CN102572290A - Detection and processing method for 2-2 film mode - Google Patents

Abstract

The invention discloses a detection and processing method for a 2-2 film mode. The method comprises the following steps of: dividing screen areas, wherein the divided screen areas comprise areas possibly with caption and a non-caption area; performing 2-2 pull-down film mode detection on the non-caption area; if the non-caption area is not in a 2-2 pull-down mode, performing entire-screen picture de-interlacing processing; and if the non-caption area is in the 2-2 pull-down mode, searching for possibly moving caption in the areas possibly with the caption, judging whether scroll caption exists or not, performing picture de-interlacing processing on a caption area and performing film mode processing on the other areas if the scroll caption exists, and if the scroll caption does not exist, performing the film mode processing. The method has the advantages that: the film mode with the scroll caption can be simply and conveniently detected and processed, and the shortcoming that the quality of a picture is degraded by the processing of a picture de-interlacing technology due to a caption feather phenomenon or film mode quitting caused by scroll caption detection incapability is overcome.

Description

The detection of 2-2 film mode and processing method

Technical field

The present invention relates to a kind of detection and processing method and device of 2-2 film mode, especially a kind of detection and processing method that the drop-down film mode of 2-2 of roll titles is arranged.

Background technology

At present, along with popularizing of progressive-scan tv, need go interlacing to video flowing; Yet in practical application; The scan video of the normal interlaced often stream that goes interlacing to handle, the common TV video flowing is interlacing, the time point of each all is different.Produce by per second 24 frames because normally press for film video, and two of every frame all is to put shooting at the same time, removes to handle film video if use traditional deinterlacing technique, and effect is not best.For movie glibly on television set, in the time of can using the 2-2 glass tube down-drawing that movie conversion is become the pal mode video, change 24 frames into 48, in the time of also can using the 3-2 glass tube down-drawing that movie conversion is become TSC-system formula video, change 24 frames into 60.

This patent is only discussed the former.

Because when televising sometimes; TV, film video stream all are to intert the mixed video stream that forms each other; Therefore in order to recover the mute quality to the full extent; Need detect in real time mixed video stream, so that when finding TV set film signal, in time switch on the de-interlaced film mode.Corresponding processing method be with its through " weave " (interweaving) method revert to the progressive scanned video stream of 24 original frame/seconds, the effect of processing is removed the effective of interlacing processing than television mode like this.

When the movie video flowing, often have the roll titles of rolling, these captions are to add by the field editor in the later stage, after film mode is handled, diplopia can occur.As shown in Figure 1, two fields below every width of cloth figure, totally 4 width of cloth figure form one eight sequence, and the captions that added usually roll on picture lower horizontal ground, and literal all is that the position is different at each.The content of film video is a circle that rolls to the right, and per two circles roll once, the zone that has mobile captions that the horizontal line below is confirmed for us; The lower direction moving literal that rolls left is every and all changes; Be detected as the drop-down field of 2-2 clocklike if having the field sequence of roll titles, just can use " weave " method to eliminate interlacing, the phenomenon of then sprouting wings significantly will be apparent in around the crawl; Shown in Figure 1B, this is definitely unallowed; If do not detect is film mode; Possibly will from film mode, jump out, get into common TV alternate line eliminating process method, relatively good the rolling literal below horizontal line; But the video above horizontal line, resolution will be direct " weave " low.

Patent has proposed a kind of method that moves interlaced text that in video sequence, detects for 200710006202.7 li.Provide the detection of moving interlaced text to make film mode decision; Can correctly detect through telecine process and be derived from the mobile interlaced text in the video sequence of film; And solved the drop-down detection of 3-2; With the drop-down detection of 2-2, but calculate more loaded down with trivial detailsly, and it is not high correctly to detect the accuracy of scan text.

Certainly, the drop-down film mode detection method of the 2-2 that also has is devoted to drop-down sequence of 2-2 of examination criteria usually, does not consider the existence of roll titles.Like 200610138752.X, 02113843.5 the drop-down film mode detection method of a kind of 2-2 has been proposed respectively also; They detect film mode according to strange of same frame generation and the characteristics of even field pattern as if height correlation; If detect the situation of roll titles with their method; Light then the phenomenon that captions are sprouted wings can when handling, occur, heavy then jump out film mode, press the processing of TV deinterlacing technique.No matter be which kind of result, all can influence the quality of picture.

