CN102186046A - Video image de-interlacing scanning method and device - Google Patents
Video image de-interlacing scanning method and device Download PDFInfo
- Publication number
- CN102186046A CN102186046A CN 201110148141 CN201110148141A CN102186046A CN 102186046 A CN102186046 A CN 102186046A CN 201110148141 CN201110148141 CN 201110148141 CN 201110148141 A CN201110148141 A CN 201110148141A CN 102186046 A CN102186046 A CN 102186046A
- Authority
- CN
- China
- Prior art keywords
- point
- output
- interpolation point
- front court
- interpolation
- 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 provides a video image de-interlacing scanning method for motion adaptive detection. The method comprises the following steps of: performing PULLDOWN mode detection on a video image; if the same or repeated field is detected and a repeated sequence PATTEN is detected, determining that interlacing scanning field signals are acquired by converting the original progressive scanning frame signals through a PULLDOWN mode, and directly and selectively outputting the brightness value of the current field or the brightness value of the corresponding field of the previous frame, or otherwise, further judging whether the point to be interpolated is a motion point; if the point to be interpolated is the motion point, outputting the brightness estimated value of the current field of the point to be interpolated; and if the point to be interpolated is a static point, outputting the brightness value of the current field. Meanwhile, the invention provides a corresponding device. By using the method and the device, high resolution of each direction is kept, and a saw tooth phenomenon and a virtual concealed phenomenon of a moving object can not be caused.
Description
Technical field
The present invention relates to video image scanning Display Technique field, especially a kind of video image deinterlacing technology.
Background technology
In the traditional analog television technology, for reducing transmission bandwidth and adapting to CRT display screen characteristic, a videoscanning frame is made up of two scanning fields of time sequencing.First field scan shows odd-numbered line 1,3,5...; Second field scan shows even number line 2,4,6....Yet for improving display quality, modern display screen such as LCD (liquid crystal display), DLP (rear-projection demonstration), PLASMA (flat panel display) all adopts progressive scanning technology, and deinterlacing is the interleaving video conversion of signals progressive scanned video picture signal for adapting to this specification requirement exactly.
Be fit to hard-wired deinterlacing technology at present and mainly can reduce following a few class:
1. staggered mixing method (WEAVING): promptly directly make up two scanning fields and become a new scanning frame to go preface.The method simply is easy to realize, but the edge of moving object has tangible crenellated phenomena, and has reduced the time (motion) resolution.
2. average mixing method (BLENDING): two adjacent parity rows in promptly average two scanning fields and generate newline, vertical direction is put to be twice and is formed then.Though the method has solved the crenellated phenomena at moving object edge, produced empty dimly visible elephant of moving object, and reduced vertical resolution and time (motion) resolution.
3. go iterative method (DOUBLING): promptly repeat each row and once make each field scan become each frame scan.Though the method has solved the crenellated phenomena at moving object edge, has reduced vertical resolution.
4. Motion Adaptive method (MOTIONADAPTIVE): promptly the moving object edge based on pixel detects, and adopts different interpolation methods adaptively and take this as a foundation.The method has not only kept the high-resolution of all directions but also can not produce the crenellated phenomena and empty dimly visible elephant of moving object substantially, is the hard-wired class deinterlacing technology preferably that is fit at present.But the spread in performance of existing various Motion Adaptive method and realization is uneven, and great majority are to the requirement height of hardware and bandwidth, and the False Rate height of motion detection or edge detection, and it is big to introduce noise.
In addition, for making the standard per second is that to be suitable for the standard per second be televising of 50 or 59.94 field frequencies for the movie contents of 24 frame frequencies, pull-down pattern such as usually adopted a kind of 2: 2,3: 2,3: 3,24: 1, the movie contents of 24 frame frequencies is become TV field frequency technology and generates the interleaving video signal with frame frequency.For LCD, DLP, the demonstration of lining by line scan of PLASMA, if directly the field frequency conversion of pull-down pattern frame and picture signal go interlacing to handle, effect on the contrary can be poorer.
Summary of the invention
The present invention is intended to solve the deficiencies in the prior art, and a kind of video image deinterlacing method that pull-down pattern (PULLDOWNPATTERN) Motion Adaptive detects that has is provided.
Simultaneously, it is a kind of low to hardware and bandwidth requirement that the present invention also provides, the video image deinterlacing device that False Rate is low.
The video image deinterlacing method that Motion Adaptive detects comprises the steps:
(1) carries out the rim detection of video image, judge whether interpolation point is object edge, calculate interpolation point in the brightness discreet value of working as the front court;
(2) carry out the motion detection of video image, calculate the motion criterion, judge that the interpolation point is motor point or rest point;
(3) carry out pull-down pattern and detect, if detected identical or repeat fields and detect repetitive sequence PATTEN after, illustrate that this interlacing scan field signal is come by script progressive frames conversion of signals by pull-down pattern PULLDOWN;
(4) carrying out interpolation selects: if step (3) is judged as: when the sequence of front court for coming by the progressive frames conversion of signals by pull-down pattern, then directly select output when the brightness value of front court or the brightness value of former frame corresponding fields, otherwise further judge according to the result of step (2):
If judge that the interpolation point is the motor point, the interpolation point of then exporting step (1) output is in the brightness discreet value when the front court; If judge that the interpolation point is a rest point, then be output as brightness value when the front court.
