CN106028019A - Method for rapidly detecting three-dimensional format of video - Google Patents

Abstract

The invention discloses a method for rapidly detecting a three-dimensional format of a video. According to the characteristics that only horizontal parallax rather than vertical parallax can exist between a left view and a right view, and that features on the aspects of brightness/chromaticity and the like of the left view and the right view need to be kept highly consistent as required during making of three-dimensional video content, format detection is performed by calculation of a distance between projection vectors of different views, and only an addition operation is performed on the calculated amount in a whole calculation process, so that rapid implementation of software and hardware can be realized conveniently/economically. Secondly, only necessary calculation/detection is performed at the start of new programs according to the fact that the same video program is unique in format, so that the method is particularly suitable for systems such as mobile equipment (tablet computers and laptops) with very strict requirements on power consumption and manufacturing cost. Through adoption of the method for rapidly detecting the three-dimensional format of the video, the three-dimensional format of the video can be detected rapidly.

Description

Video 3 dimensional format method for quick

Technical field

The invention belongs to video format detection technique field, relate to a kind of video format detection method, particularly relate to one Video 3 dimensional format method for quick.

Background technology

Along with the film of three-dimensional (3D) video the starting broadcasting of more and more universal and 3D digital television program in video website, can With prediction within following a period of time, long-term (and always) is coexisted by 2D content and 3D content.For terminal presentation facility, as What detects video frequency program form automatically is 2D or 3D, and it is requisite for correctly presenting for content.

In current 3D video, having a big class is by 3D content arrangement 2D frame, in order to existing 2D display system compatibility. According to 3D subframe array format in 2D frame, this type of 3D video is divided into following several form, and wherein 3D form in left and right is with upper and lower 3D form is the most universal form.

(1) left and right 3D form (SBS): about 3D view is arranged side by side in 2D frame；

(2) upper and lower 3D form (TB): about 3D view arrangements upper and lower in 2D frame；

(3) row intertexture 3D form (LI): about intertexture 3D, the row of view allows left and right view arrangements in 2D frame；

(4) column interleaving 3D form (CI): about intertexture 3D, the row of view allow left and right view arrangements in 2D frame；

(5) grid-staggered 3D form (CB): allow about 3D view arrangements in 2D frame so that grid pattern is staggered；

(6) 3D form (CBS) of grid-side by side: allow about 3D view be arranged side by side in 2D frame with grid pattern；

(7) 2D+ depth format (2D+D): 2D subframe and depth information subframe are arranged side by side in 2D frame.

According to viewing comfortableness requirement, except requiring that the brightness/chroma of left and right view etc. is each when 3 d video content makes Aspect feature has high consistency, the most only allows to there is horizontal disparity, it is impossible to have vertical parallax；Otherwise can cause to beholder The biggest discomfort.Meanwhile, according to the convention of existing video program production, the form of each program determines that and uniquely 's.It is 2D video, if 3D content then uses a kind of form above, does not exist and comprise multiple regarding inside a program Frequently form.

In view of this, nowadays in the urgent need to designing a kind of new video format detection mode, in order to overcome existing detection side The drawbacks described above that formula exists.

Summary of the invention

The technical problem to be solved is: provides a kind of video 3 dimensional format method for quick, can quickly examine Survey the 3 dimensional format of video.

For solving above-mentioned technical problem, the present invention adopts the following technical scheme that

A kind of video 3 dimensional format method for quick, described detection method comprises the steps:

Step one, judge whether video is upper and lower 3D format video or left and right 3D format video, including:

S1, new video program start, to the first enumerator, the second counter O reset；

S2, obtain a frame width be N highly for the image P of L, input picture is divided into 4 sub regions, formed upper left, Lower-left, upper right, bottom right 4 sub regions；Carry out view division by left and right 3D, upper and lower 3D form more simultaneously；Format by left and right 3D Time, left view is on the left side, and right view is on the right；When formatting by upper and lower 3D, on top, lower view is below for top view；

S3, for YUV mode image, only luminance component Y is calculated, the level of 4 sub regions in calculation procedure S2 Projection vector Y_hWith vertical projection vector Y_v.Respectively obtain the first subregion horizontal projection vectorFirst subregion is vertically thrown Shadow vectorSecond subregion horizontal projection vectorSecond subregion vertical projection vector3rd subregion level is thrown Shadow vector3rd subregion vertical projection vector4th subregion horizontal projection vector4th subregion is vertically thrown Shadow vectorWherein, Y_hDimension be L/2, Y_vDimension be N/2.

Second subregion horizontal projection vectorWith vertical projection vectorComputing formula as follows, wherein Y_ijIt it is the i-th row The Y-component of jth row pixel；

$Y_h^2=\left[\begin{array}{c}{Y}_{1h}\\ Y_{\mathrm{2h}}\\ Y_{\mathrm{3h}}\\ \·\\ \·\\ \·\\ Y_{\frac{L}{2}h}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{1j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{2j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{3j}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{\frac{L}{2}j}\end{array}\right];$

$Y_v^2=\left[\begin{array}{c}{Y}_{1v}\\ Y_{2v}\\ Y_{3v}\\ \·\\ \·\\ \·\\ Y_{\frac{N}{2}v}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i1}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i2}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i3}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i\frac{L}{2}}\end{array}\right];$

Other 3 regionsCalculating process same, it is only necessary to according to difference The value amendment span of region i, j.First area is3rd region is 4th region is

For RGB mode image, transfer image to YUV color space from RGB color, according still further to said process to Y Process；Or directly at rgb space, tri-components of R, G, B are all processed by the process of above-mentioned Y, thus judges lattice Formula；

S4, relation according to subregion in step S2 Yu left and right view, obtain the horizontal projection vector that left view is corresponding And vertical projection vectorThe horizontal projection vector of right viewAnd vertical projection vector

