f
LRField difference f
L, f
RAnd ask three's maximum, wherein f
L, f
RBe multiplied by vertical weight б.X is point to be inserted 2 pixel averages up and down;
X＝(f
n(i-1，j)+f
n(i+1，j))/2
f
L＝|f
n-1(i，j)-X|×б；
f
R＝|f
n+1(i，j)-X|×б；
f
LR＝(2×|f
n-1(i，j)-f
n+1(i，j)|+|f
n-1(i，j-1)-f
n+1(i，j-1)|+|f
n-1(i，j+1)-f
n+1(i，j+1)|)/4；
Kinematic parameter M=max (f
L, f
R, f
LR);
And calculating auxiliary parameter f
L2UWith f
L2D:
f
L2U＝|f
n(i-1，j)-f
n-2(i-1，j)|；
f
L2D＝|f
n(i+1，j)-f
n-2(i+1，j)|；
4, motion determination: establish T5=T2/2;
If M〉T1 and (f
L2UT5 or f
L2DT5)
Motion detection result MS=1; The motion detection result of // interpolation point is a motion state;
Else if T1〉M〉T2 and (f
L2UT5 or f
L2DT5)
Motion detection result MS=(M-T2)/T1-T2; The motion detection result of // interpolation point is that the boundary is between motion state and inactive state;
else
Motion detection result MS=0; The motion detection result of // interpolation point is an inactive state;
5, movement protection step: suppose that current some MS (n) is judged as inactive state, but 2 MS (n-1) and MS (n+1) are judged as and have moved before and after it, will be modified to motion state to this motion detection result MS (n) so.All be judged as inactive state otherwise work as 2 of front and back, and current point is judged as motion state, then will be modified to inactive state to motion detection result;
6, the interpolation result y of interpolation point is: y=MS* field interpolation+(1-MS) duplicate between the field.
Claims (10)
- [claim 1] a kind of method for testing motion that is applicable to deinterlacing technique is characterized in that, may further comprise the steps:A, the motion thresholding T1 that motion determination is set and static thresholding T2;B, vertical 2 the adjacent pixel value average of calculating with interpolation point, with the absolute difference of interpolation point pixel value of the point of correspondence position in previous field, its absolute difference result is first difference; Calculate vertical 2 adjacent pixel value average with interpolation point, with the absolute difference of interpolation point pixel value of the point of correspondence position in Hou Yichang, its absolute difference result is second difference; Calculate the absolute difference with the pixel value of adjacent two corresponding interpolation point positions, its absolute difference result is that frame is poor;C, get maximum between first difference, second difference, the frame difference as kinematic parameter;D, whether judge kinematic parameter greater than motion thresholding T1, in this way, the motion detection result of interpolation point is a motion state; As not entering step e;E, whether judge kinematic parameter less than static thresholding T2, in this way, the motion detection result of interpolation point is an inactive state; As not, the motion detection result of interpolation point is the state between motion and standstill.
- [claim 2] a kind of according to claim 1 method for testing motion that is applicable to deinterlacing technique is characterized in that, after the step a, also includes the step x that calculates the vertical direction weight coefficient before the step b;Among the described step b, calculate vertical 2 adjacent pixel value average with interpolation point, with the absolute difference of interpolation point pixel value of the point of correspondence position in previous field, its absolute difference result and vertical direction weight coefficient are multiplied each other, with multiplied result as first difference; Calculate vertical 2 adjacent pixel value average,, its absolute difference result and vertical direction weight coefficient multiplied each other with the absolute difference of interpolation point pixel value of the point of correspondence position in Hou Yichang with interpolation point, with multiplied result as second difference.
- [claim 3] is characterized in that as a kind of method for testing motion that is applicable to deinterlacing technique as described in the claim 2 the step x of described calculating vertical direction weight coefficient specifically may further comprise the steps:X-1, the 3rd threshold value T3, the 4th threshold value T4 are set;X-2, calculating interpolation point be the absolute difference V of 2 adjacent pixel value vertically;X-3, judge vertical adjacent 2 the pixel value of interpolation point absolute difference V whether greater than the 4th threshold value T4, in this way, the vertical direction weight coefficient is 0; As not, enter step x-4;X-4, judge vertical adjacent 2 the pixel value of interpolation point absolute difference V whether less than the 3rd threshold value T3, in this way, the vertical direction weight coefficient is 1; As not, the vertical direction weight coefficient is (T4-V)/(T4-T3).
- [claim 4] a kind of according to claim 1 method for testing motion that is applicable to deinterlacing technique, it is characterized in that, among the step b, also need calculate any absolute difference of pixel value of the adjacent left side, adjacent two corresponding interpolation point positions, any absolute difference of pixel value of adjacent the right, adjacent two corresponding interpolation point positions; Average together with the absolute difference of the pixel value of adjacent two corresponding interpolation point positions, the result is poor as frame with mean value.
- [claim 5] is as a kind of method for testing motion that is applicable to deinterlacing technique as described in the claim 1,2,3 or 4, it is characterized in that, among the step b, also need calculate when interpolation between front court and one the field of being separated by and put the absolute difference of the pixel value of vertical consecutive points, with this absolute difference result as auxiliary parameter;Steps d, whether judge kinematic parameter greater than motion thresholding T1, and whether auxiliary parameter is greater than the 5th threshold value T5, in this way, the motion detection result of interpolation point is a motion state; As not, enter step e;Step e, whether judge kinematic parameter less than static thresholding T2, and whether auxiliary parameter is greater than the 5th threshold value T5, in this way, the motion detection result of interpolation point is an inactive state; As not, the motion detection result of interpolation point is the state between motion and standstill.
