CN101547372A - Method and device for achieving three-dimensional video - Google Patents

Abstract

The invention discloses a method and a device for achieving three-dimensional video, which improve the problem of numerous data of the three-dimensional video, do not increase the bandwidth, reduce the algorithm complexity, have simple video processing, and reduce the cost. The technical proposal comprises that the method comprises the following steps: (1) segmenting left eye images and right eye images into a plurality of pixel points respectively; and (2) inserting and assembling the pixel points corresponding to the left eye images and the right eye images into a set of new video images in an alternation mode. The method and the device are applied in the field of the video processing.

Description

A kind of three-dimensional video-frequency implementation method and device

Technical field

The present invention relates to a kind of video implementation method and device, relate in particular to a kind of video implementation method and device of stereo format.

Background technology

Along with the continuous development of TV tech, the television video technology is changed to three-dimensional video-frequency by planar video gradually.Compare with the common planar video, three-dimensional video-frequency has increased the depth information of object, the three-dimensional information of scene is fully reproduced out, the beholder is distance, the depth of scenery in the image as can be seen, the presence and the sense true to nature of vision have been strengthened, highlight the image and content on have stronger impact and impulsive force undoubtedly, image ready to appear makes the beholder obtain sensation on the spot in person.

Human eye can not only see the width and the height of object, and can know their degree of depth when watching object, can judgment object between or the distance between beholder and object.The main cause that this 3D vision characteristic produces is: people usually always binocular watch object simultaneously, and because the spacing of two optical axis, right and left eyes is when seeing same object, received visual pattern is different, thereby brain is by the motion adjustment of eyeball, combine the information of this two width of cloth image, produced third dimension.

As shown in Figure 1, human eye can not only see the width and the height of object, and can know their degree of depth when watching object, can judgment object between or the distance between beholder and object.The main cause that this 3D vision characteristic produces be people usually always binocular watch object simultaneously, and because the spacing (about 65mm) of two optical axis, right and left eyes is when seeing same object, received visual pattern is different, thereby brain is by the motion adjustment of eyeball, combine the information of this two width of cloth image, produced third dimension.When list was watched object with left eye and right eye, the image shift that is produced felt to be called parallax.Shown in accompanying drawing one, with two identical video cameras, two planes of delineation are positioned at same plane Q.Two machine reference axis are parallel, and X-axis overlaps, and two machines are baseline B in the directions X spacing.The subpoint G of characteristic point P in two planes of delineation in the scene _L, G _RBe called conjugate pair, promptly a subpoint is another corresponding points, just in the cataloged procedure that reconfigures left-eye image and eye image about a pair of corresponding points of two width of cloth image previous generation table space same points.After two width of cloth figure are overlapping, the alternate position spike X between the conjugate pair _LOne X _RBe parallax, release z=bf/ (X according to triangular relationship _L-X _R).This shows that the depth information of object calculates by parallax, can recover the depth information of object by the parallax of computer memory o'clock in two width of cloth images.

The stereoscopic vision characteristic of human eye is the basis of stereoscopic TV.We can propose the solution thinking of stereoscopic TV by this principle.Utilize two interstations video camera at a certain distance to take stereo-picture synchronously from two suitable angles, note two parts information, after televising, by television receiver two parts information is sent into left eye and right eye respectively, guarantee to allow people's right and left eyes see corresponding image respectively, in brain, after synthetic, be shown as stereo-picture, thereby produced third dimension.

As shown in Figure 2, the stereoscopic TV video technique that generally uses adopts and transmits left-eye image and eye image by turns with certain speed at present, the video picture end shows left-eye image and eye image in turn on screen, the beholder need wear a special liquid crystal glasses and watch three-dimensional television image, glasses with one with the control of the synchronous switch of transmitting terminal, when left-eye image occurs, the liquid crystal lens printing opacity of left eye, the liquid crystal lens of right eye is light tight; On the contrary, when eye image occurs, have only the liquid crystal lens printing opacity of right eye, the liquid crystal lens of left eye is light tight, thus about making two can only see required separately image.

This method can cause data volume to increase, and bandwidth increases, and makes original 60HZ image become the 120HZ image, and the increase of bandwidth can improve demand and the cost to hardware widely.But this method can produce comparatively serious flicker.If eliminate flicker, just must improve frame per second, thereby increase bandwidth, but this can improve demand and cost to hardware undoubtedly widely.Adopt this method can cause data volume huge in addition, and sequencing control complexity, adopt the stereoscopic TV of traditional three-dimensional video-frequency form, the active liquid crystal glasses of the costliness that need wear when watching is the Photoelectric Detection formula, except that needs input video drive signal, also needs to import steering signal, the circuit structure complexity, the debugging trouble, cost is higher, and causes visual fatigue easily.