In sum; For fear of said circumstances; Improve the detection and the processing accuracy of TV-film video stream; A kind of new, the 2-2 film mode of the crawl in the FX is detected, and the invention that detects the method and apparatus that has the film mode that moves the interlacing scan captions is imperative.

Summary of the invention

The objective of the invention is to solve following technical problem:

The first, when using the 2-2 glass tube down-drawing to convert film source to the pal mode video, if belong to the drop-down film mode of the 2-2 that has roll titles, subarea processing still can detect the film mode of 2-2, and makes correct processing.

The second, the pal mode video to various use 2-2 glass tube down-drawings convert film source to can accurately detect the film mode of 2-2, and provide correct treatment.

It is a kind of with the screen separated into two parts that the present invention provides, the zone that possibly occur for captions, and one is the zone that captions impossible occur.To the zone of captions not occurring; Carry out the drop-down film mode of 2-2 and detect, based on the drop-down film mode interlace signal of 2-2 per 2 from a frame, have high correlation; Detect sequence and be rule that " size size " arranged, judge whether to belong to the 2-2 pull-down pattern with this.If non-caption area does not belong to the 2-2 pull-down pattern; Then carry out whole screen television and go interlacing to handle, as belong to, then seek possible crawl in the zone that captions possibly occur; Judged whether roll titles; If any roll titles, then caption area to be carried out TV and go interlacing to handle, the processing of film mode is carried out in other zones; As do not have roll titles, then carry out the processing of film mode.

The object of the invention is realized through following technical scheme:

A kind of detection of 2-2 film mode and processing method is characterized in that it comprises the steps:

Divide screen area, the screen area of division comprises zone and the non-caption area that captions possibly occur;

Non-caption area is carried out the drop-down film mode of 2-2 to be detected; Based on the drop-down film mode interlace signal of 2-2 per 2 from a frame; Have high correlation, detecting sequence has the rule of " size size ", judges whether to belong to the drop-down film mode of 2-2 with this;

If non-caption area does not belong to the 2-2 pull-down pattern; Then carry out whole screen television and go interlacing to handle, as belong to, then seek possible crawl in the zone that captions possibly occur; Judged whether roll titles; If any roll titles, then caption area to be carried out TV and go interlacing to handle, the processing of film mode is carried out in other zones; As do not have roll titles, then carry out the processing of film mode.

The method of seeking captions in the zone that captions possibly occur comprises:

The statistics current point with close on the field difference of the point of the individual pixel horizontal level difference of N (N=2 or 3 or 4) arranged, find forward direction same polarity field difference and the difference of back then, if difference is big to same polarity field difference;

And continuous three have these points that move of N pixel in same horizontal movement direction, calculate the quantity of these points, if greater than certain threshold value, and continuous several so, then thinking has the captions of rolling; Otherwise then thinking does not have.

The method that obtains the drop-down film mode detection of 2-2 sequence comprises:

Two the correspondence position pixel that polarity is identical is subtracted each other and is obtained two differences of same polarity, and two poor absolute values of the absolute value of two differences of its same polarity and the same polarity of previous field remake difference and obtain four differences of same polarity;

Adjacent two the correspondence position pixel that polarity is different is subtracted each other and is obtained two differences of heteropolarity; Four differences of this same polarity and two difference of heteropolarity are not obtained film mode 2-2 detected value with the back addition of not multiplying each other of two weights;

With film mode 2-2 detected value and previous film mode 2-2 detected value relatively, if greater than, then the value with the detection sequence of this correspondence is taken as 1, thinks big; If less than, then with the value of the detection sequence of this correspondence be taken as 0 think little;

If the conformance with standard sequence then meets film mode.

It is jumped into the mechanism of jumping out and comprises:

The sequential value that detects sequence set sequential value and standard detection sequence set is compared successively, when the number of times that meets each other continuously reaches the incision threshold value, be judged to be 2-2 conversion television movie mode, the interlacing of going that switches to the 2-2 conversion is handled;

The sequential value that detects sequence set and the sequential value of standard detection sequence set are compared successively, when the number of times that is not inconsistent continuously reaches when cutting out threshold value, jump out the interlaced mode of going that 2-2 changes.

