CN104580978A - Video detecting and processing method and video detecting and processing device - Google Patents
Video detecting and processing method and video detecting and processing device Download PDFInfo
- Publication number
- CN104580978A CN104580978A CN201510071223.1A CN201510071223A CN104580978A CN 104580978 A CN104580978 A CN 104580978A CN 201510071223 A CN201510071223 A CN 201510071223A CN 104580978 A CN104580978 A CN 104580978A
- Authority
- CN
- China
- Prior art keywords
- field
- pixel
- comb
- frame
- module
- 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)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention relates to a video detecting and processing method. The video detecting and processing method comprises the following steps: calculating the whole dependency of images of adjacent fields to predicate whether a current field is a film field or not; if so, combining the current field and the adjacent field to form a frame; detecting comb tooth artifacts in the synthesized frame pixel by pixel; if no comb tooth artifact is detected, determining that a current local region is a film pattern and taking the synthesized frame as a restored video frame; if the comb tooth artifacts are detected, judging that the current local region is a non-film pattern; and calculating an interpolation frame by adopting a movement self-adaptive method or a spatial interpolation method and taking the interpolation frame as a restored video frame. The video detecting and processing method can be used for correctly detecting a film region and an interlacing region in a mixed video and different interlacing removing technologies are used for processing respectively, so that the details of the film region are recovered and a comb tooth phenomenon is avoided. Meanwhile, a film pattern detection module and an interlacing removing module can use the same input so that field caching and DDR bandwidth are reduced, and the hardware cost is saved.
Description
Technical field
The present invention relates to video mode detection field, particularly relate to a kind of video and detect and processing method, device.
Background technology
Film normally press row-by-row system with per second 24 frames shooting.In order to movie program glibly on a television set, need moving-picture frame to be split as TV Field signal and carry out broadcast transmission, this technology is called drop-down conversion process, and common comprises for the drop-down conversion of 3:2 in the broadcast of TSC-system formula with for the drop-down converter technique of 2:2 in pal mode broadcast.Example is transformed to, if moving-picture frame sequence is with A, B, C, D so that 3:2 is drop-down ... represent, as shown in Figure 1, first every frame is splitted into two, as A
tand A
b, wherein subscript t and b represents field, top and field, the end respectively, and the TV Field signal obtained after the drop-down conversion process of 3:2 is A
t, A
b, A
t, B
b, B
t, C
t, C
b, C
t, D
t, D
b, D
t, its rate is 60 per second.The drop-down converter technique of 2:2 in like manner.
Film mode is detected as follow-up de interlacing process and provides necessary information.If known current pending field F
tbe film video, and also know with this from other of same moving-picture frame it is front court F
t-1or back court F
t+1, so only need simply these two weave ins can be restored original moving-picture frame, this process is called anti-drop-down conversion process.The deinterlacing technique (as field interpolation or Motion Adaptive de interlacing etc.) conventional from common TV video is different, and anti-drop-down conversion process farthest can keep image detail.But, then can cause comb pseudomorphism to doing anti-drop-down conversion process with non-movie field.It is generally acknowledged, film mode occurs that the result (being detected as film by non-movie video) that flase drop causes is more serious than undetected (being detected as non-movie by film video).
According to the general principle of above-mentioned drop-down conversion, the content from the adjacent two field picture of same moving-picture frame often has stronger correlation, and (two of synchronization record, as the A in Fig. 1
tand A
b), and from adjacent two correlations of different moving-picture frame less (two that do not record in the same time, as the A of Fig. 1
tand B
b), these strong correlations simultaneously and weak dependence show certain periodicity in continuous print film sequence.If represent that the back court correlation of current field and next-door neighbour is comparatively strong with 1, represent that the back court correlation of current field and next-door neighbour is more weak with 0, then shown in analysis chart 1, TV Field can find, 3:2 drop-down conversion film sequence has following correlation rule:
11010 11010 11010 11010 ……
Respectively with first A
t, A
b, second A
t, B
bfield and B
tas initial field, then 3:2 drop-down conversion film sequence has five kinds of possible correlation rules: 11010,10101,01011,10110,01101, defines these five kinds of patterns for next-door neighbour field correlation reference model.Visible, if the vision signal of input is the drop-down conversion film sequence of 3:2, then every continuous print five field picture must meet the one in above five kinds of next-door neighbour field correlation reference models, and often continuous five field picture afterwards all have correlation rule identical with it.
In addition, owing to often always having a content (as A in continuous five in 3:2 drop-down conversion film sequence
t) repeat, so by adding up the correlation between the front court of current field and back court continuously, also certain periodic regularity can be found.Still Figure 1 shows that example, if represent, between the front court of current field and back court, there is strong correlation with 1, and represent, between the front court of current field and back court, there is comparatively weak dependence with 0, respectively with first A
t, A
b, second A
t, B
bfield and B
tas initial field, can find five kinds of possible correlation rules: 01000,10000,00001,00010,00100, defining these five kinds of matching grating is field, interval correlation reference model.Visible, if the vision signal of input is the drop-down conversion film sequence of 3:2, then every continuous print five field picture must meet the one in the correlation reference model of field, above five kinds of intervals, and often continuous five field picture afterwards all have correlation rule identical with it.