$Y_h^{11}=Y_h^2\left|\right|Y_h^3;$

$Y_v^{11}=Y_v^2+Y_v^3;$

$Y_h^{12}=Y_h^1\left|\right|Y_h^4;$

$Y_v^{12}=Y_v^1+Y_v^4;$

Wherein,WithDimension is L,WithDimension is N/2；

The concatenation operation of two vectorial V and T of definition is:

$V\left|\right|T=\left[\begin{array}{c}{v}_1\\ v_2\\ v_3\\ \·\\ \·\\ \·\\ v_n\\ t_1\\ t_2\\ \·\\ \·\\ \·\\ t_m\end{array}\right]$

Wherein, the dimension of vector V is n, and the dimension of vector T is m, and the new vector dimension after connection is n+m；

Equally, the horizontal projection vector that top view is corresponding is obtainedAnd vertical projection vectorThe level of lower view is thrown Shadow vectorAnd vertical projection vector

$Y_h^{21}=Y_h^2+Y_h^1;$

$Y_v^{21}=Y_v^2\left|\right|Y_v^1;$

$Y_h^{22}=Y_h^3+Y_h^4;$

$Y_v^{22}=Y_v^3\left|\right|Y_v^4;$

S5, the distance calculated between left view and right view, top view and the lower each projection vector of viewWith

$D_h^1=Y_h^{11}-Y_h^{12};$

$D_v^1=Y_v^{11}-Y_v^{12};$

$D_h^2=Y_h^{21}-Y_h^{22};$

$D_v^2=Y_v^{21}-Y_v^{22};$

If S6 left view and right view horizontal projection vector distanceLess than first threshold TH_h, and projection vector away from FromLess than Second Threshold TH_v, then the first Counter Value adds 1；If top view and lower view horizontal projection vector distanceLittle In first threshold TH_h, and projection vector distanceLess than Second Threshold TH_v, then the second Counter Value adds 1；

S7, repetition step S2～S6, until step S6 has been performed m time；

If the value of S8 the first enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program form detected For left and right 3D video format；

If the value of the second enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program form detected is upper Lower 3D video format；

If the value of S9 two above enumerator is all unsatisfactory for condition, then the current video program detected is not left and right 3D Video format, is not upper and lower 3D video format；

Step 2, judge that whether video is row intertexture 3D form LI video or column interleaving 3D form CI video or grid-friendship The 3D form CBS video of wrong 3D form CB video or grid-side by side；Including: with step S2 of step one process, step S3, step S4 is similar to, and according to the rule of different-format left and right view arrangement, the region carrying out left and right view divides, and obtains the water of left and right view Flat projection and upright projection；The region, left and right of row intertexture 3D form LI divides and divides according to the parity rows of image pixel, institute The pixel having odd-numbered line carries out floor projection/upright projection, obtains the horizontal/vertical projection vector of a view；All even numbers The pixel of row carries out floor projection/upright projection, obtains the horizontal/vertical projection vector of another view；Subsequent processing steps Then identical with step S5 to the S8 of step one.

If step 3 step one, step 2 are all not detected by 3D format video, then judge the current video joint detected Mesh is 2D video format.

A kind of video 3 dimensional format method for quick, it is characterised in that described detection method comprises the steps:

Enumerator initializes；

Obtain a two field picture, according to the n kind 3D video format of required detection, image is carried out n kind left and right view and draws Point；

Calculate each view to divide, around the horizontal projection vector of the pixel component of view and vertical projection vector；

Calculate each view to divide, the horizontal projection vector distance of corresponding left and right view and vertical projection vector away from From；

Divide for each view, if horizontal projection vector distance and vertical projection vector distance are less than given threshold Value, then the 3D format counter that the division of this view is corresponding adds 1；

Repeated m time, check which kind of 3D format counter value more than k*m, k ∈ (0.5,1], more than then concluding that form is this 3D Form；If all Counter Values are all unsatisfactory for, then conclude that form is 2D form.

As a preferred embodiment of the present invention, described 3D form is one of them in following 6 kinds of forms: left and right 3D lattice The 3D lattice of formula SBS, upper and lower 3D form TB, row intertexture 3D form LI, column interleaving 3D form, grid-staggered 3D form, grid-side by side Formula.

As a preferred embodiment of the present invention, for the most universal left and right 3D and upper and lower 3D form, described inspection Survey method specifically includes following steps:

S1, new video program start, to the first enumerator, the second counter O reset；

S2, obtain a frame width be N highly for the image P of L, input picture is divided into 4 sub regions, formed upper left, Lower-left, upper right, bottom right 4 sub regions；Carry out view division by left and right 3D, upper and lower 3D form more simultaneously；Format by left and right 3D Time, left view is on the left side, and right view is on the right；When formatting by upper and lower 3D, on top, lower view is below for top view；

S3, for YUV mode image, the luminance component Y of 4 sub regions in step S2 is calculated, calculation procedure S2 In the horizontal projection vector Y of 4 sub regions_hWith vertical projection vector Y_v；

For RGB mode image, in calculation procedure S2, tri-components of R, G, B of 4 sub regions calculate, and obtain correspondence Horizontal projection vector R_h、G_h、B_hWith vertical projection vector R_v、G_v、B_v；

Above-mentioned each component Y, R, G, B are labeled as X, for component X, respectively obtain the first subregion horizontal projection vectorFirst subregion vertical projection vectorSecond subregion horizontal projection vectorSecond subregion vertical projection vector3rd subregion horizontal projection vector3rd subregion vertical projection vector4th subregion horizontal projection vector4th subregion vertical projection vectorWherein, X_hDimension be L/2, X_vDimension be N/2；