Also comprise error correction step: when the motion of the picture in the film mode is very little or static to the smaller Sequence Detection of picture motion; Detect sequence and do not meet standard sequence; If pairing is not 1 or 0; And four differences of same polarity are less, then think and can use the standard detection sequence set to replace, and do not jump out original tupe.

Film mode 2-2 processing module treatment step comprises:

Judge whether captions sign Txt22_Flag is 1, if not, then carry out the processing of 2-2 film mode, if promptly when the corresponding sequence set element in front court be 0, should with last one merge, if be 1, should and next field merge;

If captions sign Txt22_Flag is 1, then when handling, need handle area dividing, then caption area to be carried out TV and go interlacing to handle, the processing of film mode is as above carried out in other zones.

The relative prior art of the present invention has the following advantages and beneficial effect: can detect and handle the film mode that roll titles is arranged quickly and easily; Prevented that because of detecting the captions emergence phenomenon that the roll titles that do not come out occurs or jumping out film mode what cause handles and drawback that the image quality that causes descends by the TV deinterlacing technique.

Description of drawings:

Through following description, can further understand purpose, specific structural features and the advantage of its invention to its accompanying drawing of embodiments of the invention combination.Wherein, accompanying drawing is:

Figure 1A-D is: the 2-2 film mode of band roll titles is through " weave " the sketch map that occurs emergence after handling;

Fig. 2 is: the rule sketch map of four differences of same polarity during the 2-2 film mode;

Fig. 3 is: the rule sketch map of two differences of heteropolarity during the 2-2 film mode;

Fig. 4 is: captions region sketch map;

Fig. 5 is: the 2-2 film mode of band roll titles detects flow chart of steps;

Fig. 6 is: the 2-2 film mode treatment step flow chart of band roll titles.

Embodiment:

Instance below in conjunction with among the figure is done further description to the present invention.

Film mode of the present invention only detects luminance channel is detected, and other two chrominance channels will be handled according to the processing of luminance channel.

2-2 detects the method for movies signal, comprises the steps:

The heteropolarity field of the interleaved signal of A, four differences in same polarity field of calculating n-3, n-2, n-1 field and the interleaved signal of n field and n-1 field and n field is poor, and according to income value through calculating the detection sequence that 2-2 changes;

B, judge that gained detects sequence and whether meets corresponding standard detection sequence, if then be the movies signal of this standard detection sequence correspondence, otherwise be TV signal.

The principle that detects 2-2 conversion TV set film signal is as shown in Figure 2; Can find out; Because the absolute of first, the 3rd same polarity field difference and second, the 4th same polarity field difference is from two frame data; And the absolute of second, the 4th same polarity field difference and the 3rd, the 5th same polarity field difference is from three frame data, because the correlation of two frame data will be higher than three frame data, the former is littler than the latter in theory.Therefore the obvious characteristic that " size size " so alternately occurs appears in four missionary societies of same polarity.

As can be seen from Figure 3; Because from the correlation of two field data of same frame than higher; So its heteropolarity field difference data in theory should be smaller, and lower from two field data correlations of different frame, its heteropolarity field difference in theory should be bigger than the former.Therefore detect heteropolarity field difference and the obvious characteristic that " size size " so alternately occurs also can occur.With above the two combine, satisfy such characteristic if detect sequence, then decidable is a 2-2 conversion TV set film signal.

Specific practice is: two correspondence position pixels that polarity is identical subtract each other that to obtain the same polarity field poor, and the absolute value of its same polarity field difference and the same polarity field of previous field differ from absolute value and remake difference and obtain four differences of same polarity; Adjacent two the correspondence position pixel that polarity is different subtracts each other that to obtain the heteropolarity field poor; Four differences of this same polarity and heteropolarity field difference are obtained a new film mode 2-2 detected value through weighting summation, with this value and previous film mode 2-2 detected value relatively, if greater than, then the value with the detection sequence of this correspondence is taken as 1 (thinking big); If less than, then the value with the detection sequence of this correspondence is taken as 0 (thinking little).The detection sequence and the standard sequence that will obtain are then done the detection of relatively carrying out the drop-down film mode of 2-2.

When film mode was handled, the field that " 0 " is corresponding merged with the field of front, and the field that " 1 " is corresponding merges with the field of back.

Roll titles or be called crawl (down with) can below, but be not limited to the below, also can be up or side.The present invention establishes surveyed area to rectangular area, screen below, this means, will only carry out at this white space above rectangular area of screen the detection of 2-2 film mode.Be illustrated in figure 4 as caption area below sketch map.