The overall relevancy of existing most film mode detection technique also just by calculating opposite field image judges whether current field is film field.Fig. 2 gives film mode conventional at present and detects and de interlacing flow process, and wherein film mode detection module needs three field data inputs as F
t, F
t-1and F
t-2to calculate the correlation between opposite field, the output of module is to field F
t-1testing result, whether comprise it is film field and and F
tor F
t-2from same moving-picture frame.De interlacing module also needs three field data inputs as F
t-1, F
t-2and F
t-3to carry out the process of Motion Adaptive de interlacing, it exports F
t-2* F is represented
t-2the interpolation result frame of field, now needs to obtain F from film mode detection module
t-2the film mode information of field correspondence is (with M
t-2represent), if F
t-2field is film field, just adopts anti-drop-down transform method to obtain progressive frame, otherwise just adopts Motion Adaptive method to calculate interpolated frame.Visible, in prior art, the result of de interlacing module always has the delay of than the result of film mode detection module, this needs more field buffer memory and DDR bandwidth.
On the other hand, the interlacing information that later stage compilation is added often is comprised in the mute of actual transmissions, the station symbol etc. that such as captions, scroll bar, sign language window even rotate.Prior art only gives the testing result of whole field, and this makes de interlacing module also be processed as film mode in the interlacing region in these mixed videos by mistake, thus occurs comb pseudomorphism (as Figure 11).
Summary of the invention
In view of the problems of the prior art, in first aspect, the embodiment of the present invention provides a kind of video to detect and processing method, judge whether current field is film field by the overall relevancy calculating opposite field image, when being judged as film field, current field and opposite field are merged framing, then in this synthetic frame by pixel detection comb pseudomorphism, if without comb pseudomorphism, then determine that current regional area is film mode, using the frame of video of described synthetic frame as reduction; If there is comb pseudomorphism, then judge that current regional area is non-film mode, now adopt Motion Adaptive method or spatial domain interpolation method to calculate interpolated frame, as the frame of video of reduction, described current regional area comprises described current field and opposite field.
Preferably, describedly to comprise by the method for pixel detection comb pseudomorphism in this synthetic frame: obtain the pixel characteristic value merged in frame in current pixel NxN neighborhood; Calculate the pixel characteristic value intermediate value of first direction three neighborhood territory pixel of current pixel, calculate the pixel characteristic value intermediate value of two pixels first direction three neighborhood territory pixel separately the most adjacent with current pixel in a first direction; Calculate the absolute value of the difference of the pixel characteristic value of each intermediate value and each center pixel respectively; These three absolute values and threshold value are compared, if any one is greater than threshold value in them, think that first direction detects comb, namely comb pseudomorphism is had, otherwise think that this direction does not detect comb, namely without comb pseudomorphism, described N is positive integer, and described NxN represents a rectangular area of putting centered by current pixel.
Be noted that, here so-called pixel characteristic value refers to the property value of each pixel in image, such as each pixel can use Y, Cb, Cr tri-component values, or R, G, B tri-component values (generally can be provided by collecting device or head end video decoder) represent, and are separate relations between these three components (or being called passage).The present invention both can select separately Y or Cb or Cr or R or G or channel B to process as pixel characteristic value herein, also the pixel characteristic value of wherein several combination of channels Cheng Xin can be processed.That is, preferably, described pixel characteristic value is YCbCr value or rgb value.If represent described pixel characteristic value with P, in the embodiment of the present invention, can select separately certain component of YCbCr or RGB such as, as pixel characteristic value, P=Y; Also can by the pixel characteristic value of wherein several component combination Cheng Xin, such as P=(R+G+B)/3, is not particularly limited at this.
Preferably, described first direction is vertical direction, horizontal direction or other directions.
In second aspect, the present invention also provides a kind of video to detect and processing unit, comprise film mode detection module, film mode optimizes module and de interlacing module, described film mode is optimized module and is comprised generation merging frame module and comb detection module, described film mode detection module, when being film field for detecting current field by the overall relevancy of calculating opposite field image, generation testing result being sent to described film mode to optimize in module merges frame module; Described generation merges frame module, for when receiving film field testing result, current field is adjacent occasion and framing, and sends to described comb detection module by merging the frame generated; Described comb detection module, for the synthetic frame received by pixel detection comb pseudomorphism, if there is comb pseudomorphism, then determine that current regional area is non-film mode, and by determination result notify described in de interlacing module; Described de interlacing module, when being non-movie pattern for receiving determination result, adopt Motion Adaptive method or spatial domain interpolation method to calculate interpolated frame, as the frame of video of reduction, described current regional area comprises described current field and opposite field.