S4, relation according to subregion in step S2 Yu left and right view, obtain the horizontal projection vector that left view is corresponding And vertical projection vectorThe horizontal projection vector of right viewAnd vertical projection vectorWherein,WithDimension For L,WithDimension is N/2；

Equally, the horizontal projection vector that top view is corresponding is obtainedAnd vertical projection vectorThe level of lower view is thrown Shadow vectorAnd vertical projection vector

S5, the distance calculated between left view and right view, top view and the lower each projection vector of viewWith

If S6 left view and right view horizontal projection vector distanceLess than first threshold TH_h, and projection vector away from FromLess than Second Threshold TH_v, then the first Counter Value adds 1；If top view and lower view horizontal projection vector distanceLittle In first threshold TH_h, and projection vector distanceLess than Second Threshold TH_v, then the second Counter Value adds 1；

S7, repetition step S2～S6, until step S6 has been performed m time；

If the value of S8 the first enumerator is more than or equal to setting value, then the current video program form detected is left and right 3D Video format；

If the value of the second enumerator is more than or equal to setting value, then the current video program form detected is upper and lower 3D video Form；

If the value of two above enumerator is all unsatisfactory for condition, then the current video program detected is 2D video format.

As a preferred embodiment of the present invention, in described step S3, the second subregion horizontal projection vectorWith vertical Projection vectorComputing formula as follows, wherein Y_ijIt it is the Y-component of the i-th row jth row pixel；

$Y_h^2=\left[\begin{array}{c}{Y}_{1h}\\ Y_{\mathrm{2h}}\\ Y_{\mathrm{3h}}\\ \·\\ \·\\ \·\\ Y_{\frac{L}{2}h}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{1j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{2j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{3j}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{\frac{L}{2}j}\end{array}\right];$

$Y_v^2=\left[\begin{array}{c}{Y}_{1v}\\ Y_{2v}\\ Y_{3v}\\ \·\\ \·\\ \·\\ Y_{\frac{N}{2}v}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i1}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i2}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i3}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i\frac{L}{2}}\end{array}\right];$

Other 3 regionsCalculation be similar to therewith.

As a preferred embodiment of the present invention, in described step S4,

$X_h^{21}=X_h^2+X_h^1;X_v^{21}=X_v^2\left|\right|X_v^1;X_h^{22}=X_h^3+X_h^4;X_v^{22}=X_v^3\left|\right|X_v^4;$

The concatenation operation of definition two vectorial V, T is:

$V\left|\right|T=\left[\begin{array}{c}{v}_1\\ v_2\\ v_3\\ \·\\ \·\\ \·\\ v_n\\ t_1\\ t_2\\ \·\\ \·\\ \·\\ t_m\end{array}\right]$

The dimension of V be the dimension of n, T be m, the new vector dimension after connection is n+m.

As a preferred embodiment of the present invention, in described step S5,

The beneficial effects of the present invention is: the video 3 dimensional format method for quick that the present invention proposes, can quickly detect The 3 dimensional format of video.

The present invention is by a kind of index of design, and this index calculates in horizontal disparity unrelated, the most relevant to vertical parallax, then pin Left and right view to 3D content, this index has high consistency, and this point has significant difference, Jin Erjian with common 2D video Measure the 3D/2D form of video.

Can only there is horizontal disparity according to the left and right view required by 3 d video content making in the present invention, do not allow to exist Vertical parallax, and each side feature such as left and right view brightness/chroma must these features highly consistent, use and calculate difference and regard Distance between the projection vector of figure carries out format detection, and the amount of calculation of whole process only exists additive operation, the most permissible Convenient/economic applicable software and hardware quickly realizes.

Secondly, the inventive method make use of the fact that same video frequency program form be'ss unique condition, simply at each new joint Do when mesh starts necessity calculating/detection, be particularly well-suited to mobile device (panel computer, notebook computer) this kind of to power consumption, Manufacturing cost requires the harshest system.

Accompanying drawing explanation

Fig. 1 is the schematic diagram that input picture is divided into 4 sub regions.

Fig. 2 is the schematic diagram formatted input picture by left and right 3D.

Fig. 3 is the schematic diagram formatted input picture by upper and lower 3D.

Fig. 4 is the flow chart of video 3 dimensional format method for quick of the present invention.

Detailed description of the invention

Describe the preferred embodiments of the present invention below in conjunction with the accompanying drawings in detail.

Embodiment one

Referring to Fig. 4, present invention is disclosed a kind of video 3 dimensional format method for quick, the index of this method design is Pixel value in left and right view area in video is the most both horizontally and vertically projected, obtains horizontal projection vector with vertical Projection vector.Due to the horizontal disparity of left and right view, floor projection is calculated impact little, and it is big that upright projection calculates impact, Therefore the difference of the horizontal projection vector of left and right view little (in other words apart from close), and the significant difference of vertical projection vector (or distance is bigger).3D video is set up by this feature, there is not 2D video, such that it is able to distinguish 3D and 2D video Form.

The process that calculates in detail of horizontal projection vector is as follows, has the RGB mode image P of L row N row pixel to a frame (if YUV pattern, then only calculate luminance component Y), projects RGB component to horizontal direction respectively, obtains 3 Individual dimension is all the vectorial R of L_h、G_h、B_h.With R_hAs a example by, computing formula is as follows, wherein R_ijIt is that the R of the i-th row jth row pixel divides Amount.

$R_h=\left[\begin{array}{c}{R}_{1h}\\ R_{2h}\\ R_{3h}\\ \·\\ \·\\ \·\\ R_{Lh}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^N{R}_{1j}\\ {\mathrm{\Σ}}_{j=1}^N{R}_{2j}\\ {\mathrm{\Σ}}_{j=1}^N{R}_{3j}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{j=1}^N{R}_{Lj}\end{array}\right]$

VectorWithBetween distance definition be DR_h, it is calculated as follows.