Shaded rectangle zone A is judged whether the appearance of roll titles; Need find the characteristics of the roll titles that moves horizontally: the error ratio between the same polarity field is big (this judgement can mask static background); And the heteropolarity field has the characteristics of horizontal motion vector in the horizontal direction; Find the quantity of the point of roll titles, if greater than certain threshold value, then thinking has the captions of rolling.If any roll titles, the whole application television Signal Processing in this shadow region method then; As do not have, Zone Full is used the film mode mode and is handled.

Below will describe in detail step by step:

Step 1, statistics

Read in four field data, current data is the 3rd, and it is poor to calculate four differences of same polarity and heteropolarity field then.

Step 1A, four differences of statistics same polarity

Step 1A1: calculate same polarity field difference counter Same_Counter

At first statistics when the front court with next poor with the corresponding picture element of previous same polarity field respectively:

Formula 1

pixel_err_f(i，j)＝|f _n(i，j)-f _n-2(i，j)|

pixel_err_b(i，j)＝|f _n+1(i，j)-f _n-1(i，j)|

Wherein, (it is poor to be defined as the same polarity former two for i, j) poor for when the corresponding picture element of front court and previous same polarity field for pixel_err_f.(i, what be next with the corresponding picture element of its previous same polarity field j) is poor, is defined as two differences after the same polarity for pixel_err_b.。

Be provided with n same polarity field difference counter Same_Counter (1), Same_Counter (2) ... Same_Counter (n) initial value is made as 0; Be respectively applied for to pixel_err_f (i, j) and pixel_err_b (i j) is in [0; 256/n-1], [256/n; 512/n-1], [512/n, 1024/n-1] ... the number of [256-256/n, 255] scope is counted.N generally chooses 8 or 16.

Here Th [k]=256*k/n, Th [0]=0, expression threshold value.Here, txt_th is for drawing the distance on middle distance on the scene base, piece zone.Select 60.This formula shows, the picture element difference of other the regional same polarity fields except that caption area is carried out the number of each tonal range and adds up, and is stored in n the counter.

Step 1A2: the Same_Counter according to the front formula calculates, calculate four the poor field4_error of same polarity that are used for the detection of 2-2 film mode:

Formula 3 $\mathrm{Field}4\_\mathrm{Error}=\underset{k=1}{\overset{n}{\mathrm{\Σ}}}|\mathrm{Same}\_\mathrm{Counter}\left(k\right)-\mathrm{Last}\_\mathrm{Same}\_\mathrm{Counter}\left(k\right)|$

Here, field4_error is four differences, and Same_Counter is this same polarity field of trying to achieve difference counter, and last_Same_Counter is a last same polarity field difference counter of trying to achieve.

Step 1B, adjacent of statistics heteropolarity are poor

Step 1B1, at first ask corresponding picture element poor of previous field and back one heteropolarity field:

Formula 4

pixel_diff_f(i，j)＝|f _n(i，j)-f _n-1(i，j)|

pixel_diff_b(i，j)＝|f _n(i，j)-f _n+1(i，j)|

Wherein, pixel_diff_f (i, j) and pixel_diff_b (i j) is poor with the corresponding picture element of previous field and back one heteropolarity field.

Step 1B2, ask heteropolarity field difference Sum_diff_res

It is 0 that heteropolarity field difference counter diff_Counter initial value is set, and is respectively applied for that (i j) is in [0,256/n-1], [256/n, 512/n-1], [512/n, 1024/n-1] ... the number of [256-256/n, 255] scope is counted to pixel_diff_f.Obtain being used for the diff_Counter that the 2-2 film mode detects:

This formula shows, the picture element difference of other the regional heteropolarity fields except that caption area is carried out the number of each tonal range and adds up, and is stored among n the counter diff_Counter.

Formula 6

$\mathrm{Sum}\_\mathrm{diff}\_\mathrm{res}=\underset{k=2}{\overset{n}{\mathrm{\Σ}}}\mathrm{diff}\_\mathrm{Counter}\left(k\right)$

This formula shows, according to the heteropolarity field difference Sum_diff_res that the value of 2 to n counter diff_Counter is sued for peace obtain.