Preferably, described comb detection module carries out comprising by the step of pixel detection comb pseudomorphism to synthetic frame: obtain the pixel characteristic value merged in frame in current pixel NxN neighborhood; Calculate the pixel characteristic value intermediate value of first direction three neighborhood territory pixel of current pixel, calculate the pixel characteristic value intermediate value of two pixels first direction three neighborhood territory pixel separately the most adjacent with current pixel in a first direction; Calculate the absolute value of the difference of the pixel characteristic value of each intermediate value and each center pixel respectively; These three absolute values and threshold value are compared, if any one is greater than threshold value in them, think that first direction detects comb, namely comb pseudomorphism is had, otherwise think that this direction does not detect comb, namely without comb pseudomorphism, described N is positive integer, and described NxN represents a rectangular area of putting centered by current pixel.
In the third aspect, the present invention also provides a kind of video to detect and processing unit, comprise film mode detection module, film mode optimizes module and de interlacing module, described film mode is optimized module and is comprised film mode prediction module, generate and merge frame module and comb detection module, described film mode detection module, when being film field for detecting current field by the overall relevancy of calculating opposite field image, testing result is sent to the film mode prediction module in described film mode optimization module; Described film mode prediction module, for based between opposite field film mode information exist periodicity, doping next field according to aforementioned film field testing result is film mode, and this is predicted the outcome send to generate merge frame module; Described generation merges frame module, for when receiving film field prediction result, next described field is adjacent occasion and framing, and sends to described comb detection module by merging the frame generated; Described comb detection module, for the synthetic frame received by pixel detection comb pseudomorphism, if there is comb pseudomorphism, then determine that current regional area is non-film mode, and by determination result notify described in de interlacing module; Described de interlacing module, when being non-movie pattern for receiving determination result, adopt Motion Adaptive method or spatial domain interpolation method to calculate interpolated frame, as the frame of video of reduction, described current regional area comprises next described field and opposite field thereof.
Preferably, described comb detection module carries out comprising by the step of pixel detection comb pseudomorphism to synthetic frame: obtain the pixel characteristic value merged in frame in current pixel NxN neighborhood; Calculate the pixel characteristic value intermediate value of first direction three neighborhood territory pixel of current pixel, calculate the pixel characteristic value intermediate value of two pixels first direction three neighborhood territory pixel separately the most adjacent with current pixel in a first direction; Calculate the absolute value of the difference of the pixel characteristic value of each intermediate value and each center pixel respectively; These three absolute values and threshold value are compared, if any one is greater than threshold value in them, think that first direction detects comb, namely comb pseudomorphism is had, otherwise think that this direction does not detect comb, namely without comb pseudomorphism, described N is positive integer, and described NxN represents a rectangular area of putting centered by current pixel.
Preferably, described pixel characteristic value is YCbCr value or rgb value.
Preferably, described first direction is vertical direction, horizontal direction or other directions.
Accompanying drawing explanation
Fig. 1 is the drop-down conversion process of 3:2 of prior art and anti-drop-down conversion process schematic diagram;
Fig. 2 is that the film mode of prior art detects and de-interlacing apparatus schematic diagram;
Fig. 3 is a kind of device schematic diagram of the embodiment of the present invention;
Fig. 4 is film mode optimization device schematic diagram in Fig. 3;
Fig. 5 is film mode detection method flow chart in Fig. 3;
Fig. 6 is the next-door neighbour field described in the embodiment of the present invention and field, interval reference point position relationship schematic diagram;
Fig. 7 is the another kind of device schematic diagram of the embodiment of the present invention;
Fig. 8 is the film mode optimization device schematic diagram described in Fig. 7;
Fig. 9 is the another kind of device schematic diagram of the embodiment of the present invention;
Figure 10 is the another kind of device schematic diagram of the embodiment of the present invention;
Figure 11 is the comb pseudomorphism schematic diagram that film mode flase drop causes motion interlacing region to occur;
Figure 12 is that the comb described in the embodiment of the present invention detects location of pixels relation schematic diagram;
Figure 13 is the comb testing process schematic diagram described in the embodiment of the present invention;
Figure 14 is the spatial domain interpolating pixel position relationship schematic diagram described in the embodiment of the present invention.
Embodiment
Be further elaborated the embodiment of the present invention based on accompanying drawing below, these elaborations are to allow reader be easier to understand the present invention, and can not be considered as being limitation of the present invention.
The present invention proposes a kind of local film mode detection method detected based on comb, predicting the outcome of film mode is corrected, correctly can detect the movie field in mixed video and interlacing region, and adopt different deinterlacing techniques to process respectively, thus both recovered the details of movie field, avoid again producing interlacing jagged edge.Meanwhile, predict the outcome because film mode detects only need provide, film mode detection module and de interlacing module can use identical input, decrease a buffer memory and DDR bandwidth, have saved hardware implementation cost.
Owing to interlaced video to be considered as after film mode carries out anti-drop-down conversion process, there will be jagged edge.Basic ideas of the present invention are exactly after being judged as film mode based on whole field correlation, increase local comb detection module, detect the interlaced content region in mixed video.Again owing to there being comb detection module, so film mode detection module is without the need to providing accurate current field testing result, only need provides predicted value, thus saving a buffer memory and bandwidth.