${\mathrm{DR}}_h=R_h^1-R_h^2={\mathrm{\Σ}}_{j=1}^L|R_{jh}^1-R_{jh}^2|$

Equally, component G with B is obtained distance DG of similar horizontal projection vector_hAnd DB_h。

Similar with floor projection, obtain image P vertical direction projection vector R_v、G_v、B_v(its dimension is N), and vector Between distance DR_v、DG_vAnd DB_v。

The concatenation operation defining two vectorial V (dimension is n) and T (dimension is m) is that the new vector dimension after connection is n+ m

$V\left|\right|T=\left[\begin{array}{c}{v}_1\\ v_2\\ v_3\\ \·\\ \·\\ \·\\ v_n\\ t_1\\ t_2\\ \·\\ \·\\ \·\\ t_m\end{array}\right].$

As a example by the most universal left and right 3D and upper and lower 3D form, whole format detection process comprises the steps:

[step S1] new video program starts, to the first enumerator, the second counter O reset；

[step S2] obtain a frame width be N highly for the image P of L, input picture is divided into 4 sub regions, is formed Upper left, lower-left, upper right, bottom right 4 sub regions, as shown in Figure 1；Carry out view division by left and right 3D, upper and lower 3D form more simultaneously； When formatting by left and right 3D (as shown in Figure 2), left view Figure 11 is on the left side, and right view 12 is on the right；When formatting by upper and lower 3D (as Shown in Fig. 3), on top, lower view 22 is below for top view 21；

Luminance component Y, for YUV mode image, is calculated by [step S3], the water of 4 sub regions in calculation procedure S2 Flat projection vector Y_hWith vertical projection vector Y_v；Respectively obtain the first subregion horizontal projection vectorFirst subregion is vertical Projection vectorSecond subregion horizontal projection vectorSecond subregion vertical projection vector3rd subregion level Projection vector3rd subregion vertical projection vector4th subregion horizontal projection vector4th subregion is vertical Projection vectorWherein, Y_hDimension be L/2, Y_vDimension be N/2；

For RGB mode image, image can be transferred to YUV color space from RGB color, according still further to said process To Y process.In view of multiply-add operations a large amount of required for color space conversion, it is also possible to directly at rgb space, to R, G, B Three components are all processed by the process of above-mentioned Y, thus judge form.

For having the RGB mode image of L row N row pixel, in calculation procedure S2 the floor projection of 4 sub regions to Amount R_h、G_h、B_hWith vertical projection vector R_v、G_v、B_v；

Each sub regions horizontal projection vector R_h、G_h、B_hComputational methods be: R, G, B component are entered to horizontal direction respectively Row projection, obtains the horizontal projection vector R that 3 dimensions are all L_h、G_h、B_h；

$R_h=\left[\begin{array}{c}{R}_{1h}\\ R_{2h}\\ R_{\mathrm{3h}}\\ \·\\ \·\\ \·\\ R_{Lh}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^N{R}_{1j}\\ {\mathrm{\Σ}}_{j=1}^N{R}_{2j}\\ {\mathrm{\Σ}}_{j=1}^N{R}_{3j}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{j=1}^N{R}_{Lj}\end{array}\right];$

$G_h=\left[\begin{array}{c}{G}_{1h}\\ G_{2h}\\ G_{3h}\\ \·\\ \·\\ \·\\ G_{Lh}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^N{G}_{1j}\\ {\mathrm{\Σ}}_{j=1}^N{G}_{2j}\\ {\mathrm{\Σ}}_{j=1}^N{G}_{3j}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{j=1}^N{G}_{Lj}\end{array}\right];$

$B_h=\left[\begin{array}{c}{B}_{1h}\\ B_{2h}\\ B_{3h}\\ \·\\ \·\\ \·\\ B_{Lh}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^N{B}_{1j}\\ {\mathrm{\Σ}}_{j=1}^N{B}_{2j}\\ {\mathrm{\Σ}}_{j=1}^N{B}_{3j}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{j=1}^N{B}_{Lj}\end{array}\right];$

Wherein, R_ijIt is the R component of the i-th row jth row pixel, G_ijIt is the G component of the i-th row jth row pixel, B_ijIt is i-th The B component of row jth row pixel；

Each sub regions vertical projection vector R_v、G_v、B_vComputational methods be: R, G, B component are entered to vertical direction respectively Row projection, obtains the horizontal projection vector R that 3 dimensions are all N_v、G_v、B_v；

$R_v=\left[\begin{array}{c}{R}_{1v}\\ R_{2v}\\ R_{3v}\\ \·\\ \·\\ \·\\ R_{Nv}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^L{R}_{1i}\\ {\mathrm{\Σ}}_{i=1}^L{R}_{2i}\\ {\mathrm{\Σ}}_{i=1}^L{R}_{3i}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{i=1}^L{R}_{Li}\end{array}\right];$

$G_v=\left[\begin{array}{c}{G}_{1v}\\ G_{2v}\\ G_{3v}\\ \·\\ \·\\ \·\\ G_{Nv}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^L{G}_{1i}\\ {\mathrm{\Σ}}_{i=1}^L{G}_{2i}\\ {\mathrm{\Σ}}_{i=1}^L{G}_{3i}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{i=1}^L{G}_{Li}\end{array}\right];$

$B_v=\left[\begin{array}{c}{B}_{1v}\\ B_{2v}\\ B_{3v}\\ \·\\ \·\\ \·\\ B_{Nv}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^L{B}_{1i}\\ {\mathrm{\Σ}}_{i=1}^L{B}_{2i}\\ {\mathrm{\Σ}}_{i=1}^L{B}_{3i}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{i=1}^L{B}_{Li}\end{array}\right];$

Wherein, R_ijIt is the R component of the i-th row jth row pixel, G_ijIt is the G component of the i-th row jth row pixel, B_ijIt is i-th The B component of row jth row pixel；