Step 1C, the statistics that moves horizontally captions point and the judgement of caption area

Step 1C1, statistics move horizontally the captions point

Below the zone; Characteristics according to the roll titles that moves horizontally: promptly two poor pixel_err_b are bigger after same polarity former two difference pixel_err_f and the same polarity; And continuous three have moving of N pixel in the same direction of motion; Find the quantity of the point of these roll titles, if greater than certain threshold value, then thinking has the captions of rolling.

At first add up current point and close on a difference (N=2 or 3 or 4, getting 2 with N here is example) that the point of N pixel horizontal level difference is arranged:

pixel_diff_fl2(i，j)＝|f _n(i，j)-f _n-1(i，j+2)|

pixel_diff_bl2(i，j)＝|f _n(i，j)-f _n+1(i，j-2)|

Formula 7

pixel_diff_fr2(i，j)＝|f _n(i，j)-f _n-1(i，j-2)|

pixel_diff_br2(i，j)＝|f _n(i，j)-f _n+1(i，j+2)|

Wherein, pixel_diff_fl2 (i, j) poor for the corresponding picture element of two picture elements on the right of the level of front court and previous field.Pixel_diff_bl2 (i, j) for when the front court with after one corresponding picture element poor of two picture elements in the horizontal left side.Pixel_diff_fr2 (i, j) poor for when the corresponding picture element of front court and two picture elements in the horizontal left side of previous field.Pixel_diff_br2 (i, j) for when the front court with after corresponding picture element poor of two picture elements on the right of one the level.

Here, txt_th is for drawing the distance on middle distance on the scene base, piece zone.Select 60.Th_txt_b is a threshold value of thinking that crawl is different with background, selects 50 usually, and Th_txt_s thinks the threshold value that crawl is similar.Usually select 32.Here Num_txt is a number of picture elements of thinking to move horizontally captions.

Step 1C2: the captions situation that judges whether to belong to 2-2

According to the Num_txt that obtains,, then think to belong to subtitling mode if all be like this greater than certain threshold value and continuous several:

Here Th_txt_num is the number of picture elements threshold value that moves horizontally captions, chooses 40 usually.Txt_flag_Num is Num_txt greater than the counting of the continous field of certain threshold value, if during middle Txt_flag=0, interrupts counting, restarts.This formula shows, all is Num_txt during greater than threshold value as 3 fields, just thinks the crawl pattern that got into, and this can effectively evade the pixel (but not being crawl) that some were motion originally.

The detection of the television movie mode of step 2,2-2 conversion

Step 2A: calculate 2-2 transition detection sequence seq22 according to four differences of same polarity and heteropolarity field difference

Step 2A1: calculate film mode 2-2 detected value FilmMode22:

Formula 10 FilmMode22=th_f4*field4_error+th_diff*Sum_diff_res

Th_f4 is a coefficient, is made as 10.For the influence that makes field4_error and Sum_diff_res on an one magnitude.Th_diff is another coefficient, is made as 1.The weighting purpose is the ratio weight of adjustment field4_error and Sum_diff_res.

Step 2A2: ask the element element_22 that detects sequence set

Gained film mode 2-2 detected value FilmMode22 and previous film mode 2-2 detected value last_FilmMode22 are compared; If less than previous last_FilmMode22; Explain that current FilmMode22 is little; Element_22 is changed to 0 with the current sequence value, otherwise element_22 is changed to 1; Element_22 is for detecting the element of sequence set seq22.

Step 2A3: continuous 10 sequential values are arranged in order, generate 2-2 transition detection sequence set seq22.

Step 2B: relatively gained detects sequence set seq22 and standard detection sequence

If the prediction initial value is 0, predict22=0; Relatively the non-caption area of gained detects sequence set seq22 and standard detection sequence set seq22_s, when the sequential value in detecting sequence set is identical with the first sequential value in the standard detection sequence set, gets into detected state, establishes predict22=1;

Wherein seq22_s is 0101010101 or 1010101010.

Step 2C: correct the little detection sequence of motion picture

When the picture in film mode motion was very little, at this moment the result that detects of field4_error was smaller, and what occur that the situation of size maybe be with prediction is different, and jumping out of blindness there is no need like this.Need screen this situation.