The specific implementation method of above-mentioned steps is:
(1) carry out the rim detection of video image:
In the capable pixel of the I-1 of previous field, to choose from the nearest odd number of pixels point of interpolation point, its brightness value is designated as Y (F-1, I-1, J-M) ..., Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1) ..., Y (F-1, I-1, J+M), and in the capable pixel of the I+1 of previous field, choose pixel with the capable correspondence position of I-1, its brightness value note Y (F-1, I+1, J-M) ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), wherein M is a positive integer greater than 1
If D (I, J)=| Y (F-1, I-1, J)-Y (F-1, I+1, J) | and＜=ED_TH0, interpolation point is non-object edge so, interpolation point in the brightness discreet value when the front court is:
Y ' (F, I, J)=(Y (F-1, I-1, J)+Y (F-1, I+1, J))/2, wherein said ED_TH0 is configurable first threshold value that is obtained by experiment; Otherwise the interpolation point is taken as object edge, treats that following formula is satisfied in the judgement of difference point position:
MINIMUM (| Y (F-1, I-1, J-i)-Y (F-1, I+1, J+i) |), i is 1 to M integer, and treats that the difference point in the brightness discreet value when the front court is:
Y’(F，I，J)＝(Y(F-1，I-1，J-i)+Y(F-1，I+1，J+i))/2。
(2) carry out the motion detection of video image:
The luminance difference of the interpolation point periphery of present frame is:
M(T，I，J)＝∑|Y(T，I+P，J+Q)-Y(T，I，J)|；
The luminance difference of the interpolation point periphery of former frame is:
M(T-1，I，J)＝∑|Y(T-1，I+P，J+Q)-Y(T-1，I，J)|；
P=-1 wherein, 0,1; Q=-1,0,1;
The homologous field of adjacent two frames, the even field of the odd field of promptly adjacent two frames or adjacent two frames in the tested pixel luminance difference of the capable correspondence of I is:
PD(T，I，J)＝|Y(T，I，J)-Y(T-1，I，J)|；
The motion criterion of present frame be MV (T, I, J)=PD (T, I, J)＞=MD_TH0 or
MV (T, I, J)=| M (T, I, J)-M (T-1, I, J) |＞=MD_TH1, MD_TH0, MD_TH1 are come configurable second threshold value and the 3rd threshold value by experiment here.
If (J)=1, then the interpolation point is the motor point to MV for T, I, otherwise the interpolation point is a rest point.
(3) carrying out pull-down pattern detects:
The homologous field of adjacent two frames, the even field of the odd field of promptly adjacent two frames or adjacent two frames in the tested pixel luminance difference of the capable correspondence of I is:
PD (T, I, J)=| Y (T, I, J)-Y (T-1, I, J) |; ((J)＜=PD_TH0)＜PD_TH1, then ruling is identical with the corresponding fields of preceding frame when the front court, otherwise is difference for T, I for PD if when all pixel luminance differences of front court are crossed thresholding accumulative total PD (T)=∑.In case detected identical or repeat fields and detect repetitive sequence PATTEN after, illustrate that this interlacing scan field signal is come by script progressive frames conversion of signals by pull-down pattern PULLDOWN.
Here PD_TH0, PD_TH1 are respectively the 5th threshold values of being come configurable the 4th threshold value and being adjusted according to the image size adaptation by experiment.
(4) carrying out interpolation selects:
If the sequence of working as the front court is then directly selected brightness value Y (F, the I of output when the front court for being come by the progressive frames conversion of signals by pull-down pattern, J) or the brightness value Y of former frame corresponding fields (F-1, I, J), judge that else if the interpolation point is the motor point, then output when the brightness discreet value Y ' of front court (F, I, J), if judge that the interpolation point is a rest point, then be output as brightness value Y when the front court (F, I, J).
The video image deinterlacing method that Motion Adaptive detects also further comprises denoising filtering:
In the capable pixel of the I-1 of previous field, to choose from the odd number of pixels point for the treatment of that difference point is nearest, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel with the capable correspondence position of I-1, its brightness value note Y (F-1, I+1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is the positive integer greater than 1, and choose when the more preceding denoising filtering of front court interpolation point as a result YO (F, I, J-1), get the rightmost initial value YO of every row (F, I ,-1)=Y " (F; I, 0).If
Y”(F，I，J)+DN_TH0＜＝MINIMUM(Y(F-1，I-1，J-M)、…、Y(F-1，I-1，J-1)、Y(F-1，I-1，J)、Y(F-1，I-1，J+1)、…、Y(F-1，I-1，J+M)，(F-1，I+1，J-M)、、…、Y(F-1，I+1，J-1)、Y(F-1，I+1，J)、Y(F-1，I+1，J+1)、…、Y(F-1，I+1，J+M)、YO(F，I，J-1))；
Or Y " (F, I, J)-DN_TH0＞=MAXIMUM (Y (F-1, I-1, J-M) ..., Y (F-1; I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1; J+1) ..., Y (F-1, I-1, J+M), (F-1; I, J-M), ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), YO (F, I, J-1));
DN_TH0 is come configurable the 6th threshold value by experiment.
Then denoising filtering result is: and YO (F, I, J)=Y (F-1, I-1, J)/4+Y (F-1, I+1, J)/4+YO (F, I, J-1)/2; Otherwise be YO (F, I, J)=Y " (F, I, J)
The video image deinterlacing device that Motion Adaptive detects comprises:
DMA (direct memory visit) controller connects the DRAM memory;
Line buffer, described line buffer under the control of dma controller from the DRAM memory read and the pixel brightness value of buffer memory output previous field and former frame corresponding fields pixel brightness value, when the brightness value of front court pixel and the peripheral luminance difference of interpolation point of former frame;
Edge detection module ED, input is from the capable pixel of I-1 of the previous field of line buffer and the capable pixel of I+1 of previous field, in the capable pixel of the I-1 of previous field F-1, choose from the nearest odd number of pixels point of interpolation point, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel, its brightness value note Y (F-1, I+1 with the capable correspondence position of I-1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is a positive integer greater than 1.