Fig. 3 is a kind of device schematic diagram of the embodiment of the present invention.It represents that the film mode that the present invention can be proposed is optimized module and is appended to after existing film mode detection module, can detect current interpolation field F
t-2in interlacing region.Wherein M
t-2to F
t-2the overall film mode Detection Information of field; The film mode provided at Fig. 4 is optimized in module, first according to M
t-2information is by current field and opposite field (specifically F
t-1field or F
t-3, by M
t-2information indicate) merge framing (anti-drop-down conversion process), then in this synthetic frame by pixel detection comb pseudomorphism, according to judging with or without comb pseudomorphism whether current regional area is film mode, M
t-2* be exactly the local film pattern information exporting to follow-up de interlacing module.
Fig. 5 provides a kind of film mode detection method of the embodiment of the present invention, and the method comprises the steps:
Step 1: definition correlation reference model ordered set, comprise next-door neighbour field correlation reference model collection PATTERN_FIELD={11010,10101,01011,10110,01101} and field, interval correlation reference model collection PATTERN_FRAME={01000,10000,00001,00010,00100}, wherein each reference model is a 5bit number;
Step 2: the opposite field correlation detecting every input picture, comprises next-door neighbour's field correlation and field, interval correlation;
Step 3: the relevance detection results of Coutinuous store five field picture, forms correlation sequence, is compared by the correlation reference model defined, the film mode testing result of acquisition current field in every correlation sequence and the step 1 of continuous five.
The next-door neighbour field correlation of the every field picture of the detection described in above-mentioned steps 2, also comprises following concrete steps:
Step 2.1: with the F inputted
t-1field is current field, with its back court F
tfield is next-door neighbour field, calculates the pending point of current field and the absolute difference being close to field reference point;
Step 2.2: a little, absolute difference step 2.1 calculated adds up the institute in traversal current field, calculates the absolute difference sum Sum_diff_field of next-door neighbour field;
Step 2.3: the absolute difference sum of next-door neighbour field and a threshold value T1 are compared, determine the next-door neighbour field relevance detection results of current field: if absolute difference sum is less than threshold value T1, represent that current field is comparatively strong with next-door neighbour field correlation, field relevance detection results will be close to and represent with " 1 "; Otherwise represent that correlation is more weak, field relevance detection results will be close to and represent with " 0 ".
Calculating current field described in above-mentioned steps 2.1 and the absolute difference being close to field reference point, concrete grammar also comprises:
With the F of input
t-1field is current field, with its back court F
tfield is next-door neighbour field, the characteristic value of the pending point of current field is represented with P0, and choose up and down two points adjacent in the vertical direction with current pending point in next-door neighbour field, its characteristic value represents with P1 and P2 respectively.The position relationship of P0, P1 and P2 as shown in Figure 6.So the pending point of current field is as follows with absolute difference (representing with the diff_field) computing formula of next-door neighbour field reference point:
diff_field＝abs(P0-(P1+P2)/2)，
Wherein abs (.) represents the operation that takes absolute value.
Field, the interval correlation of the every field picture of the detection described in above-mentioned steps 2, also comprises following concrete steps:
Step 2.4: with the F inputted
t-1field is current field, with the front court F of current field
t-2field and back court F
tfield is field, interval, calculates the absolute difference of front court reference point and back court reference point;
Step 2.5: a little, absolute difference step 2.4 calculated adds up the institute in traversal current field, the absolute difference sum Sum_diff_frame of counting period field;
Step 2.6: the absolute difference sum of field, interval and a threshold value are compared, determine field, the interval relevance detection results of current field: if absolute difference sum is less than this threshold value, represent that front court and the back court correlation of current field are comparatively strong, field, interval relevance detection results is represented with " 1 "; Otherwise represent that correlation is more weak, field, interval relevance detection results is represented with " 0 ".
Calculating current field described in above-mentioned steps 2.4 and the absolute difference being close to field reference point, concrete grammar also comprises:
With the F of input
t-1field is current field, with the front court F of current field
t-2field and back court F
tfield is field, interval, and in field, interval, choose up and down two points adjacent in the vertical direction with current pending some P0 respectively, its characteristic value represents with P1 and P2 and P3 and P4 respectively.The position relationship of P0, P1, P2, P3 and P4 as shown in Figure 6.So the pending point of current field is as follows with absolute difference (representing with the diff_frame) computing formula of next-door neighbour field reference point:
diff_frame＝abs((P1+P2)/2-(P3+P4)/2)，
Wherein abs (.) represents the operation that takes absolute value.