Summary of the invention

The objective of the invention is to address the above problem, a kind of three-dimensional video-frequency implementation method is provided, improve the huge problem of stereo video data amount, can not increase bandwidth, reduced the complexity of algorithm, and Video processing is simple, has reduced cost.

Another object of the present invention has provided a kind of three-dimensional video-frequency implement device.

Technical scheme of the present invention is: the present invention has disclosed a kind of three-dimensional video-frequency implementation method, comprising:

(1) respectively left-eye image and eye image are divided into a plurality of pixels;

(2) pixel of the correspondence of left-eye image and eye image is inserted in the mode of alternate intervals be combined into one group of new video image.

Above-mentioned three-dimensional video-frequency implementation method, wherein, in step (1), left-eye image is by pixel L _{M, n}Constitute, eye image is by pixel R _{M, n}Constitute;

In step (2), the pixel P of new video image _{M, n}=L _{M, n}Or R _{M, n}

Above-mentioned three-dimensional video-frequency implementation method wherein, also comprises afterwards in step (2):

Video image is carried out compressed encoding.

Above-mentioned three-dimensional video-frequency implementation method wherein, also comprises afterwards in step (2):

When playing, new video image is split as left-eye image and eye image with the anti-process of step (2).

Above-mentioned three-dimensional video-frequency implementation method wherein, also comprises:

The process that when playing, decompresses earlier.

The present invention has also disclosed a kind of three-dimensional video-frequency implement device, comprising:

The image synthesis unit is divided into a plurality of pixels with left-eye image and eye image respectively, the pixel of the correspondence of left-eye image and eye image is inserted in the mode of alternate intervals be combined into one group of new video image.

Above-mentioned three-dimensional video-frequency implement device, wherein, this device also comprises:

The image encoding unit connects this image synthesis unit, and the video image that this image synthesis unit is exported carries out compressed encoding.

Above-mentioned three-dimensional video-frequency implement device, wherein, this device also comprises:

Memory connects this image encoding unit, the video image after the store compressed.

Above-mentioned three-dimensional video-frequency implement device, wherein, this device also comprises:

The picture decoding unit connects this memory, and the video image after the compression in this memory is carried out decompression.

Above-mentioned three-dimensional video-frequency implement device, wherein, this device also comprises:

The separation of images unit connects this picture decoding unit, and the video image after will decompressing with the anti-process in this image synthesis unit is divided into left-eye image and eye image.

The present invention contrasts prior art following beneficial effect: three-dimensional video-frequency implementation method of the present invention is separated into pixel one by one with left-eye image and eye image, then the corresponding pixel points of left-eye image and eye image being encoded to calculate to reconfigure according to the algorithm process of characteristics obtains one group of new video image information, and a frame one frame ground is play again.The contrast prior art, method and apparatus of the present invention has solved the problem that data volume is huge, bandwidth is too big that the conventional stereo TV tech causes, and has reduced the cost of stereoscopic TV.And solved the problem of conventional stereo digital TV picture flicker.The pixel of the relevance that the present invention proposes is handled and is made Video processing oversimplify, and convenient and other video formats are changed.In addition, technical scheme of the present invention makes video data have good confidentiality, has possessed anti-piracy function.The invention enables relevant supporting use passive glasses to watch, reduce cost, and the eyes indefatigability.

Description of drawings

Fig. 1 is the principle schematic of stereoscopic vision.

Fig. 2 is the schematic diagram of traditional three-dimensional video-frequency implementation method.

Fig. 3 is the schematic diagram of three-dimensional video-frequency implementation method of the present invention.

Fig. 4 is the schematic diagram of three-dimensional video-frequency implementation method of the present invention.

Fig. 5 is the schematic diagram of interval of the present invention interpolation encryption algorithm.

Fig. 6 is the schematic diagram of three-dimensional video-frequency implement device of the present invention.

Embodiment

The invention will be further described below in conjunction with drawings and Examples.