If D (I, J)=| Y (F-1, I-1, J)-and Y (F-1, I+1, J) |＜=ED_TH0, ED_TH0 is configurable first threshold value that is obtained by experiment, and interpolation point is non-object edge so, and the interpolation point of described edge detection module output in the brightness discreet value when the front court is:
Y’(F，I，J)＝(Y(F-1，I-1，J)+Y(F-1，I+1，J))/2。Otherwise the interpolation point of described edge detection module output is an object edge, and following formula is satisfied in the judgement for the treatment of difference point position of described edge detection module output:
MINIMUM (| Y (F-1, I-1, J-i)-Y (F-1, I+1, J+i) |), i is 1 to M integer, the difference point for the treatment of of described edge detection module output in the brightness discreet value when the front court is:
Y’(F，I，J)＝(Y(F-1，I-1，J-i)+Y(F-1，I+1，J+i))/2。
Motion detection block MD, input from the luminance difference M of the interpolation point periphery of the former frame of line buffer (J) and the pixel of the interpolation point of present frame T periphery, the peripheral luminance difference of interpolation point of calculating present frame T is for T-1, I:
M (T, I, J)=∑ | Y (T, I+P, J+Q)-Y (T, I, J) |, P=-1 wherein, 0,1; Q=-1,0,1;
The luminance difference of the interpolation point periphery of described motion detection block output present frame T is saved in the DRAM memory through dma controller;
Described motion detection block is exported the homologous field of adjacent two frames, the even field of the odd field of promptly adjacent two frames or adjacent two frames, and in the tested pixel luminance difference of the capable correspondence of I, its value is:
PD(T，I，J)＝|Y(T，I，J)-Y(T-1，I，J)|；
The motion criterion of present frame be MV (T, I, J), if PD (T, I, J)＞=MD_TH0 or
MV (T, I, J)=| M (T, I, J)-M (T-1, I, J) |＞=MD_TH1, MD_TH0, MD_TH1 are come configurable second threshold value and the 3rd threshold value, then motion detection block MD output movement criterion MV (T by experiment here, I, J)=1, the interpolation point is the motor point, otherwise motion detection block MD output movement criterion MV (T, I, J)=0, the interpolation point is a rest point.
Pull-down pattern detection module PD, input is from the homologous field of adjacent two frames of line buffer, the even field of the odd field of promptly adjacent two frames or adjacent two frames, in the tested pixel luminance difference of the capable correspondence of I is PD (T, I J), adds up if cross thresholding when all pixel luminance differences of front court:
PD (T)=∑ (PD (T, I, J)＜=PD_TH0)＜and PD_TH1, it is identical with the corresponding fields of preceding frame that then the front court is worked as in ruling, otherwise be different.In case detected identical or repeat fields and detect repetitive sequence PATTEN after, illustrate that this interlacing scan field signal is come by script progressive frames conversion of signals by pull-down pattern PULLDOWN.
Here PD_TH0, PD_TH1 are respectively the 5th threshold values of being come configurable the 4th threshold value and being adjusted according to the image size adaptation by experiment.
Interpolation is selected module WS, the signal of input is: the interpolation point of edge detection module output is at brightness discreet value the Y ' (F when the front court, I, J), motion detection block MD output movement criterion MV (T, I, J), the brightness value Y (F when the front court of the output of pull-down pattern detection module PD and line buffer output, I, J) or brightness value Y (F-1, the I of former frame corresponding fields, J), if pull-down pattern detection module PD is output as very, promptly when the sequence of front court for coming by the progressive frames conversion of signals by pull-down pattern, then described interpolation selects module WS directly to select to export (the F as the brightness value Y of front court according to the PATTERN of detection, I, J) or the brightness value Y of former frame corresponding fields (F-1, I, J); If pull-down pattern detection module PD be output as vacation and motion detection block MD output movement criterion MV (J)=0, promptly the interpolation point is static for T, I, then described interpolation select module WS be output as line buffer output the brightness value Y when the front court (F, I, J); If pull-down pattern detection module PD be output as vacation and motion detection block MD output movement criterion MV (T, I, J)=1, be that the interpolation point is the motor point, then described interpolation select interpolation point that module WS is output as edge detection module output when the brightness discreet value Y ' of front court (F, I, J).
The video image deinterlacing device that Motion Adaptive detects, also comprise the denoising filtration module, brightness value Y (F, the I when the front court of described denoising filtration module line of input buffer output, J) and interpolation select the output of module, post-processing unit is given in the filtering of output denoising.
In the capable pixel of the I-1 of previous field, to choose from the odd number of pixels point for the treatment of that difference point is nearest, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel with the capable correspondence position of I-1, its brightness value note Y (F-1, I+1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is the positive integer greater than 1, and choose preceding 1 YO when front court F interpolation point (F, I, J-1), get the rightmost initial value YO of every row (F, I ,-1)=Y " (F; I, 0).If
Y”(F，I，J)+DN_TH0＜＝MINIMUM(Y(F-1，I-1，J-M)、…、Y(F-1，I-1，J-1)、Y(F-1，I-1，J)、Y(F-1，I-1，J+1)、…、Y(F-1，I-1，J+M)，(F-1，I+1，J-M)、、…、Y(F-1，I+1，J-1)、Y(F-1，I+1，J)、Y(F-1，I+1，J+1)、…、Y(F-1，I+1，J+M)、YO(F，I，J-1))；
Or Y " (F, I, J)-DN_TH0＞=MAXIMUM (Y (F-1, I-1, J-M) ..., Y (F-1; I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1; J+1) ..., Y (F-1, I-1, J+M), (F-1; I+1, J-M) ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), YO (F, I, J-1));
DN_TH0 is come configurable the 6th threshold value by experiment.
Then denoising filtering result is: and YO (F, I, J)=Y (F-1, I-1, J)/4+Y (F-1, I+1, J)/4+YO (F, I, J-1)/2; Otherwise be YO (F, I, J)=Y " (F, I, J).
The scope of first threshold value of mentioning in the present invention is: 15～25, the scope of second threshold value is that the scope of (5～10), the 3rd threshold value is: 15～25, the scope of the 4th threshold value is: 10～20, the 5th thresholding according to the scope of the adjustment of image size adaptation, the 6th threshold value is: 15～25.