[step S4], according to the relation of subregion in step S2 Yu left and right view, obtains the floor projection that left view Figure 11 is corresponding VectorAnd vertical projection vectorThe horizontal projection vector of right view 12And vertical projection vector

$Y_h^{11}=Y_h^2\left|\right|Y_h^3;$

$Y_v^{11}=Y_v^2+Y_v^3;$

$Y_h^{12}=Y_h^1\left|\right|Y_h^4;$

$Y_v^{12}=Y_v^1+Y_v^4;$

Wherein,WithDimension is L,WithDimension is N/2；

The concatenation operation of two vectorial V and T of definition is:

$V\left|\right|T=\left[\begin{array}{c}{v}_1\\ v_2\\ v_3\\ \·\\ \·\\ \·\\ v_n\\ t_1\\ t_2\\ \·\\ \·\\ \·\\ t_m\end{array}\right]$

Wherein, the dimension of vector V is n, and the dimension of vector T is m, and the new vector dimension after connection is n+m；

Equally, the horizontal projection vector of top view 21 correspondence is obtainedAnd vertical projection vectorThe water of lower view 22 Flat projection vectorAnd vertical projection vector

$Y_h^{21}=Y_h^2+Y_h^1;$

$Y_v^{21}=Y_v^2\left|\right|Y_v^1;$

$Y_h^{22}=Y_h^3+Y_h^4;$

$Y_v^{22}=Y_v^3\left|\right|Y_v^4;$

[step S5] calculates the distance between left view Figure 11 and right view 12, top view 21 and the lower each projection vector of view 22With

$D_h^1=Y_h^{11}-Y_h^{12};$

$D_v^1=Y_v^{11}-Y_v^{12};$

$D_h^2=Y_h^{21}-Y_h^{22};$

$D_v^2=Y_v^{21}-Y_v^{22};$

[step S6] if left view Figure 11 and right view 12 horizontal projection vector distanceLess than first threshold TH_h, and Projection vector distanceLess than Second Threshold TH_v, then the first Counter Value adds 1；If top view 21 and lower view 22 level are thrown Shadow vector distanceLess than first threshold TH_h, and projection vector distanceLess than Second Threshold TH_v, then the second Counter Value Add 1；

[step S7] repeats step S2～S6, until step S6 has been performed m time；

[step S8] if the value of the first enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program detected Form is left and right 3D video format；

[step S9] if the value of the second enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program detected Form is upper and lower 3D video format.

Can also be according to above-mentioned flow process, it is judged that whether current video program form is other several 3D forms, such as row is handed over Knit 3D form (LI), column interleaving 3D form (CI), grid-staggered 3D form (CB), 3D form (CBS) of grid-side by side.Detected Journey can be similar with S2, S3, S4 of said process, according to the rule of different-format left and right view arrangement, carries out left and right view Region divides, and obtains floor projection and the upright projection of left and right view.Such as going intertexture 3D form (LI), region, left and right divides and presses Parity rows according to image pixel divides, and the pixel of all odd-numbered lines carries out floor projection/upright projection, obtains a view Horizontal/vertical projection vector；The pixel of all even number lines carries out floor projection/upright projection, obtains the water of another view Flat/vertical projection vector.Subsequent processing steps then catches up with that to state S5 to S8 identical.

If be all not detected by 3D format video according to said process, then judge that the current video program detected is 2D video Form.

Embodiment two

A kind of video 3 dimensional format method for quick, described detection method comprises the steps:

Step one, judge whether video is upper and lower 3D format video or left and right 3D format video, including:

S1, new video program start, to the first enumerator, the second counter O reset；

S2, obtain a frame width be N highly for the image P of L, input picture is divided into 4 sub regions, formed upper left, Lower-left, upper right, bottom right 4 sub regions；Carry out view division by left and right 3D, upper and lower 3D form more simultaneously；Format by left and right 3D Time, left view Figure 11 is on the left side, and right view 12 is on the right；When formatting by upper and lower 3D, top view 21 on top, lower view 22 under Limit；

S3, for YUV mode image, only luminance component Y is calculated, the level of 4 sub regions in calculation procedure S2 Projection vector Y_hWith vertical projection vector Y_v.Respectively obtain the first subregion horizontal projection vectorFirst subregion is vertically thrown Shadow vectorSecond subregion horizontal projection vectorSecond subregion vertical projection vector3rd subregion level is thrown Shadow vector3rd subregion vertical projection vector4th subregion horizontal projection vector4th subregion is vertically thrown Shadow vectorWherein, Y_hDimension be L/2, Y_vDimension be N/2.

Second subregion horizontal projection vectorWith vertical projection vectorComputing formula as follows, wherein Y_ijIt it is the i-th row The Y-component of jth row pixel；

$Y_h^2=\left[\begin{array}{c}{Y}_{1h}\\ Y_{\mathrm{2h}}\\ Y_{\mathrm{3h}}\\ \·\\ \·\\ \·\\ Y_{\frac{L}{2}h}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{1j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{2j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{3j}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{\frac{L}{2}j}\end{array}\right];$

$Y_v^2=\left[\begin{array}{c}{Y}_{1v}\\ Y_{2v}\\ Y_{3v}\\ \·\\ \·\\ \·\\ Y_{\frac{N}{2}v}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i1}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i2}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i3}\\ \·\\ \·\\ \·\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i\frac{L}{2}}\end{array}\right];$

Other 3 regionsCalculating process same, it is only necessary to according to difference (those skilled in the art can obtain other 3 regions according to description of the invention to the value amendment span of region i, j Do not repeat).First area is3rd region For4th region is

For RGB mode image, transfer image to YUV color space from RGB color, according still further to said process to Y Process；Or directly at rgb space, tri-components of R, G, B are all processed by the process of above-mentioned Y, thus judges lattice Formula；