The embodiment of three-dimensional video-frequency implementation method

Fig. 3 shows the principle of three-dimensional video-frequency implementation method of the present invention.The embodiment of three-dimensional video-frequency implementation method of the present invention is divided into a plurality of pixels with left-eye image and eye image respectively earlier.In Fig. 3, left-eye image 10 is divided into 1920 * 1080 pixel L _{M, n}(wherein m is the arbitrary value in 1～1080, and n is the arbitrary value in 1～1920), eye image 11 is divided into 1920 * 1080 pixel R _{M, n}Then the corresponding pixel points of left-eye image and eye image is calculated to reconfigure through coding according to special algorithm and obtain one group of new video image information and carry out transfer of data and broadcast.For example, in Fig. 3, the pixel P11 ' of the video image that reconfigures is that the pixel L11 of left-eye image and the pixel R11 of eye image calculate through special algorithm.The pixel P12 ' of the video image that reconfigures is that the pixel L12 of left-eye image and the pixel R12 of eye image calculate through special algorithm.And the like, the pixel Pmn ' of the video image that reconfigures is that the pixel Lmn of left-eye image and the pixel Rmn of eye image calculate through special algorithm,

$\mathrm{<figure-callout\; id="11"\; label="P"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">P</figure-callout>}11\&prime;=\mathrm{<figure-callout\; id="11"\; label="L"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">L</figure-callout>}11\&CircleTimes;\mathrm{<figure-callout\; id="11"\; label="R"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">R</figure-callout>}11,$

$\mathrm{<figure-callout\; id="12"\; label="P"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">P</figure-callout>}12\&prime;=\mathrm{<figure-callout\; id="12"\; label="L"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">L</figure-callout>}12\&CircleTimes;\mathrm{<figure-callout\; id="12"\; label="R"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">R</figure-callout>}12,$

$P13\&prime;=L13\&CircleTimes;R13,$

$P14\&prime;=L14\&CircleTimes;R14,$

...

$\mathrm{<figure-callout\; id="21"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}21\&prime;=\mathrm{<figure-callout\; id="21"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}21\&CircleTimes;\mathrm{<figure-callout\; id="21"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}21,$

$\mathrm{<figure-callout\; id="22"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}22\&prime;=\mathrm{<figure-callout\; id="22"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}22\&CircleTimes;\mathrm{<figure-callout\; id="22"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}22,$

$\mathrm{<figure-callout\; id="23"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}23\&prime;=\mathrm{<figure-callout\; id="23"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}23\&CircleTimes;\mathrm{<figure-callout\; id="23"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}23,$

$\mathrm{<figure-callout\; id="24"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}24\&prime;=\mathrm{<figure-callout\; id="24"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}24\&CircleTimes;\mathrm{<figure-callout\; id="24"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}24,$

...

$\mathrm{Pmn}\&prime;=\mathrm{Lmn}\&CircleTimes;\mathrm{Rmn},$

Adopt above-mentioned method, because adjacent pixels point all has relevance, so the correlation height of data processing can reduce the complexity of algorithm.And, therefore can change with other form easily because each adjacent pixels point all is to have relevance.

For above-mentioned described specific algorithm, be interval interpolation encryption algorithm in the present embodiment, can represent with following logical expression:

P11′＝L11*1+R11*0，P12′＝L12*0+R12*1，P13′＝L13*1+R13*0…

P21′＝L21*0+R21*1，P22′＝L22*1+R22*0，P23′＝L23*0+R23*1…

P31′＝L31*1+R31*0，P32′＝L32*0+R32*1，P33′＝L33*1+R33*0…

P41′＝L41*0+R41*1，P42′＝L42*1+R42*0，P43′＝L43*0+R43*1…

The essence of this interval interpolation encryption algorithm is that the pixel with the pixel of left-eye image and eye image inserts in the mode of alternate intervals and is combined into one group of new video image, as shown in Figure 5.In same pixel column or same pixel column, any adjacent pixels point can all not belong to left-eye image or belong to eye image.Interval interpolation encryption algorithm in the present embodiment, the pixel arrangement mode of the new video image that it is reassembled into is similar to chess chessboard pattern, therefore also with this interval interpolation encryption algorithm popular be called chess gridiron pattern algorithm.

New video image according to above-mentioned interval interpolation encryption algorithm obtains can not broadcast a frame left-eye image and a frame eye image in turn as traditional three-dimensional video-frequency form is realized, but the broadcast of a frame one frame in regular turn, as shown in Figure 4.Therefore, if in data transmission procedure, intercept one section video flowing, do not pass through the corresponding decoding algorithm of this programme, the video image that obtains so is can not watching of bluring, thereby make three-dimensional video-frequency form of the present invention have good confidentiality, possessed anti-piracy function.