Video image deinterlacing method beneficial effect with the detection of pull-down pattern (PULLDOWN PATTERN) Motion Adaptive provided by the invention is that it had both solved the motion detection of art methods existence and the high big problem of noise of introducing of False Rate of edge detection, keep the high-resolution of all directions again and can not produce the crenellated phenomena and empty dimly visible elephant of moving object, simultaneously the various interlacing scan field signals that come by script progressive frames conversion of signals by pull-down pattern have been had the undistorted frame signal restore funcitons of lining by line scan of self adaptation.It is a kind of low to hardware and bandwidth requirement that the present invention also provides, the video image deinterlacing device that False Rate is low.
Description of drawings
Fig. 1 is an edge detection algorithm schematic diagram
Fig. 2 is the motion detection algorithm schematic diagram
Fig. 3 is a denoising filtering algorithm schematic diagram
The video image deinterlacing device schematic diagram that Fig. 4 detects for Motion Adaptive
Embodiment
Below in conjunction with accompanying drawing content of the present invention is further specified.
The video image deinterlacing device that Motion Adaptive detects as shown in Figure 4, comprising:
DMA (direct memory visit) controller connects the DRAM memory;
Line buffer, described line buffer under the control of dma controller from the DRAM memory access and the pixel brightness value of buffer memory output previous field and former frame corresponding fields pixel brightness value, when the brightness value of front court pixel and the peripheral luminance difference of interpolation point of former frame;
Edge detection module ED, input is from the capable pixel of I-1 of the previous field of line buffer and the capable pixel of I+1 of previous field, in the capable pixel of the I-1 of previous field F-1, choose from the nearest odd number of pixels point of interpolation point, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel, its brightness value note Y (F-1, I+1 with the capable correspondence position of I-1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is a positive integer greater than 1.
In the present embodiment, as shown in Figure 1, choose the detection judgement of carrying out 5 directions from nearest 5 pixels of 5 pixels of interpolation point, obtain interpolation point and direction when the front court:
If D (I, J)=| Y (F-1, I-1, J)-Y (F-1, I+1, J) | and＜=ED_TH0, interpolation point is non-object edge so, the interpolation point of described edge detection module output in the brightness discreet value when the front court is:
Y ' (F, I, J)=(Y (F-1, I-1, J)+Y (F-1, I+1, J))/2, wherein said ED_TH0 is configurable first threshold value that is obtained by experiment, gets ED_TH0=20 in the present embodiment.
Otherwise the interpolation point of described edge detection module output is an object edge, and following formula is satisfied in the judgement for the treatment of difference point position of described edge detection module output:
MINIMUM (| Y (F-1, I-1, J-i)-Y (F-1, I, J+i) |), i is 1 to M integer, the difference point for the treatment of of described edge detection module output in the brightness discreet value when the front court is:
Y’(F，I，J)＝(Y(F-1，I-1，J-i)+Y(F-1，I+1，J+i))/2。
As shown in Figure 1, establish: A=|Y (F-1, I-1, J-2)-Y (F-1, I+1, J+2) |;
B＝|Y(F-1，I-1，J-1)-Y(F-1，I+1，J+1)|；
C＝|Y(F-1，I-1，J)-Y(F-1，I+1，J)|；
D＝|Y(F-1，I-1，J+1)-Y(F-1，I+1，J-1)|；
E＝|Y(F-1，I-1，J+2)-Y(F-1，I+1，J-2)|；
Because the object edge direction has maximum correlation and promptly has minimum absolute difference, get discreet value Y ' (F, I, J) be 2 of the edge direction on average promptly:
If MINIMUM (A, B, C, D, E)=A, then
Y’(F，I，J)＝(Y(F-1，I-1，J-2)+Y(F-1，I+1，J+2))/2；
If MINIMUM (A, B, C, D, E)=B, then
Y’(F，I，J)＝(Y(F-1，I-1，J-1)+Y(F-1，I+1，J+1))/2；
If MINIMUM (A, B, C, D, E)=C, then
Y’(F，I，J)＝(Y(F-1，I-1，J)+Y(F-1，I+1，J))/2；
If MINIMUM (A, B, C, D, E)=D, then
Y’(F，I，J)＝(Y(F-1，I-1，J+1)+Y(F-1，I+1，J-1))/2；
If MINIMUM (A, B, C, D, E)=E, then
Y’(F，I，J)＝(Y(F-1，I-1，J+2)+Y(F-1，I+1，J-2))/2；
Motion detection block MD, as shown in Figure 2, because when the moving object edge enters interpolation point, the absolute difference of the luminance difference of the former frame of interpolation point periphery and the luminance difference of present frame can obviously strengthen, and perhaps the brightness value of interpolation point has been compared obvious variation at present frame with former frame.Based on this, motion detection block MD utilizes in two frames near 8 pixels interpolation point to carry out following detection to judge.
Input from the luminance difference M of the interpolation point of the former frame of line buffer periphery (J) and the pixel of the interpolation point of present frame periphery, the peripheral luminance difference of interpolation point of calculating present frame is for T-1, I:
M (T, I, J)=∑ | Y (T, I+P, J+Q)-Y (T, I, J) |, P=-1 wherein, 0,1; Q=-1,0,1;
The luminance difference of the interpolation point periphery of described motion detection block output present frame is kept at the DRAM memory through dma controller;
Described motion detection block is exported the homologous field of adjacent two frames, the even field of the odd field of promptly adjacent two frames or adjacent two frames, and in the tested pixel luminance difference of the capable correspondence of I, its value is:
PD(T，I，J)＝|Y(T，I，J)-Y(T-1，I，J)|；
The motion criterion of present frame be MV (T, I, J), if PD (T, I, J)＞=MD_TH0 or MV (T, I, J)=| M (T, I, J)-M (T-1, I, J) |＞=MD_TH1, here MD_TH0, MD_TH1 are come configurable second threshold value and the 3rd threshold value, then motion detection block MD output movement criterion MV (T, I by experiment, J)=1, the interpolation point is the motor point, otherwise motion detection block MD output movement criterion MV (T, I, J)=0, the interpolation point is a rest point.
Get MD_TH0=8 in the present embodiment, MD_TH1=20.