S4, relation according to subregion in step S2 Yu left and right view, obtain the horizontal projection vector that left view Figure 11 is correspondingAnd vertical projection vectorThe horizontal projection vector of right view 12And vertical projection vector

$Y_h^{11}=Y_h^2\left|\right|Y_h^3;$

$Y_v^{11}=Y_v^2+Y_v^3;$

$Y_h^{12}=Y_h^1\left|\right|Y_h^4;$

$Y_v^{12}=Y_v^1+Y_v^4;$

Wherein,WithDimension is L,WithDimension is N/2；

The concatenation operation of two vectorial V and T of definition is:

$V\left|\right|T=\left[\begin{array}{c}{v}_1\\ v_2\\ v_3\\ \·\\ \·\\ \·\\ v_n\\ t_1\\ t_2\\ \·\\ \·\\ \·\\ t_m\end{array}\right]$

Wherein, the dimension of vector V is n, and the dimension of vector T is m, and the new vector dimension after connection is n+m；

Equally, the horizontal projection vector of top view 21 correspondence is obtainedAnd vertical projection vectorThe water of lower view 22 Flat projection vectorAnd vertical projection vector

$Y_h^{21}=Y_h^2+Y_h^1;$

$Y_v^{21}=Y_v^2\left|\right|Y_v^1;$

$Y_h^{22}=Y_h^3+Y_h^4;$

$Y_v^{22}=Y_v^3\left|\right|Y_v^4;$

S5, the distance calculated between left view Figure 11 and right view 12, top view 21 and the lower each projection vector of view 22 With

$D_h^1=Y_h^{11}-Y_h^{12};$

$D_v^1=Y_v^{11}-Y_v^{12};$

$D_h^2=Y_h^{21}-Y_h^{22};$

$D_v^2=Y_v^{21}-Y_v^{22};$

If S6 left view Figure 11 and right view 12 horizontal projection vector distanceLess than first threshold TH_h, and project to Span fromLess than Second Threshold TH_v, then the first Counter Value adds 1；If top view 21 and lower view 22 horizontal projection vector DistanceLess than first threshold TH_h, and projection vector distanceLess than Second Threshold TH_v, then the second Counter Value adds 1；

S7, repetition step S2～S6, until step S6 has been performed m time；

If the value of S8 the first enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program form detected For left and right 3D video format；

If the value of the second enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program form detected is upper Lower 3D video format；

If the value of S9 two above enumerator is all unsatisfactory for condition, then the current video program detected is not left and right 3D Video format, is not upper and lower 3D video format；

Step 2, judge that whether video is row intertexture 3D form LI video or column interleaving 3D form CI video or grid-friendship The 3D form CBS video of wrong 3D form CB video or grid-side by side；Including: with step S2 of step one process, step S3, step S4 is similar to, and according to the rule of different-format left and right view arrangement, the region carrying out left and right view divides, and obtains the water of left and right view Flat projection and upright projection；The region, left and right of row intertexture 3D form LI divides and divides according to the parity rows of image pixel, institute The pixel having odd-numbered line carries out floor projection/upright projection, obtains the horizontal/vertical projection vector of a view；All even numbers The pixel of row carries out floor projection/upright projection, obtains the horizontal/vertical projection vector of another view；Subsequent processing steps Then identical with step S5 to the S8 of step one.

If step 3 step one, step 2 are all not detected by 3D format video, then judge the current video joint detected Mesh is 2D video format.

Embodiment three

A kind of video 3 dimensional format method for quick, described detection method comprises the steps:

Enumerator initializes；

Obtaining a two field picture, according to the n kind 3D video format of required detection, (described 3D form is in following 6 kinds of forms One of them: left and right 3D form SBS, upper and lower 3D form TB, row intertexture 3D form LI, column interleaving 3D form, grid-staggered 3D The 3D form of form, grid-side by side), image is carried out the division of n kind left and right view；

Calculate each view to divide, around the horizontal projection vector of the pixel component of view and vertical projection vector；

Calculate each view to divide, the horizontal projection vector distance of corresponding left and right view and vertical projection vector away from From；

Divide for each view, if horizontal projection vector distance and vertical projection vector distance are less than given threshold Value, then the 3D format counter that the division of this view is corresponding adds 1；

Repeated m time, check which kind of 3D format counter value more than k*m, k ∈ (0.5,1], more than then concluding that form is this 3D Form；If all Counter Values are all unsatisfactory for, then conclude that form is 2D form.

In sum, the video 3 dimensional format method for quick that the present invention proposes, can quickly detect the three-dimensional lattice of video Formula.The present invention is by a kind of index of design, and this index calculates in horizontal disparity unrelated, the most relevant to vertical parallax, then for 3D The left and right view of content, this index has high consistency, and this point has significant difference with common 2D video, and then detects The 3D/2D form of video.

Can only there is horizontal disparity according to the left and right view required by 3 d video content making in the present invention, do not allow to exist Vertical parallax, and each side feature such as left and right view brightness/chroma must these features highly consistent, use and calculate difference and regard Distance between the projection vector of figure carries out format detection, and the amount of calculation of whole process only exists additive operation, the most permissible Convenient/economic applicable software and hardware quickly realizes.

Secondly, the inventive method make use of the fact that same video frequency program form be'ss unique condition, simply at each new joint Do when mesh starts necessity calculating/detection, be particularly well-suited to mobile device (panel computer, notebook computer) this kind of to power consumption, Manufacturing cost requires the harshest system.