Above-mentioned steps 3 also comprises following concrete steps:
Step 3.1: using the next-door neighbour field relevance detection results of current field as lowest order, forms a 5bit number with the next-door neighbour field relevance detection results of continuous four above, is namely close to field correlation sequence S;
Step 3.2: using field, the interval relevance detection results of current field as lowest order, and field, the interval relevance detection results of continuous four forms a 5bit number, i.e. field, interval correlation sequence Q above;
Step 3.3: if meet following three conditions simultaneously, can judge that current field is film mode: being 1. close to field correlation sequence S, identical with i-th pattern in the next-door neighbour field correlation reference model collection PATTERN_FIELD defined in step 1 (wherein i is the mode index in PATTERN_FIELD, i=1,2 ... 5); 2. the jth pattern in field, the interval correlation reference model collection PATTERN_FRAME defined in field, interval correlation sequence Q and step 1 identical (wherein j is the mode index in PATTERN_FRAME, j=1,2 ... 5); 3. i=j.Such as, if S=10101, Q=10000, because of S=PATTERN_FIELD (i), Q=PATTERN_FRAME (j), and i=j=2, therefore can determine that current field is film mode.If above three conditions can not meet simultaneously, can judge that current field is non-film mode.Such as, if S=10101, Q=01000, although there is S=PATTERN_FIELD (i), i=2, and Q=PATTERN_FRAME (j), j=1, because of i ≠ j, therefore think that current field is non-movie pattern.
Step 3.4: determine current field F
t-1film mode testing result M
t-1: if current field is film mode, M
t-1=i, wherein i is the mode index described in step 3.3; If current field non-movie pattern, then M
t-1=0.
In figure 3, film mode optimization module and de interlacing module are all with the F of input
t-2field is current pending field, F
t-2the film mode testing result of field correspondence is M
t-2, namely by M
t-1result after postponing through one.If it is film mode that current field belongs to, according to M
t-2information can be determined and F
t-2field source is its front court F from same moving-picture frame
t-3or back court F
t-1.Specifically comprise the steps:
Step 1: by M
t-2as mode index, in above-mentioned next-door neighbour field correlation reference model collection PATTERN_FIELD, choose corresponding reference model E=PATTERN_FIELD (M
t-2);
Step 2: the lowest order according to reference model E judges F
t-2field merges framing by with its front court or back court: if the lowest order of E is 1, then by F
t-2field and its back court F
t-1merge; If the lowest order of E is 0, then by F
t-2field and its front court F
t-3merge.
For example, if M
t-2=3, then reference model E=PATTERN_FIELD (3)=01011, the lowest order due to E is 1, then by F
t-2field and its back court F
t-1merge; If M
t-2=4, then reference model E=PATTERN_FIELD (3)=10110, the lowest order due to E is 0, then by F
t-2field and its front court F
t-3merge.
Fig. 7 is the another kind of device schematic diagram of the embodiment of the present invention.The comb pseudomorphism in picture can be detected due to comb detection module, even if the film mode detection module generation flase drop before it, also can correct by comb detection module, therefore can input three identical field data (as F to film mode detection and de interlacing module simultaneously
t, F
t-1and F
t-2), thus save a buffer memory and respective bandwidth.
The film mode provided at Fig. 8 is optimized in module, and from film mode detection module is F
t-2the film mode information M of field
t-2, first according to the periodicity that the film mode information between opposite field exists, by M
t-2dope next F
t-1film mode information M
t-1', then according to M
t-1' information is by current field F
t-1with F
tor F
t-2merge framing, finally by the comb region in this synthetic frame of pixel detection, revise overall M
t-1' information is local film pattern information M
t-1*, for follow-up to F
t-1field uses when doing de interlacing process.
Film mode prediction module in Fig. 8 uses following computing formula by M
t-2predict next F
t-1film mode information M
t-1':
For example, if from the F of film mode detection module
t-2the film mode information M of field
t-2be 2, then film mode prediction module is according to next F of above-mentioned formula predictions
t-1film mode information M
t-1' be 3.
Fig. 9 and Figure 10 is other two kinds of device schematic diagrames of the embodiment of the present invention.They represent that film mode can be optimized Module-embedding uses in existing film mode detection module or de interlacing module, thus realize the detection of local film pattern information.
When there is film mode flase drop, opposite field merging treatment is done by as film mode in interlacing region, thus occurs jagged edge.Figure 11 shows the comb pseudomorphism of the three kinds of angles often occurred in the moving object merged in frame, respectively corresponding (a) 135 degree of comb, (b) 90 degree of comb, (c) 45 degree of comb.Certainly can also detect other more low-angle comb, the comb angle of detection is more, and result is more accurate, but realizes cost and can increase.
Figure 13 provides the flow chart of the algorithm that a kind of local comb detects, and is detected as example with 90 degree of direction comb:
First, the pixel characteristic value merged in frame in current pixel 5x5 neighborhood is obtained.Suppose that current pixel is (x, y) merging the position coordinates in frame, the location of pixels relation in its 5x5 neighborhood as shown in figure 12.In figure, square represents and belongs to current field (as F
t-1) pixel, circle represents and the pixel belonging to opposite field, and opposite field may be that front court is (as F
t-2) also can be back court (F
t), this decides according to film detection module Output rusults.Pixel characteristic value P (x, y) can be YCbCr value or rgb value etc., and wherein Y represents the luminance component of pixel, Cb and Cr represents two chromatic components of pixel respectively, and R, G, B then represent the red component of pixel, green component and blue component respectively.