Pull-down pattern detection module PD, equally as shown in Figure 3, input is from the homologous field of adjacent two frames of line buffer, and the even field of the odd field of promptly adjacent two frames or adjacent two frames is PD (T in the tested pixel luminance difference of the capable correspondence of I, I, J), add up PD (T)=∑ (PD (T, I if cross thresholding when all pixel luminance differences of front court, J)＜=PD_TH0)＜PD_TH1, then when the front court identical with the corresponding fields of former frame.Here PD_TH0, PD_TH1 are respectively the 5th threshold values of being come configurable the 4th threshold value and being adjusted according to the image size adaptation by experiment.Get PD_TH0=15 in the present embodiment, PD_TH1=300.In case detected identical or repeat fields and detect repetitive sequence PATTEN after, illustrate that this interlacing scan field signal is come by script progressive frames conversion of signals by pull-down pattern PULLDOWN.Otherwise the sequence of deserving the front court is not by the progressive frames conversion of signals.
Interpolation is selected module WS, the signal of input is: the interpolation point of edge detection module output is at brightness discreet value the Y ' (F when the front court, I, J), motion detection block MD output movement criterion MV (T, I, J), the brightness value Y (F when the front court of the output of pull-down pattern detection module PD and line buffer output, I, J) or brightness value Y (F-1, the I of former frame corresponding fields, J), if pull-down pattern detection module PD is output as very, promptly when the sequence of front court for coming by the progressive frames conversion of signals by pull-down pattern, then described interpolation selection module WS directly selects to export (the F as the brightness value Y of front court, I, J) or the brightness value Y of former frame corresponding fields (F-1, I, J); If pull-down pattern detection module PD be output as vacation and motion detection block MD output movement criterion MV (J)=0, promptly the interpolation point is static for T, I, then described interpolation select module WS be output as line buffer output the brightness value Y when the front court (F, I, J); If pull-down pattern detection module PD is output as vacation and motion detection block MD output movement criterion MV (T, I, J)=1, be that the interpolation point is the motor point, then described interpolation selects module WS to be output as the interpolation point of edge detection module output at brightness discreet value the Y ' (F when the front court, I ', J).
The algorithm of any deinterlacing and realization are understood some unavoidably and are introduced noise. for the performance that further improves deinterlacing is taked the denoising filtering technique, the video image deinterlacing device that Motion Adaptive detects, also comprise the denoising filtration module, the brightness value Y (F when the front court of described denoising filtration module line of input buffer output, I, J) and interpolation select the output of module, post-processing unit is given in the filtering of output denoising.
In the capable pixel of the I-1 of previous field, to choose from the odd number of pixels point for the treatment of that difference point is nearest, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel with the capable correspondence position of I-1, its brightness value note Y (F-1, I+1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is the positive integer greater than 1, and choose preceding 1 YO when front court F interpolation point (F, I, J-1), get the rightmost initial value YO of every row (F, I ,-1)=Y " (F; I, 0).If
Y”(F，I，J)+DN_TH0＜＝MINIMUM(Y(F-1，I-1，J-M)、…、Y(F-1，I-1，J-1)、Y(F-1，I-1，J)、Y(F-1，I-1，J+1)、…、Y(F-1，I-1，J+M)，(F-1，I+1，J-M)、、…、Y(F-1，I+1，J-1)、Y(F-1，I+1，J)、Y(F-1，I+1，J+1)、…、Y(F-1，I+1，J+M)、YO(F，I，J-1))；
Or Y " (F, I, J)-DN_TH0＞=MAXIMUM (Y (F-1, I-1, J-M) ..., Y (F-1; I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1; J+1) ..., Y (F-1, I-1, J+M), (F-1; I+1, J-M) ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), YO (F, I, J-1));
DN_TH0 is come configurable the 6th threshold value by experiment.
Then denoising filtering result is: and YO (F, I, J)=Y (F-1, I-1, J)/4+Y (F-1, I+1, J)/4+YO (F, I, J-1)/2; Otherwise be YO (F, I, J)=Y " (F, I, J).
DN_TH0=10 in this enforcement.YO in this illustration 3 (F, I, J)=Y (F-1, I-1, J)/4+Y (F-1, I+1, J)/4+YO (F, I, J-1)/and 2=62/4+50/4+80/2=68, the result is more level and smooth more than initial value 215.
The invention discloses the video image deinterlacing method and apparatus that Motion Adaptive detects, and describe the specific embodiment of the present invention and effect with reference to the accompanying drawings.What should be understood that is that the foregoing description is just to explanation of the present invention; rather than limitation of the present invention; any innovation and creation that do not exceed in the connotation scope of the present invention, and the replacement of other unsubstantialities or modification all fall within the protection range of the present invention.
Claims (5)
1. the video image deinterlacing method of Motion Adaptive detection comprises the steps:
(1) carries out the rim detection of video image, judge whether interpolation point is object edge, calculate interpolation point in the brightness discreet value of working as the front court;
(2) carry out the motion detection of video image, calculate the motion criterion, judge that the interpolation point is motor point or rest point;
(3) carry out pull-down pattern and detect, if detected identical or repeat fields and detect repetitive sequence PATTEN after, illustrate that this interlacing scan field signal is come by script progressive frames conversion of signals by pull-down pattern PULLDOWN;
(4) carrying out interpolation selects, if step (3) is judged as: the sequence when the front court is to be come by the progressive frames conversion of signals by pull-down pattern, then directly select output as the brightness value of front court or the brightness value Y of former frame corresponding fields, otherwise further the result according to step (2) judges: if judge that the interpolation point is the motor point, the interpolation point of then exporting step (1) output is in the brightness discreet value when the front court; If judge that the interpolation point is a rest point, then be output as brightness value when the front court.