Here description of the invention and application is illustrative, is not wishing to limit the scope of the invention to above-described embodiment In.The deformation of embodiments disclosed herein and change are possible, real for those skilled in the art The various parts with equivalence of replacing executing example are known.It should be appreciated by the person skilled in the art that without departing from the present invention Spirit or essential characteristics in the case of, the present invention can in other forms, structure, layout, ratio, and with other assembly, Material and parts realize.In the case of without departing from scope and spirit of the present invention, embodiments disclosed herein can be entered Other deformation of row and change.

Claims (7)

1. a video 3 dimensional format method for quick, it is characterised in that described detection method comprises the steps:

Step one, judge whether video is upper and lower 3D format video or left and right 3D format video, including:

S1, new video program start, to the first enumerator, the second counter O reset；

S2, obtain a frame width be N highly for the image P of L, input picture is divided into 4 sub regions, formed upper left, lower-left, Upper right, bottom right 4 sub regions；Carry out view division by left and right 3D, upper and lower 3D form more simultaneously；When formatting by left and right 3D, left View (11) is on the left side, and right view (12) is on the right；When formatting by upper and lower 3D, top view (21) on top, lower view (22) Below；

S3, for YUV mode image, only luminance component Y is calculated, the floor projection of 4 sub regions in calculation procedure S2 Vector Y_hWith vertical projection vector Y_v；Respectively obtain the first subregion horizontal projection vectorFirst subregion upright projection to AmountSecond subregion horizontal projection vectorSecond subregion vertical projection vector3rd subregion floor projection to Amount3rd subregion vertical projection vector4th subregion horizontal projection vector4th subregion upright projection to AmountWherein, Y_hDimension be L/2, Y_vDimension be N/2；

Second subregion horizontal projection vectorWith vertical projection vectorComputing formula as follows, wherein Y_ijIt it is the i-th row jth The Y-component of row pixel；

$Y_h^2=\left[\begin{array}{c}{Y}_{1h}\\ Y_{2h}\\ Y_{3h}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ Y_{\frac{L}{2}h}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{1j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{2j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{3j}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{\frac{L}{2}j}\end{array}\right];$

$Y_v^2=\left[\begin{array}{c}{Y}_{1v}\\ Y_{2v}\\ Y_{3v}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ Y_{\frac{N}{2}v}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i1}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i2}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i3}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i\frac{L}{2}}\end{array}\right];$

Other 3 regionsCalculating process same, it is only necessary to according to zones of different The value amendment span of i, j；First area is3rd region is 4th region is

For RGB mode image, transfer image to YUV color space from RGB color, according still further to said process, Y is carried out Process；Or directly at rgb space, tri-components of R, G, B are all processed by the process of above-mentioned Y, thus judges form；

S4, relation according to subregion in step S2 Yu left and right view, obtain the horizontal projection vector that left view (11) is corresponding And vertical projection vectorThe horizontal projection vector of right view (12)And vertical projection vector

$Y_h^{11}=Y_h^2\left|\right|Y_h^3;$

$Y_v^{11}=Y_v^2+Y_v^3;$

$Y_h^{12}=Y_h^1\left|\right|Y_h^4;$

$Y_v^{12}=Y_v^1+Y_v^4;$

Wherein,WithDimension is L,WithDimension is N/2；

The concatenation operation of two vectorial V and T of definition is:

$V\left|\right|T=\left[\begin{array}{c}{v}_1\\ v_2\\ v_3\\ \begin{array}{c}.\\ .\\ .\end{array}\\ v_n\\ t_1\\ t_2\\ \begin{array}{c}.\\ .\\ .\end{array}\\ t_m\end{array}\right]$

Wherein, the dimension of vector V is n, and the dimension of vector T is m, and the new vector dimension after connection is n+m；

Equally, the horizontal projection vector that top view (21) is corresponding is obtainedAnd vertical projection vectorThe level of lower view (22) Projection vectorAnd vertical projection vector

$Y_h^{21}=Y_h^2+Y_h^1;$

$Y_v^{21}=Y_v^2\left|\right|Y_v^1;$

$Y_h^{22}=Y_h^3+Y_h^4;$

$Y_v^{22}=Y_v^3\left|\right|Y_v^4;$

S5, the distance calculated between left view (11) and right view (12), top view (21) and lower view (22) each projection vectorWith

$D_h^1=Y_h^{11}-Y_h^{12};$

$D_v^1=Y_v^{11}-Y_v^{12};$

$D_h^2=Y_h^{21}-Y_h^{22};$

$D_v^2=Y_v^{21}-Y_v^{22};$

If S6 left view (11) and right view (12) horizontal projection vector distanceLess than first threshold TH_h, and project to Span fromLess than Second Threshold TH_v, then the first Counter Value adds 1；If top view (21) and lower view (22) floor projection Vector distanceLess than first threshold TH_h, and projection vector distanceLess than Second Threshold TH_v, then the second Counter Value adds 1；

S7, repetition step S2～S6, until step S6 has been performed m time；

If the value of S8 the first enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program form detected for a left side Right 3D video format；

If the value of the second enumerator more than or equal to k*m, k ∈ (0.5,1], then the current video program form detected is upper and lower 3D Video format；

If the value of S9 two above enumerator is all unsatisfactory for condition, then the current video program detected is not left and right 3D video Form, is not upper and lower 3D video format；

Step 2, judge that whether video is row intertexture 3D form LI video or column interleaving 3D form CI video or grid-staggered 3D The 3D form CBS video of form CB video or grid-side by side；Including: with step S2 of step one process, step S3, step S4 class Seemingly, according to the rule of different-format left and right view arrangement, the region carrying out left and right view divides, and the level obtaining left and right view is thrown Shadow and upright projection；The region, left and right of row intertexture 3D form LI divides and divides according to the parity rows of image pixel, and institute is odd The pixel of several rows carries out floor projection/upright projection, obtains the horizontal/vertical projection vector of a view；All even number lines Pixel carries out floor projection/upright projection, obtains the horizontal/vertical projection vector of another view；Subsequent processing steps then with Step S5 to the S8 of step one is identical；

If step 3 step one, step 2 are all not detected by 3D format video, then judge that the current video program detected is 2D video format.