Secondly, respectively centered by (x, y-1), (x, y) and (x, y+1), calculate three neighborhood intermediate values of vertical direction pixel:
M_90_1＝median(P(x,y-2),P(x,y-1),P(x,y))
M_90_2＝median(P(x,y-1),P(x,y),P(x,y+1))
M_90_3＝median(P(x,y),P(x,y+1),P(x,y+2))
Wherein, median (.) gets median function.
Again, the absolute value of the difference of each intermediate value and each center pixel is calculated respectively:
D_90_1＝abs(M_90_1–P(x,y-1))
D_90_2＝abs(M_90_2–P(x,y))
D_90_3＝abs(M_90_3–P(x,y+1))
Wherein, abs (.) is the function that takes absolute value.
Finally, these three absolute values and threshold value are compared, if any one is greater than threshold value in them, just think that this direction detects comb:
if(D_90_1>T_90||D_90_2>T_90||D_90_3>T_90)
Cm_90＝1；
else
Cm_90＝0.
The threshold value of defined when wherein T_90 is 90 degree of direction comb detections, whether Cm_90 indicating positions (x, y) place exists 90 degree of direction comb pseudomorphisms.
Comb Cleaning Principle on other directions (as 45 degree and 135 degree) and said process similar.Testing result in last comprehensive all directions, as long as any one direction detects comb, just judges that comb has appearred in the region at current pixel place, and is labeled as non-movie pattern, i.e. Cm (x, y)=1; Otherwise Cm (x, y)=0.
According to local comb pseudomorphism indicated value Cm (x, y) calculated, when carrying out de interlacing process, if Cm (x, y)=0, adopting merging frame method to obtain result line by line, thus maintaining the details of movie field; If otherwise Cm (x, y)=1, adopt common deinterlacing technique (as spatial domain interpolation or Motion Adaptive method etc.) to obtain result line by line, thus avoid the appearance of comb pseudomorphism.
In an example, the common deinterlacing technique in the embodiment of the present invention adopts a kind of spatial domain interpolation method, comprises following concrete steps:
Step 1: with the F inputted
t-2field is current field, and the location of pixels relation in its interpolation point (coordinate is (x, y)) neighborhood as shown in figure 14.In figure, circle represents the pixel belonging to current field, and cross then represents the pixel of interpolation.Choose up and down two points adjacent in vertical direction with interpolation point, suppose that its characteristic value is respectively P (x, y-1) and P (x, y+1);
Step 2: the characteristic value P (x, y) asking for interpolation point:
P(x,y)＝(P(x,y-1)+P(x,y+1))/2；
Step 3: calculated the characteristic value of all interpolations point by step 1 and step 2 after, by acquisition interpolation field picture, merges framing by this interpolation field picture and current field, obtains result line by line.
Advantage of the present invention is: 1) can detect movie contents region and TV interlaced content region in mixed video, so that de interlacing module energy differentiating and processing; 2) Double Data Rate synchronous DRAM DDR buffer memory and bandwidth resources are saved.
Those skilled in the art can recognize, in conjunction with each exemplary unit/module that embodiment disclosed herein describes, can realize with electronic hardware, computer software or the combination of the two, in order to the interchangeability of hardware and software is clearly described, generally describe the composition of each example in the above description according to function.These functions perform with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme.Professional and technical personnel can use distinct methods to realize described function to each specifically should being used for, but this realization should not thought and exceeds scope of the present invention.
It should be noted last that, above embodiment is only in order to illustrate technical scheme of the present invention and unrestricted.Although with reference to embodiment to invention has been detailed description, those skilled in the art are to be understood that, modify to technical scheme of the present invention or equivalent replacement, do not depart from the spirit and scope of technical solution of the present invention, it all should be encompassed in the middle of right of the present invention.
Claims (10)
1. a video detects and processing method, judge whether current field is film field by the overall relevancy calculating opposite field image, it is characterized in that, when being judged as film field, current field and opposite field are merged framing, then in this synthetic frame by pixel detection comb pseudomorphism, if without comb pseudomorphism, then determine that current regional area is film mode, using the frame of video of described synthetic frame as reduction; If there is comb pseudomorphism, then judge that current regional area is non-film mode, now adopt Motion Adaptive method or spatial domain interpolation method to calculate interpolated frame, as the frame of video of reduction, described current regional area comprises described current field and opposite field.
2. video according to claim 1 detects and processing method, it is characterized in that: describedly comprise by the method for pixel detection comb pseudomorphism in this synthetic frame: obtain the pixel characteristic value merged in frame in current pixel NxN neighborhood; Calculate the pixel characteristic value intermediate value of first direction three neighborhood territory pixel of current pixel, calculate the pixel characteristic value intermediate value of two pixels first direction three neighborhood territory pixel separately the most adjacent with current pixel in a first direction; Calculate the absolute value of the difference of the pixel characteristic value of each intermediate value and each center pixel respectively; These three absolute values and threshold value are compared, if any one is greater than threshold value in them, think that first direction detects comb, namely comb pseudomorphism is had, otherwise think that this direction does not detect comb, namely without comb pseudomorphism, described N is positive integer, and described NxN represents a rectangular area of putting centered by current pixel.