2. the video image deinterlacing method that detects of Motion Adaptive according to claim 1 is characterized in that the specific implementation step is:
(1) carry out the rim detection of video image:
In the capable pixel of the I-1 of previous field, to choose from the nearest odd number of pixels point of interpolation point, its brightness value is designated as Y (F-1, I-1, J-M) ..., Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1) ..., Y (F-1, I-1, J+M), and in the capable pixel of the I+1 of previous field, choose pixel with the capable correspondence position of I-1, its brightness value note Y (F-1, I+1, J-M) ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), wherein M is a positive integer greater than 1
If D (I, J)=| Y (F-1, I-1, J)-Y (F-1, I+1, J) | and＜=ED_TH0, interpolation point is non-object edge so, interpolation point in the brightness discreet value when the front court is:
Y ' (F, I, J)=(Y (F-1, I-1, J)+Y (F-1, I+1, J))/2, wherein, described ED_TH0 is configurable first threshold value that is obtained by experiment; Otherwise the interpolation point is taken as object edge, treats that following formula is satisfied in the judgement of difference point position:
MINIMUM (| Y (F-1, I-1, J-i)-Y (F-1, I+1, J+i) |), i is 1 to M integer, and treats that the difference point in the brightness discreet value when the front court is:
Y’(F，I，J)＝(Y(F-1，I-1，J-i)+Y(F-1，I+1，J+i))/2；
(2) carry out the motion detection of video image:
The luminance difference of the interpolation point periphery of present frame is:
M(T，I，J)＝∑|Y(T，I+P，J+Q)-Y(T，I，J)|；
The luminance difference of the interpolation point periphery of former frame is:
M(T-1，I，J)＝∑|Y(T-1，I+P，J+Q)-Y(T-1，I，J)|；
P=-1 wherein, 0,1; Q=-1,0,1;
The homologous field of adjacent two frames, the even field of the odd field of promptly adjacent two frames or adjacent two frames in the tested pixel luminance difference of the capable correspondence of I is:
PD(T，I，J)＝|Y(T，I，J)-Y(T-1，I，J)|；
The motion criterion of present frame be MV (T, I, J)=PD (T, I, J)＞=MD_TH0 or MV (T, I, J)=| M (T, I, J)-and M (T-1, I, J) | and＞=MD_TH1, MD_TH0, MD_TH1 are come configurable second threshold value and the 3rd threshold value by experiment here;
If (J)=1, then the interpolation point is the motor point to MV for T, I, otherwise the interpolation point is a rest point;
(3) carrying out pull-down pattern detects:
The homologous field of adjacent two frames, the even field of the odd field of promptly adjacent two frames or adjacent two frames in the tested pixel luminance difference of the capable correspondence of I is:
PD (T, I, J)=| Y (T, I, J)-Y (T-1, I, J) |; ((J)＜=PD_TH0)＜PD_TH1, then ruling is identical with the corresponding fields of preceding frame when the front court, otherwise is difference for T, I for PD if when all pixel luminance differences of front court are crossed thresholding accumulative total PD (T)=∑; In case detected identical or repeat fields and detect repetitive sequence PATTEN after, illustrate that this interlacing scan field signal is come by script progressive frames conversion of signals by pull-down pattern PULLDOWN;
Here PD_TH0, PD_TH1 are respectively the 5th threshold values of being come configurable the 4th threshold value and being adjusted according to the image size adaptation by experiment;
(4) carrying out interpolation selects:
If the sequence of working as the front court is then directly selected brightness value Y (F, the I of output when the front court for being come by the progressive frames conversion of signals by pull-down pattern, J) or the brightness value Y of former frame corresponding fields (F-1, I, J), judge that else if the interpolation point is the motor point, then export interpolation point that step (1) obtains when the brightness discreet value Y ' of front court (F, I, J), if judge that the interpolation point is a rest point, then be output as brightness value Y when the front court (F, I, J).
3. the video image deinterlacing method that detects as Motion Adaptive as described in the claim 2, its feature also further comprises denoising filtering:
In the capable pixel of the I-1 of previous field, to choose from the odd number of pixels point for the treatment of that difference point is nearest, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel with the capable correspondence position of I-1, its brightness value note Y (F-1, I+1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is the positive integer greater than 1, and choose when the more preceding denoising filtering of front court interpolation point as a result YO (F, I, J-1), get the rightmost initial value YO of every row (F, I ,-1)=Y " (F; I, 0);
If Y " (F, I, J)+DN_TH0＜=MINIMUM (Y (F-1, I-1, J-M) ..., Y (F-1; I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1; J+1) ..., Y (F-1, I-1, J+M), (F-1; I+1, J-M), ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), YO (F, I, J-1));
Or Y " (F, I, J)-DN_TH0＞=MAXIMUM (Y (F-1, I-1, J-M) ..., Y (F-1; I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1; J+1) ..., Y (F-1, I-1, J+M), (F-1; I, J-M), ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), YO (F, I, J-1));
DN_TH0 is come configurable the 6th threshold value by experiment;
Then denoising filtering result is: and YO (F, I, J)=Y (F-1, I-1, J)/4+Y (F-1, I+1, J)/4+YO (F, I, J-1)/2; Otherwise be YO (F, I, J)=Y " (F, I, J).