- [claim 6] is characterized in that as a kind of method for testing motion that is applicable to deinterlacing technique as described in the claim 5 described the 5th threshold value T5 is 1/2 times of static thresholding T2.
- [claim 7] a kind of according to claim 1 method for testing motion that is applicable to deinterlacing technique is characterized in that, when motion detection result is a motion state, then the Motion Adaptive weight coefficient is 1; When motion detection result is an inactive state, then the Motion Adaptive weight coefficient is 0; When motion detection result is a state between motion and standstill, then the Motion Adaptive weight coefficient is (kinematic parameter T-2)/(T1-T2).
- [claim 8] a kind of according to claim 1 method for testing motion that is applicable to deinterlacing technique, it is characterized in that, in the steps d, after the motion detection result of interpolation point is judged as motion state, whether the consecutive points of judging this interpolation point again are judged as inactive state, in this way, the motion detection result of then revising this interpolation point is an inactive state; As not, then keep motion detection result;Among the step e, after the motion detection result of interpolation point is judged as inactive state, judge whether the consecutive points of this interpolation point are judged as motion state, in this way, the motion detection result of then revising this interpolation point is a motion state again; As not, then keep motion detection result.
- [claim 9] a kind of according to claim 1 method for testing motion that is applicable to deinterlacing technique, it is characterized in that described motion thresholding T1=TM+TM/8, described static thresholding T1=TM-TM/8, wherein TM is the human eye sensitivity threshold, and human eye sensitivity threshold TM value is between 12-24.
- [claim 10] is characterized in that as a kind of method for testing motion that is applicable to deinterlacing technique as described in the claim 3 described the 3rd threshold value T3 value is between 0-100, and the 4th threshold value T4 value is between 100-200.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008103064508A CN101483747B (en) | 2008-12-22 | 2008-12-22 | Movement detection method suitable for deinterlacing technique |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008103064508A CN101483747B (en) | 2008-12-22 | 2008-12-22 | Movement detection method suitable for deinterlacing technique |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101483747A true CN101483747A (en) | 2009-07-15 |
| CN101483747B CN101483747B (en) | 2010-12-22 |
Family
ID=40880657
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008103064508A Expired - Fee Related CN101483747B (en) | 2008-12-22 | 2008-12-22 | Movement detection method suitable for deinterlacing technique |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101483747B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102611829A (en) * | 2012-02-03 | 2012-07-25 | 安徽云森物联网科技有限公司 | Image de-interleaving method based on block partitioning |
| CN105657317A (en) * | 2014-11-14 | 2016-06-08 | 澜起科技（上海）有限公司 | Interlaced video motion detection method and system in video de-interlacing |
-
2008
- 2008-12-22 CN CN2008103064508A patent/CN101483747B/en not_active Expired - Fee Related
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102611829A (en) * | 2012-02-03 | 2012-07-25 | 安徽云森物联网科技有限公司 | Image de-interleaving method based on block partitioning |
| CN102611829B (en) * | 2012-02-03 | 2014-09-24 | 安徽云森物联网科技有限公司 | Image de-interleaving method based on block partitioning |
| CN105657317A (en) * | 2014-11-14 | 2016-06-08 | 澜起科技（上海）有限公司 | Interlaced video motion detection method and system in video de-interlacing |
| CN105657317B (en) * | 2014-11-14 | 2018-10-16 | 澜至电子科技（成都）有限公司 | A kind of interlaced video method for testing motion in video release of an interleave and its system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101483747B (en) | 2010-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6104755A (en) | Motion detection using field-difference measurements | |
| US7423691B2 (en) | Method of low latency interlace to progressive video format conversion | |
| CN101483746B (en) | Deinterlacing method based on movement detection | |
| US20080304568A1 (en) | Method for motion-compensated frame rate up-conversion | |
| JPH1032729A (en) | Interpolation filter for video signal | |
| US8115867B2 (en) | Image processing device | |
| US7405766B1 (en) | Method and apparatus for per-pixel motion adaptive de-interlacing of interlaced video fields | |
| US6452972B1 (en) | Motion detection using field-difference measurements | |
| CN102946505A (en) | Self-adaptive motion detection method based on image block statistics | |
| JP4031389B2 (en) | Image conversion apparatus and image conversion method | |
| US20070053013A1 (en) | Image signal processing apparatus and interlace-to-progressive conversion method | |
| US20080165278A1 (en) | Human visual system based motion detection/estimation for video deinterlacing | |
| CN101483747B (en) | Movement detection method suitable for deinterlacing technique | |
| US7616693B2 (en) | Method and system for detecting motion between video field of same and opposite parity from an interlaced video source | |
| JP4031390B2 (en) | Image conversion apparatus and image conversion method | |
| CN201222771Y (en) | High speed edge self-adapting de-interlaced interpolation device | |
| US8532177B2 (en) | Motion adaptive image processing | |
| CN102186046A (en) | Video image de-interlacing scanning method and device | |
| US20080151107A1 (en) | Progressive scanning conversion apparatus and progressive scanning conversion method | |
| CN100425054C (en) | Film mode extrapolation | |
| JP5178477B2 (en) | Video signal processing apparatus and method, and video display apparatus | |
| US8553145B2 (en) | Method and related apparatus for image de-interlacing | |
| CN102497492B (en) | Detection method for subtitle moving in screen | |
| US8644379B2 (en) | De-interlacing method and method of compensating a de-interlaced pixel | |
| US8922711B2 (en) | Method and apparatus for de-interlacing video data |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20101222 Termination date: 20161222 |