3. video according to claim 2 detects and processing method, it is characterized in that: described pixel characteristic value is YCbCr value or rgb value.
4. the video according to Claims 2 or 3 detects and processing method, it is characterized in that: described first direction is vertical direction, horizontal direction or other directions.
5. a video detects and processing unit, comprise film mode detection module, film mode optimizes module and de interlacing module, described film mode is optimized module and is comprised generation merging frame module and comb detection module, it is characterized in that: described film mode detection module, when being film field for detecting current field by the overall relevancy of calculating opposite field image, generation testing result being sent to described film mode to optimize in module merges frame module; Described generation merges frame module, for when receiving film field testing result, current field is adjacent occasion and framing, and sends to described comb detection module by merging the frame generated; Described comb detection module, for the synthetic frame received by pixel detection comb pseudomorphism, if there is comb pseudomorphism, then determine that current regional area is non-film mode, and by determination result notify described in de interlacing module; Described de interlacing module, when being non-movie pattern for receiving determination result, adopt Motion Adaptive method or spatial domain interpolation method to calculate interpolated frame, as the frame of video of reduction, described current regional area comprises described current field and opposite field.
6. video according to claim 5 detects and processing unit, it is characterized in that: described comb detection module carries out comprising by the step of pixel detection comb pseudomorphism to synthetic frame: obtain the pixel characteristic value merged in frame in current pixel NxN neighborhood; Calculate the pixel characteristic value intermediate value of first direction three neighborhood territory pixel of current pixel, calculate the pixel characteristic value intermediate value of two pixels first direction three neighborhood territory pixel separately the most adjacent with current pixel in a first direction; Calculate the absolute value of the difference of the pixel characteristic value of each intermediate value and each center pixel respectively; These three absolute values and threshold value are compared, if any one is greater than threshold value in them, think that first direction detects comb, namely comb pseudomorphism is had, otherwise think that this direction does not detect comb, namely without comb pseudomorphism, described N is positive integer, and described NxN represents a rectangular area of putting centered by current pixel.
7. a video detects and processing unit, comprise film mode detection module, film mode optimizes module and de interlacing module, described film mode is optimized module and is comprised film mode prediction module, generate and merge frame module and comb detection module, it is characterized in that: described film mode detection module, when being film field for detecting current field by the overall relevancy of calculating opposite field image, testing result is sent to the film mode prediction module in described film mode optimization module; Described film mode prediction module, for based between opposite field film mode information exist periodicity, doping next field according to aforementioned film field testing result is film mode, and this is predicted the outcome send to generate merge frame module; Described generation merges frame module, for when receiving film field prediction result, next described field is adjacent occasion and framing, and sends to described comb detection module by merging the frame generated; Described comb detection module, for the synthetic frame received by pixel detection comb pseudomorphism, if there is comb pseudomorphism, then determine that current regional area is non-film mode, and by determination result notify described in de interlacing module; Described de interlacing module, when being non-movie pattern for receiving determination result, adopt Motion Adaptive method or spatial domain interpolation method to calculate interpolated frame, as the frame of video of reduction, described current regional area comprises next described field and opposite field thereof.
8. video according to claim 7 detects and processing unit, it is characterized in that: described comb detection module carries out comprising by the step of pixel detection comb pseudomorphism to synthetic frame: obtain the pixel characteristic value merged in frame in current pixel NxN neighborhood; Calculate the pixel characteristic value intermediate value of first direction three neighborhood territory pixel of current pixel, calculate the pixel characteristic value intermediate value of two pixels first direction three neighborhood territory pixel separately the most adjacent with current pixel in a first direction; Calculate the absolute value of the difference of the pixel characteristic value of each intermediate value and each center pixel respectively; These three absolute values and threshold value are compared, if any one is greater than threshold value in them, think that first direction detects comb, namely comb pseudomorphism is had, otherwise think that this direction does not detect comb, namely without comb pseudomorphism, described N is positive integer, and described NxN represents a rectangular area of putting centered by current pixel.
9. video according to claim 8 detects and processing unit, it is characterized in that: described pixel characteristic value is YCbCr value or rgb value.