4. the video image deinterlacing device that detects of Motion Adaptive comprises:
DMA (direct memory visit) controller connects the DRAM memory;
Line buffer, described line buffer under the control of dma controller from the DRAM memory read and the pixel brightness value of buffer memory output previous field and former frame corresponding fields pixel brightness value, when the brightness value of front court pixel and the peripheral luminance difference of interpolation point of former frame;
Edge detection module ED, input is from the capable pixel of I-1 of the previous field of line buffer and the capable pixel of I+1 of previous field, in the capable pixel of the I-1 of previous field F-1, choose from the nearest odd number of pixels point of interpolation point, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel, its brightness value note Y (F-1, I+1 with the capable correspondence position of I-1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is a positive integer greater than 1;
If D (I, J)=| Y (F-1, I-1, J)-and Y (F-1, I+1, J) |＜=ED_TH0, ED_TH0 is configurable first threshold value that is obtained by experiment, and interpolation point is non-object edge so, and the interpolation point of described edge detection module output in the brightness discreet value when the front court is:
Y ' (F, I, J)=(Y (F-1, I-1, J)+Y (F-1, I+1, J))/2, otherwise the interpolation point of described edge detection module output is an object edge, and following formula is satisfied in the judgement for the treatment of difference point position of described edge detection module output: MINIMUM (| Y (F-1, I-1, J-i)-Y (F-1, I+1, J+i) |), i is 1 to M integer, and the difference point for the treatment of of described edge detection module output in the brightness discreet value when the front court is:
Y’(F，I，J)＝(Y(F-1，I-1，J-i)+Y(F-1，I+1，J+i))/2；
Motion detection block MD, input from the luminance difference M of the interpolation point periphery of the former frame of line buffer (J) and the pixel of the interpolation point of present frame T periphery, the peripheral luminance difference of interpolation point of calculating present frame T is for T-1, I:
M (T, I, J)=∑ | Y (T, I+P, J+Q)-Y (T, I, J) |, P=-1 wherein, 0,1; Q=-1,0,1;
The luminance difference of the interpolation point periphery of described motion detection block output present frame T is saved in the DRAM memory through dma controller;
Described motion detection block is exported the homologous field of adjacent two frames, the even field of the odd field of promptly adjacent two frames or adjacent two frames, and in the tested pixel luminance difference of the capable correspondence of I, its value is:
PD(T，I，J)＝|Y(T，I，J)-Y(T-1，I，J)|；
The motion criterion of present frame be MV (T, I, J), if PD (T, I, J)＞=MD_TH0 or MV (T, I, J)=| M (T, I, J)-M (T-1, I, J) |＞=MD_TH1, here MD_TH0, MD_TH1 are come configurable second threshold value and the 3rd threshold value, then motion detection block MD output movement criterion MV (T, I by experiment, J)=1, the interpolation point is the motor point, otherwise motion detection block MD output movement criterion MV (T, I, J)=0, the interpolation point is a rest point;
Pull-down pattern detection module PD, input is from the homologous field of adjacent two frames of line buffer, the even field of the odd field of promptly adjacent two frames or adjacent two frames, in the tested pixel luminance difference of the capable correspondence of I is PD (T, I J), adds up if cross thresholding when all pixel luminance differences of front court:
PD (T)=∑ (PD (T, I, J)＜=PD_TH0)＜and PD_TH1, it is identical with the corresponding fields of preceding frame that then the front court is worked as in ruling, otherwise be different; In case detected identical or repeat fields and detect repetitive sequence PATTEN after, illustrate that this interlacing scan field signal is come by script progressive frames conversion of signals by pull-down pattern PULLDOWN;
Here PD_TH0, PD_TH1 are respectively the 5th threshold values of being come configurable the 4th threshold value and being adjusted according to the image size adaptation by experiment;
Interpolation is selected module WS, the signal of input is: the interpolation point of edge detection module output is at brightness discreet value the Y ' (F when the front court, I, J), motion detection block MD output movement criterion MV (T, I, J), the brightness value Y (F when the front court of the output of pull-down pattern detection module PD and line buffer output, I, J) and brightness value Y (F-1, the I of former frame corresponding fields, J), if pull-down pattern detection module PD is output as very, promptly when the sequence of front court for coming by the progressive frames conversion of signals by pull-down pattern, then described interpolation selects module WS directly to select to export (the F as the brightness value Y of front court according to the PATTERN of detection, I, J) or the brightness value Y of former frame corresponding fields (F-1, I, J); If pull-down pattern detection module PD be output as vacation and motion detection block MD output movement criterion MV (J)=0, promptly the interpolation point is static for T, I, then described interpolation select module WS be output as line buffer output the brightness value Y when the front court (F, I, J); If pull-down pattern detection module PD be output as vacation and motion detection block MD output movement criterion MV (T, I, J)=1, be that the interpolation point is the motor point, then described interpolation select interpolation point that module WS is output as edge detection module output when the brightness discreet value Y ' of front court (F, I, J).
5. the video image deinterlacing device that detects as Motion Adaptive as described in the claim 4, its feature also comprises the denoising filtration module, the brightness value Y (F when the front court of described denoising filtration module line of input buffer output, I, J) and interpolation select the output of module, post-processing unit is given in the filtering of output denoising;
In the capable pixel of the I-1 of previous field, to choose from the odd number of pixels point for the treatment of that difference point is nearest, its brightness value is designated as Y (F-1, I-1, J-M), Y (F-1, I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1, J+1), Y (F-1, I-1, J+M), in the capable pixel of the I+1 of previous field, choose pixel with the capable correspondence position of I-1, its brightness value note Y (F-1, I+1, J-M), Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1), Y (F-1, I+1, J+M), wherein M is the positive integer greater than 1, and choose preceding 1 YO when front court F interpolation point (F, I, J-1), get the rightmost initial value YO of every row (F, I ,-1)=Y " (F; I, 0);