Watch the stereoscopic TV of above-mentioned three-dimensional video-frequency form, only need wear dog-cheap passive type polarising glass, simple, convenient making of circuit structure and cost are lower.

The embodiment of three-dimensional video-frequency implement device

Corresponding to above-mentioned three-dimensional video-frequency implementation method, the present invention also provides a kind of three-dimensional video-frequency implement device, as shown in Figure 6.The three-dimensional video-frequency implement device comprises image synthesis unit 20, image encoding unit 21, memory 22, picture decoding unit 23, separation of images unit 24.Wherein image synthesis unit 20 left-eye image of respectively left video camera being taken and the right video camera eye image of taking is divided into a plurality of pixels, the pixel of left-eye image and eye image correspondence is inserted in the mode of alternate intervals be combined into one group of new video image.

At present embodiment, see also Fig. 3, left-eye image 10 is divided into 1920 * 1080 pixel L _{M, n}(wherein m is the arbitrary value in 1～1080, and n is the arbitrary value in 1～1920), eye image 11 is divided into 1920 * 1080 pixel R _{M, n}Then the corresponding pixel points of left-eye image and eye image is calculated to reconfigure through coding according to special algorithm and obtain one group of new video image information and carry out transfer of data and broadcast.For example, in Fig. 3, the pixel P11 ' of the video image that reconfigures is that the pixel L11 of left-eye image and the pixel R11 of eye image calculate through special algorithm.The pixel P12 ' of the video image that reconfigures is that the pixel L12 of left-eye image and the pixel R12 of eye image calculate through special algorithm.And the like, the pixel Pmn ' of the video image that reconfigures is that the pixel Lmn of left-eye image and the pixel Rmn of eye image calculate through special algorithm,

$\mathrm{<figure-callout\; id="11"\; label="P"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">P</figure-callout>}11\&prime;=\mathrm{<figure-callout\; id="11"\; label="L"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">L</figure-callout>}11\&CircleTimes;\mathrm{<figure-callout\; id="11"\; label="R"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">R</figure-callout>}11,$

$\mathrm{<figure-callout\; id="12"\; label="P"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">P</figure-callout>}12\&prime;=\mathrm{<figure-callout\; id="12"\; label="L"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">L</figure-callout>}12\&CircleTimes;\mathrm{<figure-callout\; id="12"\; label="R"\; filenames="A200910050475E00131.png,A200910050475E00141.png"\; state="\{\{state\}\}">R</figure-callout>}12,$

$P13\&prime;=L13\&CircleTimes;R13,$

$P14\&prime;=L14\&CircleTimes;R14,$

...

$\mathrm{<figure-callout\; id="21"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}21\&prime;=\mathrm{<figure-callout\; id="21"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}21\&CircleTimes;\mathrm{<figure-callout\; id="21"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}21,$

$\mathrm{<figure-callout\; id="22"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}22\&prime;=\mathrm{<figure-callout\; id="22"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}22\&CircleTimes;\mathrm{<figure-callout\; id="22"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}22,$

$\mathrm{<figure-callout\; id="23"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}23\&prime;=\mathrm{<figure-callout\; id="23"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}23\&CircleTimes;\mathrm{<figure-callout\; id="23"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}23,$

$\mathrm{<figure-callout\; id="24"\; label="P"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">P</figure-callout>}24\&prime;=\mathrm{<figure-callout\; id="24"\; label="L"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">L</figure-callout>}24\&CircleTimes;\mathrm{<figure-callout\; id="24"\; label="R"\; filenames="A200910050475E00142.png"\; state="\{\{state\}\}">R</figure-callout>}24,$

...

$\mathrm{Pmn}\&prime;=\mathrm{Lmn}\&CircleTimes;\mathrm{Rmn},$

Adopt above-mentioned method, because adjacent pixels point all has relevance, so the correlation height of data processing can reduce the complexity of algorithm.And, therefore can change with other form easily because each adjacent pixels point all is to have relevance.