2. a video 3 dimensional format method for quick, it is characterised in that described detection method comprises the steps:

Enumerator initializes；

Obtain a two field picture, according to the n kind 3D video format of required detection, image is carried out the division of n kind left and right view；

Calculate each view to divide, around the horizontal projection vector of the pixel component of view and vertical projection vector；

Calculate each view to divide, the horizontal projection vector distance of corresponding left and right view and vertical projection vector distance；

Divide for each view, if horizontal projection vector distance and vertical projection vector distance are less than given threshold value, The 3D format counter that then division of this view is corresponding adds 1；

Repeated m time, check which kind of 3D format counter value more than k*m, k ∈ (0.5,1], more than then concluding that form is these 3D lattice Formula；If all Counter Values are all unsatisfactory for, then conclude that form is 2D form.

Video 3 dimensional format method for quick the most according to claim 2, it is characterised in that:

Described 3D form is one of them in following 6 kinds of forms: left and right 3D form SBS, upper and lower 3D form TB, row intertexture 3D lattice The 3D form of formula LI, column interleaving 3D form, grid-staggered 3D form, grid-side by side.

Video 3 dimensional format method for quick the most according to claim 2, it is characterised in that:

For the most universal left and right 3D and upper and lower 3D form, described detection method specifically includes following steps:

S1, new video program start, to the first enumerator, the second counter O reset；

S2, obtain a frame width be N highly for the image P of L, input picture is divided into 4 sub regions, formed upper left, lower-left, Upper right, bottom right 4 sub regions；Carry out view division by left and right 3D, upper and lower 3D form more simultaneously；When formatting by left and right 3D, left View (11) is on the left side, and right view (12) is on the right；When formatting by upper and lower 3D, top view (21) on top, lower view (22) Below；

S3, for YUV mode image, the luminance component Y of 4 sub regions in step S2 is calculated, in calculation procedure S2 4 The horizontal projection vector Y of subregion_hWith vertical projection vector Y_v；

For RGB mode image, in calculation procedure S2, tri-components of R, G, B of 4 sub regions calculate, and obtain the water of correspondence Flat projection vector R_h、G_h、B_hWith vertical projection vector R_v、G_v、B_v；

Above-mentioned each component Y, R, G, B are labeled as X, for component X, respectively obtain the first subregion horizontal projection vector First subregion vertical projection vectorSecond subregion horizontal projection vectorSecond subregion vertical projection vector3rd subregion horizontal projection vector3rd subregion vertical projection vector4th subregion horizontal projection vector4th subregion vertical projection vectorWherein, X_hDimension be L/2, X_vDimension be N/2；

S4, relation according to subregion in step S2 Yu left and right view, obtain the horizontal projection vector that left view (11) is corresponding And vertical projection vectorThe horizontal projection vector of right view (12)And vertical projection vectorWherein,With Dimension is L,ProfitDimension is N/2；

Equally, the horizontal projection vector that top view (21) is corresponding is obtainedAnd vertical projection vectorThe water of lower view (22) Flat projection vectorAnd vertical projection vector

S5, the distance calculated between left view (11) and right view (12), top view (21) and lower view (22) each projection vectorWith

If S6 left view (11) and right view (12) horizontal projection vector distanceLess than first threshold TH_h, and project to Span fromLess than Second Threshold TH_v, then the first Counter Value adds 1；If top view (21) and lower view (22) floor projection Vector distanceLess than first threshold TH_h, and projection vector distanceLess than Second Threshold TH_v, then the second Counter Value adds 1；

S7, repetition step S2～S6, until step S6 has been performed m time；

If the value of S8 the first enumerator is more than or equal to setting value, then the current video program form detected is left and right 3D video Form；

If the value of the second enumerator is more than or equal to setting value, then the current video program form detected is upper and lower 3D video lattice Formula；

If the value of two above enumerator is all unsatisfactory for condition, then the current video program detected is 2D video format.

Video 3 dimensional format method for quick the most according to claim 2, it is characterised in that:

In described step S3, the second subregion horizontal projection vectorWith vertical projection vectorComputing formula as follows, wherein Y_ijIt it is the Y-component of the i-th row jth row pixel；

$Y_h^2=\left[\begin{array}{c}{Y}_{1h}\\ Y_{2h}\\ Y_{3h}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ Y_{\frac{L}{2}h}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{1j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{2j}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{3j}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ {\mathrm{\Σ}}_{j=1}^{\frac{N}{2}}{Y}_{\frac{L}{2}j}\end{array}\right];$

$Y_v^2=\left[\begin{array}{c}{Y}_{1v}\\ Y_{2v}\\ Y_{3v}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ Y_{\frac{N}{2}v}\end{array}\right]=\left[\begin{array}{c}{\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i1}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i2}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i3}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ {\mathrm{\Σ}}_{i=1}^{\frac{L}{2}}{Y}_{i\frac{L}{2}}\end{array}\right];$

Other 3 regionsCalculation be similar to therewith.

Video 3 dimensional format method for quick the most according to claim 2, it is characterised in that:

In described step S4,

The concatenation operation of definition two vectorial V, T is:

$V\left|\right|T=\left[\begin{array}{c}{v}_1\\ v_2\\ v_3\\ \begin{array}{c}.\\ .\\ .\end{array}\\ v_n\\ t_1\\ t_2\\ \begin{array}{c}.\\ .\\ .\end{array}\\ t_m\end{array}\right]$

The dimension of V be the dimension of n, T be m, the new vector dimension after connection is n+m.

Video 3 dimensional format method for quick the most according to claim 2, it is characterised in that:

In described step S5,