False code is following:

Here, Th_noise thinks the threshold value of discredited field4_error, chooses 10 usually, and bitxor representes the step-by-step XOR.This formula shows; Its seq22 and the pattern seq22_s of prediction are pressed bit comparison; Although if certain pairing be not desirable big or little; If but four poor field4_error of same polarity are smaller, can think that this situation deficiency wins the confidence, still replace with the standard detection sequence set.Through above-mentioned processing, can improve so that the very little picture of some motions is compelled to jump out film mode.

Step 2D: jump into film mode and jump out film mode

When predict22=1; When the sequential value of sequential value that detects sequence set seq22 and standard detection sequence set seq22_s is compared successively; When the number of times film_22counter that meets each other continuously reaches incision threshold value Th_film_in; Be judged to be 2-2 conversion television movie mode, the interlacing of going that switches to the 2-2 conversion is handled;

Sequential value that detects sequence set seq22 and the sequential value of standard detection sequence set seq22_s are compared successively, when the times N ofilm_22counter that is not inconsistent continuously reaches when cutting out threshold value Th_film_out, jump out the interlaced mode of going that 2-2 changes.

Wherein Film_final_result is a 2-2 translative mode sign.If be that 1 expression is the 2-2 film mode, if 0, then expression is not.Th_film_out is a film mode incision threshold value, is made as 15, and Th_film_in is that film mode cuts out threshold value, is made as 5; Film_counter_22 is the counter that meets the field of 2-2 film mode continuously, and Nofilm_counter_22 is the counter that does not meet the field of 2-2 film mode continuously.It is as shown in Figure 5 more than to detect flow chart of steps.

Step 3, film mode Processing Algorithm realize

After coming out from the film mode detection module, be 2-2 signal (being Film_final_result=1) if detect, interlace signal gets into film mode 2-2 processing module.Carry out following steps:

Step 3A: judge Txt22_Flag=1? If, then when handling, need handle area dividing, the interlacing algorithm is removed in the latter half application television, and the first half is handled as follows:

As the corresponding sequence set element element_22 in front court,, should merge with last one the field if be 0.If be 1, should merge with next the field.

Step 3B: like Txt22_Flag=0, whole screen area is handled as follows:

As the corresponding sequence set element element_22 in front court,, should merge with last one the field if be 0.If be 1, should merge with next the field.The film mode handling process is as shown in Figure 6.

Remarkable result of the present invention is; To 2-2 pattern film source in tv edit because of adding roll titles; Detect not come out and film mode occurs; Thereby cause that screen resolution is not high, or to detect be film mode, but make captions shortcomings such as emergence phenomenon occur, disclosing a kind of consideration has the drop-down film mode of 2-2 of roll titles to detect and processing method.The present invention detects the method for film mode, and there is roll titles in the zone in that the captions of screen possibly occur like judgement, then TV is carried out in this zone and goes interlacing to handle, and the processing of film mode is then still carried out in other zones.After handling like this, resolution that both can the most of screen of undistorted reduction can make mobile captions sawtooth or emergence not occur again.

In sum: the present invention provides a kind of with the screen separated into two parts, the zone that possibly occur for captions, and one is the zone that captions impossible occur.To the zone of captions not occurring; Carry out the drop-down film mode of 2-2 and detect, based on the drop-down film mode interlace signal of 2-2 per 2 from a frame, have high correlation; Detect sequence and be rule that " size size " arranged, judge whether to belong to the 2-2 pull-down pattern with this.If non-caption area does not belong to the 2-2 pull-down pattern, then carry out whole screen television and go interlacing to handle, as belong to; Then whether roll titles is arranged in the region decision that possibly occur; If any roll titles, then caption area to be carried out TV and go interlacing to handle, the processing of film mode is carried out in other zones; As do not have roll titles, then carry out the processing of film mode.So; Can detect and handle the film mode that roll titles is arranged quickly and easily; Prevented that because of detecting the captions emergence phenomenon that the roll titles that do not come out occurs or jumping out film mode what cause handles and drawback that the image quality that causes descends by the TV deinterlacing technique.

Those of ordinary skill in the art will be appreciated that; Above embodiment is used for explaining the object of the invention; And be not with opposing qualification of the present invention; As long as in essential scope of the present invention, all will drop in the scope of claim of the present invention variation, the modification of the above embodiment.

Claims (8)

1. the detection of a 2-2 film mode and processing method is characterized in that, it comprises the steps:

Divide screen area, the screen area of division comprises zone and the non-caption area that captions possibly occur;

Non-caption area is carried out the drop-down film mode of 2-2 detect,, then carry out whole screen television and go interlacing to handle if non-caption area does not belong to the drop-down film mode of 2-2; As belong to; Then seek captions, judged whether roll titles, if any roll titles in the zone that captions possibly occur; Then caption area is carried out TV and go interlacing to handle, the processing of film mode is carried out in other zones; As do not have roll titles, then carry out the processing of film mode.

2. the detection of 2-2 film mode according to claim 1 and processing method; It is characterized in that: saidly non-caption area is carried out the drop-down film mode of 2-2 detect; Be based on the drop-down film mode interlace signal of 2-2 per 2 from a frame; Have high correlation, detecting sequence has the rule of " size size ", judges whether to belong to the drop-down film mode of 2-2 with this.

3. the detection of 2-2 film mode according to claim 1 and processing method is characterized in that: the method for seeking captions in the zone that captions possibly occur comprises:

The statistics current point with close on the field difference of the point of the individual pixel horizontal level difference of N (N=2,3 or 4) arranged, find forward direction same polarity field difference and the difference of back then, if difference is big to same polarity field difference;

And continuous three have these points that move of N pixel in same horizontal movement direction, calculate the quantity of these points, if greater than certain threshold value, and continuous several so, then thinking has the captions of rolling; Otherwise then thinking does not have.

4. the detection of 2-2 film mode according to claim 1 and processing method is characterized in that, the method that obtains the drop-down film mode detection of 2-2 sequence comprises:

Two the correspondence position pixel that polarity is identical is subtracted each other and is obtained two differences of same polarity, and two poor absolute values of the absolute value of two differences of its same polarity and the same polarity of previous field remake difference and obtain four differences of same polarity;

Adjacent two the correspondence position pixel that polarity is different is subtracted each other and is obtained two differences of heteropolarity; Four differences of this same polarity and two difference of heteropolarity are not obtained film mode 2-2 detected value with the back addition of not multiplying each other of two weights;

With film mode 2-2 detected value and previous film mode 2-2 detected value relatively, if greater than, then the value with the detection sequence of this correspondence is taken as 1, thinks big; If less than, then with the value of the detection sequence of this correspondence be taken as 0 think little;

If the conformance with standard sequence then meets film mode.

5. the detection of 2-2 film mode according to claim 4 and processing method is characterized in that, 2-2 detects the method for movies signal, comprises the steps:

The heteropolarity field of the interleaved signal of A, four differences in same polarity field of calculating n-3, n-2, n-1 field and the interleaved signal of n field and n-1 field and n field is poor, and according to income value through calculating the detection sequence that 2-2 changes, n representes to work as the front court;

B, judge that gained detects sequence and whether meets corresponding standard detection sequence, if then be the movies signal of this standard detection sequence correspondence, otherwise be TV signal.

6. the detection of 2-2 film mode according to claim 1 and processing method is characterized in that, it is jumped into the mechanism of jumping out and comprises:

The sequential value that detects sequence set sequential value and standard detection sequence set is compared successively, when the number of times that meets each other continuously reaches the incision threshold value, be judged to be 2-2 conversion television movie mode, the interlacing of going that switches to the 2-2 conversion is handled;

The sequential value that detects sequence set and the sequential value of standard detection sequence set are compared successively, when the number of times that is not inconsistent continuously reaches when cutting out threshold value, jump out the interlaced mode of going that 2-2 changes.

7. the detection of 2-2 film mode according to claim 6 and processing method; It is characterized in that, also comprise error correction step: when the motion of the picture in the film mode is very little, detects sequence and do not meet standard sequence to the smaller Sequence Detection of picture motion; If pairing is not 1 or 0; And four differences of same polarity are less, then think and can use the standard detection sequence set to replace, and do not jump out original tupe.

8. the detection of 2-2 film mode according to claim 1 and processing method is characterized in that, film mode 2-2 processing module treatment step comprises:

Judge whether captions sign Txt22_Flag is 1, if not, then carry out the processing of 2-2 film mode, if promptly when the corresponding sequence set element in front court be 0, should with last one merge, if be 1, should and next field merge;

If captions sign Txt22_Flag is 1, then when handling, need handle area dividing, then caption area to be carried out TV and go interlacing to handle, the processing of film mode is as above carried out in other zones.

Images