For above-mentioned described specific algorithm, be interval interpolation encryption algorithm in the present embodiment, can represent with following logical expression:

P11′＝L11*1+R11*0，P12′＝L12*0+R12*1，P13′＝L13*1+R13*0…

P21′＝L21*0+R21*1，P22′＝L22*1+R22*0，P23′＝L23*0+R23*1…

P31′＝L31*1+R31*0，P32′＝L32*0+R32*1，P33′＝L33*1+R33*0…

P41′＝L41*0+R41*1，P42′＝L42*1+R42*0，P43′＝L43*0+R43*1…

The essence of this interval interpolation encryption algorithm is that the pixel with the pixel of left-eye image and eye image inserts in the mode of alternate intervals and is combined into one group of new video image, as shown in Figure 5.In same pixel column or same pixel column, any adjacent pixels point can all not belong to left-eye image or belong to eye image.Interval interpolation encryption algorithm in the present embodiment, the pixel arrangement mode of the new video image that it is reassembled into is similar to chess chessboard pattern, therefore also with this interval interpolation encryption algorithm popular be called chess gridiron pattern algorithm.

New video image according to above-mentioned interval interpolation encryption algorithm obtains can not broadcast a frame left-eye image and a frame eye image in turn as traditional three-dimensional video-frequency form is realized, but the broadcast of a frame one frame in regular turn, as shown in Figure 4.Therefore, if in data transmission procedure, intercept one section video flowing, do not pass through the corresponding decoding algorithm of this programme, the video image that obtains so is can not watching of bluring, thereby make three-dimensional video-frequency form of the present invention have good confidentiality, possessed anti-piracy function.

Image encoding unit 21 connection layouts are as synthesis unit 20, and the video image that image synthesis unit 20 is exported carries out compressed encoding, for example MPEG compression or the compression of HD high definition etc.Memory 22 connection layouts are as coding unit 21, the video image after the store compressed.Picture decoding unit 23 connected storages 22 carry out decompression to the video image after the compression in the memory 22, and decompression is the inverse process that the compression of image encoding unit 21 is handled.Separation of images unit 24 connection layouts are split as left-eye image and eye image as decoding unit 23 with the video image after 23 decompressions of picture decoding unit, and split process is the inverse process of the anabolic process of image synthesis unit 20.

The foregoing description provides to those of ordinary skills and realizes or use of the present invention; those of ordinary skills can be under the situation that does not break away from invention thought of the present invention; the foregoing description is made various modifications or variation; thereby protection scope of the present invention do not limit by the foregoing description, and should be the maximum magnitude that meets the inventive features that claims mention.

Claims (10)

1, a kind of three-dimensional video-frequency implementation method comprises:

(1) respectively left-eye image and eye image are divided into a plurality of pixels;

(2) pixel of the correspondence of left-eye image and eye image is inserted in the mode of alternate intervals be combined into one group of new video image.

2, three-dimensional video-frequency implementation method according to claim 1 is characterized in that, in step (1), left-eye image is by pixel L _{M, n}Constitute, eye image is by pixel R _{M, n}Constitute;

In step (2), the pixel P of new video image _{M, n}=L _{M, n}Or R _{M, n}

3, three-dimensional video-frequency implementation method according to claim 1 is characterized in that, also comprises afterwards in step (2):

Video image is carried out compressed encoding.

4, three-dimensional video-frequency implementation method according to claim 1 is characterized in that, also comprises afterwards in step (2):

When playing, new video image is split as left-eye image and eye image with the anti-process of step (2).

5, three-dimensional video-frequency implementation method according to claim 3 is characterized in that, also comprises:

The process that when playing, decompresses earlier.

6, a kind of three-dimensional video-frequency implement device comprises:

The image synthesis unit is divided into a plurality of pixels with left-eye image and eye image respectively, the pixel of the correspondence of left-eye image and eye image is inserted in the mode of alternate intervals be combined into one group of new video image.

7, three-dimensional video-frequency implement device according to claim 6 is characterized in that, this device also comprises:

The image encoding unit connects this image synthesis unit, and the video image that this image synthesis unit is exported carries out compressed encoding.

8, three-dimensional video-frequency implement device according to claim 7 is characterized in that, this device also comprises:

Memory connects this image encoding unit, the video image after the store compressed.

9, three-dimensional video-frequency implement device according to claim 8 is characterized in that, this device also comprises:

The picture decoding unit connects this memory, and the video image after the compression in this memory is carried out decompression.

10, three-dimensional video-frequency implement device according to claim 9 is characterized in that, this device also comprises:

The separation of images unit connects this picture decoding unit, and the video image after will decompressing with the anti-process in this image synthesis unit is divided into left-eye image and eye image.

Images