If Y " (F, I, J)+DN_TH0＜=MINIMUM (Y (F-1, I-1, J-M) ..., Y (F-1; I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1; J+1) ..., Y (F-1, I-1, J+M), (F-1; I+1, J-M), ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), YO (F, I, J-1));
Or Y " (F, I, J)-DN_TH0＞=MAXIMUM (Y (F-1, I-1, J-M) ..., Y (F-1; I-1, J-1), Y (F-1, I-1, J), Y (F-1, I-1; J+1) ..., Y (F-1, I-1, J+M), (F-1; I+1, J-M) ..., Y (F-1, I+1, J-1), Y (F-1, I+1, J), Y (F-1, I+1, J+1) ..., Y (F-1, I+1, J+M), YO (F, I, J-1));
DN_TH0 is come configurable the 6th threshold value by experiment;
Then denoising filtering result is: and YO (F, I, J)=Y (F-1, I-1, J)/4+Y (F-1, I+1, J)/4+YO (F, I, J-1)/2; Otherwise be YO (F, I, J)=Y " (F, I, J).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110148141 CN102186046A (en) | 2011-05-31 | 2011-05-31 | Video image de-interlacing scanning method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110148141 CN102186046A (en) | 2011-05-31 | 2011-05-31 | Video image de-interlacing scanning method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102186046A true CN102186046A (en) | 2011-09-14 |
Family
ID=44572107
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110148141 Pending CN102186046A (en) | 2011-05-31 | 2011-05-31 | Video image de-interlacing scanning method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102186046A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102496165A (en) * | 2011-12-07 | 2012-06-13 | 四川九洲电器集团有限责任公司 | Method for comprehensively processing video based on motion detection and feature extraction |
| CN106303338A (en) * | 2016-08-19 | 2017-01-04 | 天津大学 | A kind of in-field deinterlacing method based on the multi-direction interpolation of bilateral filtering |
| CN111294545A (en) * | 2019-03-13 | 2020-06-16 | 展讯通信（上海）有限公司 | Image data interpolation method and device, storage medium and terminal |
| CN113286107A (en) * | 2021-02-07 | 2021-08-20 | 合肥移瑞通信技术有限公司 | Video de-interlacing method, system, device and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1984305A (en) * | 2005-12-12 | 2007-06-20 | 深圳艾科创新微电子有限公司 | Three-dimensional self-adaptive motion inspection for eliminating space |
| CN101031029A (en) * | 2007-04-04 | 2007-09-05 | 华为技术有限公司 | Method and apparatus for processing video image |
| CN101309385A (en) * | 2008-07-09 | 2008-11-19 | 北京航空航天大学 | Alternate line eliminating process method based on motion detection |
| CN101442649A (en) * | 2009-01-08 | 2009-05-27 | 杭州华三通信技术有限公司 | Processing method and apparatus for removing alternate line and FPGA chip |
| CN101483745A (en) * | 2008-12-22 | 2009-07-15 | 四川虹微技术有限公司 | Video mode recognition method based on edge detection |
-
2011
- 2011-05-31 CN CN 201110148141 patent/CN102186046A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1984305A (en) * | 2005-12-12 | 2007-06-20 | 深圳艾科创新微电子有限公司 | Three-dimensional self-adaptive motion inspection for eliminating space |
| CN101031029A (en) * | 2007-04-04 | 2007-09-05 | 华为技术有限公司 | Method and apparatus for processing video image |
| CN101309385A (en) * | 2008-07-09 | 2008-11-19 | 北京航空航天大学 | Alternate line eliminating process method based on motion detection |
| CN101483745A (en) * | 2008-12-22 | 2009-07-15 | 四川虹微技术有限公司 | Video mode recognition method based on edge detection |
| CN101442649A (en) * | 2009-01-08 | 2009-05-27 | 杭州华三通信技术有限公司 | Processing method and apparatus for removing alternate line and FPGA chip |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102496165A (en) * | 2011-12-07 | 2012-06-13 | 四川九洲电器集团有限责任公司 | Method for comprehensively processing video based on motion detection and feature extraction |
| CN106303338A (en) * | 2016-08-19 | 2017-01-04 | 天津大学 | A kind of in-field deinterlacing method based on the multi-direction interpolation of bilateral filtering |
| CN106303338B (en) * | 2016-08-19 | 2019-03-22 | 天津大学 | A kind of in-field deinterlacing method based on the multi-direction interpolation of bilateral filtering |
| CN111294545A (en) * | 2019-03-13 | 2020-06-16 | 展讯通信（上海）有限公司 | Image data interpolation method and device, storage medium and terminal |
| CN111294545B (en) * | 2019-03-13 | 2021-03-23 | 展讯通信（上海）有限公司 | Image data interpolation method and device, storage medium and terminal |
| CN113286107A (en) * | 2021-02-07 | 2021-08-20 | 合肥移瑞通信技术有限公司 | Video de-interlacing method, system, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7769089B1 (en) | Method and system for reducing noise level in a video signal | |
| RU2419243C1 (en) | Device and method to process images and device and method of images display | |
| JP2005208613A (en) | Adaptive display controller | |
| US20080174694A1 (en) | Method and apparatus for video pixel interpolation | |
| JP2002125201A (en) | Motion adaptive median filter for interlace to progressive scan conversion | |
| CN101600061A (en) | De-interlaced method of video motion-adaptive and device | |
| US10440318B2 (en) | Motion adaptive de-interlacing and advanced film mode detection | |
| CN102186046A (en) | Video image de-interlacing scanning method and device | |
| US20080165278A1 (en) | Human visual system based motion detection/estimation for video deinterlacing | |
| US7362377B2 (en) | Spatio-temporal adaptive video de-interlacing | |
| TWI255136B (en) | De-interlacing module and related method capable of adaptively de-interlacing a video field | |
| US8704945B1 (en) | Motion adaptive deinterlacer | |
| CN101282415B (en) | Motion adaptive image processing | |
| KR100540380B1 (en) | Apparatus and method for intra field interpolation of deinterlacer | |
| CN101232570B (en) | Apparatus and method of motion adaptive image processing | |
| GB2444532A (en) | Motion adaptive image processing detecting motion at different levels of sensitivity | |
| JP2008009182A (en) | Liquid crystal display device | |
| CN101483747B (en) | Movement detection method suitable for deinterlacing technique | |
| Zhang et al. | An efficient motion adaptive deinterlacing algorithm using improved edge-based line average interpolation | |
| US8045820B2 (en) | System and method for edge direction detection for spatial deinterlace | |
| CN101247469B (en) | Motion adaptive image process | |
| CN103686043B (en) | The method and apparatus that reciprocal cross mistake is carried out to video | |
| TW201325218A (en) | Image processing method and associated image processing apparatus | |
| US8373798B2 (en) | Text protection device and related motion adaptive de-interlacing device | |
| TW201021565A (en) | Apparatus and method for motion adaptive interlacing |
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: 20110914 |