10. video according to claim 8 or claim 9 detects and processing unit, it is characterized in that: described first direction is vertical direction, horizontal direction or other directions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510071223.1A CN104580978B (en) | 2015-02-11 | 2015-02-11 | A kind of video detection and processing method, device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510071223.1A CN104580978B (en) | 2015-02-11 | 2015-02-11 | A kind of video detection and processing method, device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104580978A true CN104580978A (en) | 2015-04-29 |
| CN104580978B CN104580978B (en) | 2017-12-22 |
Family
ID=53096060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510071223.1A Active CN104580978B (en) | 2015-02-11 | 2015-02-11 | A kind of video detection and processing method, device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104580978B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105227887A (en) * | 2015-08-26 | 2016-01-06 | 华为技术有限公司 | A kind of drop-down film source processing method and device |
| CN106488312A (en) * | 2015-08-31 | 2017-03-08 | 展讯通信（上海）有限公司 | The detection method of video source and device |
| WO2017101348A1 (en) * | 2015-12-14 | 2017-06-22 | 乐视控股（北京）有限公司 | Method and device for deinterlacing interlaced videos |
| CN111476803A (en) * | 2020-04-14 | 2020-07-31 | 展讯通信（上海）有限公司 | Video processing method and related equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1937753A (en) * | 2006-10-25 | 2007-03-28 | 北京中星微电子有限公司 | Foldable field sequence detecting method and digital video deinterleaving device and method |
| CN101483790A (en) * | 2008-12-22 | 2009-07-15 | 四川虹微技术有限公司 | Movie mode video signal detection method |
| CN101729841A (en) * | 2008-10-27 | 2010-06-09 | 富士通株式会社 | Comb-like pseudoscopic image detection apparatus and method |
| US20130002957A1 (en) * | 2011-06-30 | 2013-01-03 | Stmicroelectronics Asia Pacific Pte Ltd. | Apparatus and method for local video detector for mixed cadence sequence |
| CN103826082A (en) * | 2014-01-21 | 2014-05-28 | 华为技术有限公司 | Video processing method and device |
-
2015
- 2015-02-11 CN CN201510071223.1A patent/CN104580978B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1937753A (en) * | 2006-10-25 | 2007-03-28 | 北京中星微电子有限公司 | Foldable field sequence detecting method and digital video deinterleaving device and method |
| CN101729841A (en) * | 2008-10-27 | 2010-06-09 | 富士通株式会社 | Comb-like pseudoscopic image detection apparatus and method |
| CN101483790A (en) * | 2008-12-22 | 2009-07-15 | 四川虹微技术有限公司 | Movie mode video signal detection method |
| US20130002957A1 (en) * | 2011-06-30 | 2013-01-03 | Stmicroelectronics Asia Pacific Pte Ltd. | Apparatus and method for local video detector for mixed cadence sequence |
| CN103826082A (en) * | 2014-01-21 | 2014-05-28 | 华为技术有限公司 | Video processing method and device |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105227887A (en) * | 2015-08-26 | 2016-01-06 | 华为技术有限公司 | A kind of drop-down film source processing method and device |
| CN105227887B (en) * | 2015-08-26 | 2018-08-21 | 华为技术有限公司 | A kind of drop-down film source processing method and processing device |
| CN106488312A (en) * | 2015-08-31 | 2017-03-08 | 展讯通信（上海）有限公司 | The detection method of video source and device |
| CN106488312B (en) * | 2015-08-31 | 2019-04-30 | 展讯通信（上海）有限公司 | The detection method and device of video source |
| WO2017101348A1 (en) * | 2015-12-14 | 2017-06-22 | 乐视控股（北京）有限公司 | Method and device for deinterlacing interlaced videos |
| CN111476803A (en) * | 2020-04-14 | 2020-07-31 | 展讯通信（上海）有限公司 | Video processing method and related equipment |
| CN111476803B (en) * | 2020-04-14 | 2022-08-26 | 展讯通信（上海）有限公司 | Video processing method and related equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104580978B (en) | 2017-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8526502B2 (en) | Method and apparatus for line based vertical motion estimation and compensation | |
| US7075988B2 (en) | Apparatus and method of converting frame and/or field rate using adaptive motion compensation | |
| JP4997281B2 (en) | Method for determining estimated motion vector in image, computer program, and display device | |
| US9148622B2 (en) | Halo reduction in frame-rate-conversion using hybrid bi-directional motion vectors for occlusion/disocclusion detection | |
| US7519230B2 (en) | Background motion vector detection | |
| US20050232357A1 (en) | Motion vector estimation at image borders | |
| US20090161011A1 (en) | Frame rate conversion method based on global motion estimation | |
| US7949205B2 (en) | Image processing unit with fall-back | |
| CN101640783B (en) | De-interlacing method and de-interlacing device for interpolating pixel points | |
| JP2004312680A (en) | Motion estimation apparatus and method for detecting scrolling text or graphic data | |
| JP2005318621A (en) | Ticker process in video sequence | |
| KR20100114499A (en) | Image interpolation with halo reduction | |
| CN104580978A (en) | Video detecting and processing method and video detecting and processing device | |
| CN103369208A (en) | Self-adaptive de-interlacing method and device | |
| KR20100063655A (en) | Method for detection of film mode or camera mode | |
| CN102509311B (en) | Motion detection method and device | |
| JP2009533887A (en) | Motion vector field correction apparatus and method | |
| CN102447870A (en) | Detection method for static objects and motion compensation device | |
| US20120274845A1 (en) | Image processing device and method, and program | |
| CN111294545B (en) | Image data interpolation method and device, storage medium and terminal | |
| US10659724B2 (en) | Method and apparatus for providing dropped picture image processing | |
| JP5188272B2 (en) | Video processing apparatus and video display apparatus | |
| US10587840B1 (en) | Image processing method capable of deinterlacing the interlacing fields | |
| US8811658B2 (en) | Interpolation method for image pictures and image processing apparatus thereof | |
| US8432976B2 (en) | System and method for detecting field order of video sequences |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |