CN106231397A - Main broadcaster's class interaction platform main broadcaster's end method for changing scenes and device, Zhu Boduan - Google Patents

Abstract

The invention discloses a kind of main broadcaster's class interaction platform main broadcaster's end method for changing scenes and device thereof, main broadcaster's end, wherein, a kind of main broadcaster's class interaction platform main broadcaster's end method for changing scenes, including: under current scene, scene list is asked to scene library, receive the scene list that described scene library returns, and from described scene list, select handoff scenario information, described handoff scenario information is asked to described scene library, receive the handoff scenario information that described scene library returns, send to client after the handoff scenario information of main broadcaster's information Yu return is merged, so that client obtains and presents content after handoff scenario, or, main broadcaster's information and handoff scenario information are sent to client, content is presented after handoff scenario to obtain after client merges.Present invention achieves main broadcaster's class interaction platform main broadcaster's end scene handoff functionality, main broadcaster during performance can random handoff scenario, and achieve and all client scenes under main broadcaster's platform switched, improve Consumer's Experience.

Description

Main broadcaster's class interaction platform main broadcaster's end method for changing scenes and device, Zhu Boduan

Technical field

The present invention relates to technical field of media play, particularly relate to main broadcaster's class interaction platform main broadcaster's end field Scape changing method and device, Zhu Boduan.

Background technology

The interaction platform of the live class of main broadcaster is all in 2D plane at present, without third dimension.Current Depending on live scene is typically all according to the living broadcast environment of main broadcaster, unless main broadcaster changes place residing for it And the position in place, such as, main broadcaster is transferred to bedroom by parlor, or is turned by the southern side in parlor Moving to north side ... to change living broadcast environment, otherwise user what the direct broadcasting room of this main broadcaster was seen is Same scene, single dull.

Inventor finds under study for action, cannot be random according to the wish of main broadcaster in currently available technology Handoff scenario, scene content lacks change, and, can't be to all clients under main broadcaster's platform End carries out scene switching.

Summary of the invention

The present invention provides main broadcaster's class interaction platform to broadcast end method for changing scenes and device, Zhu Boduan, Achieving main broadcaster's class interaction platform main broadcaster's end scene handoff functionality, main broadcaster not only may be used during performance With random handoff scenario, and achieve the switching of the scene to clients all under main broadcaster's platform, carry Rise Consumer's Experience.

The embodiment of the present invention provides a kind of main broadcaster's class interaction platform main broadcaster's end method for changing scenes, described Method is applied to main broadcaster's end of main broadcaster's interaction platform, and described main broadcaster's end is connected with scene library, described side Method includes:

Under current scene, ask scene list to scene library；

Receive the scene list that described scene library returns, and from described scene list, select switching field Scape information；

Described handoff scenario information is asked to described scene library；

Receive the handoff scenario information that described scene library returns；

Send to client after the handoff scenario information of main broadcaster's information Yu return is merged, so that Client obtains the content that presents after handoff scenario, or, main broadcaster's information and handoff scenario information are sent out Deliver to client, present content to obtain after client merges after handoff scenario.

Alternatively, described to scene library request scene list, particularly as follows:

Main broadcaster's end sends scene list request information to scene library, in described scene list request information Comprise main broadcaster's end mark.

Alternatively, described to the described scene library described handoff scenario information of request, specifically include:

Main broadcaster's end sends handoff scenario solicited message to scene library, in described handoff scenario solicited message Comprise scene display position, scene display size and scene identity so that described scene library according to Described handoff scenario solicited message returns handoff scenario information.

Alternatively, described method also includes:

Main broadcaster's information is processed,

The described handoff scenario information by main broadcaster's information Yu return sends to client after merging, Particularly as follows: be sent to visitor after being merged with main broadcaster's information after treatment by scene information after switching Family end,

Described main broadcaster's information and handoff scenario information are sent to client, particularly as follows: field will be switched Scape information and main broadcaster's information after treatment send to client, and wherein, described scene information includes The parallax information of scene image.

Alternatively, described method also includes:

Described main broadcaster's information is processed, specifically includes:

Main broadcaster's information is carried out FIG pull handle to obtain pending foreground image；

Described pending foreground image is carried out parallax process, to obtain described pending foreground image Initial parallax information；

Obtain the parallax letter of the reference parallax information of described pending foreground image, handoff scenario image Breath；The parallax information of described handoff scenario image includes: the handoff scenario image parallax free degree of depth；

Obtain the described pending foreground image benchmark parallax when being positioned at the shooting parallax free degree of depth, its Described in shoot the parallax free degree of depth equal to the described handoff scenario image parallax free degree of depth；

According to described with reference to parallax information, the parallax information of described handoff scenario image, described benchmark The parallactic shift amount of pending foreground image described in parallax, described initial parallax acquisition of information；According to Described pending foreground image is adjusted by described parallactic shift amount.

Alternatively, the described FIG pull handle that carries out main broadcaster's information, to obtain pending foreground image, has Body includes:

Obtaining the Background statistic information of main broadcaster's scene image, described Background statistic information includes described master Broadcast the first color area in scene at the background mean value of color space and standard deviation；

According to described background mean value and standard deviation, described main broadcaster's scene image is carried out FIG pull handle, Obtain the pending foreground image information of described main broadcaster's scene image.

Alternatively, described pending foreground image is carried out parallax process, to obtain pending prospect The initial parallax information of image, particularly as follows:

Obtain the first mask figure and the second mask figure of personage in current anchor scene image；

Described first mask figure and described second mask figure are carried out dimension-reduction treatment respectively, obtains and institute State the first row matrix corresponding to the first mask figure and second row corresponding with described second mask figure Matrix；

In described second row matrix, determine corresponding with each sample point in described the first row matrix Similitude；

Utilize the parallax of described sample point and described similitude, determine people in described current scene image The initial parallax information of thing.

Alternatively, the reference parallax of described pending foreground image includes maximum reference parallax and minimum With reference to parallax, the parallax information of described handoff scenario image include the maximum disparity of handoff scenario image, Minimum parallax and parallax free.

Alternatively, described according to described reference parallax information, the parallax letter of described handoff scenario image The parallax of pending foreground image described in breath, described benchmark parallax, described initial parallax acquisition of information Side-play amount specifically includes:

According to described with reference to parallax information, the parallax information of described handoff scenario image, described benchmark Parallax obtains the parallax regulation coefficient of described pending foreground image；

According to described initial parallax, described parallax regulation coefficient, described benchmark parallax and described reference Parallax information determines the parallactic shift amount of described pending foreground image.

Alternatively, described according to described reference parallax information, the parallax letter of described handoff scenario image Breath, described benchmark parallax obtain the parallax regulation coefficient of described pending image and include:

Described initial parallax and described benchmark parallax are compared；

When described initial parallax is more than or equal to described benchmark parallax, described handoff scenario figure will be utilized The maximum disparity of the parallax information of picture is divided by described maximum reference parallax and the difference of described benchmark parallax Result as described parallax regulation coefficient；

When described initial parallax is less than described benchmark parallax, described handoff scenario image will be utilized The minimum parallax of parallax information is made divided by the result of described benchmark parallax and the minimum difference with reference to parallax For described parallax regulation coefficient.

Alternatively, described method also includes: creating three-dimensional scenic, described establishment three-dimensional scenic is concrete Including:

Three-dimensional scenic is synthesized by two-dimension picture,

Directly shooting obtains three-dimensional scenic；Or

Import model and create three-dimensional scenic.

Alternatively, described method also includes: uploaded onto the server by new scene, and server is by new field In scape write scene library, and scene library is updated.

The embodiment of the present invention provides a kind of main broadcaster's class interaction platform main broadcaster's end device for changing scenes, described Device is applied to main broadcaster's end of main broadcaster's interaction platform, and described main broadcaster's end is connected with scene library, described dress Put and include:

First transmitting element, for asking scene list, Yi Jixiang to scene library under current scene Scene library request handoff scenario information,

Receive unit, for receiving the scene list that scene library returns, and receive scene library return Handoff scenario information,

Select unit, for selecting handoff scenario information from described scene list,

Integrated unit, for the handoff scenario information of main broadcaster's information with return is merged, and/ Or,

Second transmitting element, after merging the handoff scenario information of main broadcaster's information Yu return Send to client, or, the handoff scenario information of main broadcaster's information with return is sent to client.

Alternatively, described first transmitting element specifically for:

Send scene list request information to scene library, described scene list request information comprises master Broadcast end mark, and,

Send handoff scenario solicited message to scene library, described handoff scenario solicited message comprises field Scape display position, scene display size and scene identity so that described scene library according to described in cut Carry over scape solicited message returns handoff scenario information.

Alternatively, described device also includes:

Main broadcaster's information process unit, for main broadcaster's information carried out FIG pull handle with obtain pending before Scape image, carries out parallax process to described pending foreground image, to obtain described pending prospect The initial parallax information of image；Obtain the reference parallax information of described pending foreground image, switching The parallax information of scene image；The parallax information of described handoff scenario image includes: handoff scenario figure As the parallax free degree of depth；Obtain the described pending foreground image base when being positioned at the shooting parallax free degree of depth Quasi-parallax, the wherein said shooting parallax free degree of depth is equal to the described handoff scenario image parallax free degree of depth； According to described with reference to parallax information, the parallax information of described handoff scenario image, described benchmark parallax, The parallactic shift amount of pending foreground image described in described initial parallax acquisition of information, and according to described Described pending foreground image is adjusted by parallactic shift amount.

The embodiment of the present invention also provides for a kind of main broadcaster's class interaction platform main broadcaster's end, including:

Central processing unit, under current scene, asks scene list to scene library, receives institute State the scene list that scene library returns, and from described scene list, select handoff scenario information, to Described scene library asks described handoff scenario information, receives the handoff scenario letter that described scene library returns Breath, carries out merging and sending to client by the handoff scenario information of main broadcaster's information Yu return, so that Client obtains the content that presents after handoff scenario, or, main broadcaster's information and handoff scenario information are sent out Deliver to client, present content to obtain after client merges after handoff scenario.

Alternatively, described main broadcaster's end is personal computer.

In embodiments of the present invention, by asking scene information to scene library under current scene, and The handoff scenario information of request is sent to server after main broadcaster's information fusion, it is achieved that main broadcaster's class Interaction platform main broadcaster's end scene handoff functionality, main broadcaster is possible not only to during performance arbitrarily switch field Scape, and achieve the switching of the scene to clients all under main broadcaster's platform, improve Consumer's Experience.

Accompanying drawing explanation

Fig. 1 is embodiment of the present invention main broadcaster's interaction platform structural representation；

Fig. 2 is the signal of embodiment of the present invention main broadcaster's class interaction platform main broadcaster's end method for changing scenes flow process Figure；

Fig. 3 is embodiment of the present invention main broadcaster's class interaction platform main broadcaster's end scene handoff procedure schematic diagram；

Fig. 4 is the flow chart that embodiment of the present invention manager updates scene library；

Fig. 5 is the flow chart that embodiment of the present invention supplier updates scene library；

Fig. 6 is embodiment of the present invention supplier and scene management background framework figure；

Fig. 7 is embodiment of the present invention one main broadcaster's class interaction platform main broadcaster's end device for changing scenes structure Schematic diagram；

Fig. 8 is embodiment of the present invention one main broadcaster's class interaction platform main broadcaster's end structure schematic diagram.

Detailed description of the invention

The present invention is described in detail with embodiment below in conjunction with the accompanying drawings.

If it should be noted that do not conflict, each in the embodiment of the present invention and embodiment is special Levy and can be combined with each other, all within protection scope of the present invention.Although it addition, illustrating at device Figure carries out functional unit division, has shown logical order in flow charts, but in some feelings Under condition, can be with the dividing elements being different from device, or order in flow chart performs shown Or the step described.

Before illustrating the embodiment of the present invention, for being more fully understood that the embodiment of the present invention, to the present invention First it is briefly described.In the embodiment of the present invention, main broadcaster's class interaction platform generally includes client, clothes Business device, Zhu Boduan, scene management backstage and scene library, scene is uploaded to by scene management backstage Server, is write scene library by server.During handoff scenario, can be asked by main broadcaster's end Scene information, and according to the scene information handoff scenario returned.

In embodiments of the present invention, transmit to client to realize main broadcaster's content, need bag Including polytype server to participate in, Fig. 1 is main broadcaster's interaction platform structural representation.Such as Fig. 1 institute Showing, the video flowing of main broadcaster transfers to main broadcaster's end, and such as, PC end, by PC end through the people of software As after FIG pull handle, the video flowing of main broadcaster is transmitted to upload server, through and through upload server Transmission is to transcoding server, source server, Edge Server；Then by Edge Server by video Stream synthesizes with scene, transmission to client, after finally video flowing and scene being merged by client in Now give user.

In all embodiments of the invention, client includes such as personal computer, PDA, flat board The subscriber terminal equipments such as computer, MP4, smart mobile phone, three dimensional display box and intelligent television.

Below in conjunction with the accompanying drawings the embodiment of the present invention is further elaborated.

Fig. 2 is the signal of embodiment of the present invention main broadcaster's class interaction platform main broadcaster's end method for changing scenes flow process Figure, as in figure 2 it is shown, the method for the embodiment of the present invention, comprises the following steps:

S21, under current scene, asks scene list to scene library.

In embodiments of the present invention, under conditions of having shown scene, main broadcaster can be by main Broadcast the scene listing function on end, send the solicited message of scene list, wherein, field to scene library Including main broadcaster's end mark in the solicited message of scape list, main broadcaster's end mark may be used for scene library and returns The recipient of mark scene list when returning scene list.

S22, receives the scene list that described scene library returns, and selects to cut from described scene list Change scene information.

In embodiments of the present invention, scene library receives the scene list request information that main broadcaster's end sends After, scene list request information is processed, and the scene list that main broadcaster's end is asked is returned to Main broadcaster's end.After main broadcaster's termination receives corresponding scene list information, showing, main broadcaster can be aobvious The scene list shown selects a certain scene.

S23, asks described handoff scenario information to described scene library.

In embodiments of the present invention, main broadcaster's end, according to the handoff scenario information selected, is sent out to scene library Send scene request information, scene request information comprises scene display position, scene display size and Scene identity, so that described scene library returns scene information according to described handoff scenario solicited message.

S24, receives the handoff scenario information that described scene library returns.

In embodiments of the present invention, after scene library receives handoff scenario solicited message, will be according to field Scape handover request information selects handoff scenario information to return to main broadcaster's end, if current anchor end has selected Select handoff scenario identification information, then the information obtained from scene library includes that scene shows position, scene The information such as display size, are not repeated to obtain the existing information of scene identity.

S25, sends to client after the handoff scenario information of main broadcaster's information Yu return being merged, So that client obtains the content that presents after handoff scenario, or, main broadcaster's information and handoff scenario are believed Breath sends to client, to present content after obtaining handoff scenario after client merges.

In embodiments of the present invention, after main broadcaster's end receives the handoff scenario information that scene library returns, by The handoff scenario information of main broadcaster's information with return is merged by main broadcaster's end, such as, through main broadcaster's portrait FIG pull handle and the video flowing synthesis content that presents of client, particularly as follows: can be anti-by main broadcaster's end Feedback main broadcaster's distance away from photographic head, selects the scene depth image close with feedback information through server, Main broadcaster's image is blended with scene image, sends to server after generating the content that client presents, It is sent to client by server, so that client obtains the content that presents after handoff scenario, or, Main broadcaster's information and handoff scenario information are sent to client, by client by main broadcaster's information and return Handoff scenario information merge, to obtain after client merges in presenting after handoff scenario Hold.

By asking scene information to scene library under current scene, and by the handoff scenario letter of request Send to client after breath and main broadcaster's information fusion, or main broadcaster's information and handoff scenario information are sent To client, it is achieved that main broadcaster's class interaction platform main broadcaster's end scene handoff functionality, and further, The realization of scene handoff functionality is to complete at main broadcaster's end, alleviates the data processing pressure of other equipment, Especially alleviate the data processing pressure of client, so that user is in random handoff scenario Also can obtain the experience of smooth switching simultaneously, and, in certain embodiments, it is also possible to it is right to realize Under main broadcaster's platform, the scene switching of all clients, improves Consumer's Experience.

It is specifically addressed, such as Fig. 3 below in conjunction with main broadcaster's plateform system structure and above-mentioned changing method It show main broadcaster's class interaction platform main broadcaster's end scene handoff procedure schematic diagram.As it is shown on figure 3, scene The idiographic flow of switching is as follows:

1, main broadcaster is to connected main broadcaster's end, such as PC end, transmitting video-frequency flow, is simultaneously emitted by scene The solicited message of list, including the main broadcaster's end mark etc. sending request；

2, PC end sends the request of scene list to scene library, and is returned to PC end by scene library；

3, PC end receives and shows corresponding scene list information；

4, main broadcaster selects a certain scene in scene list, and scene ID of this scene, scene show The solicited messages such as position, scene display size, and are regarded through the FIG pull handle of main broadcaster's portrait by PC end Frequency stream synthesis, such as, by PC end feed back main broadcaster's distance away from photographic head, through server select with The scene depth image that feedback information is close, blends main broadcaster's image with scene image, then sends out Deliver to upload server；

5, synthetic video flows through upload server, transcoding server, source server, and transmission is to limit Edge server；

6, process and load synthetic video stream through Edge Server；

7, client is judged and sends by dispatch server or Edge Server letter that website should be read Breath；

8, client reads respective edges server info, processes and shows synthetic video stream.

In embodiments of the present invention, described method for changing scenes also includes: at main broadcaster's information Reason, the most described handoff scenario information by main broadcaster's information Yu return sends to client after merging Particularly as follows: send to client after handoff scenario information is merged with main broadcaster's information after treatment End, the described handoff scenario information by main broadcaster's information with return sends to client particularly as follows: will cut Change scene information and main broadcaster's information after treatment sends to client, wherein, described handoff scenario Information includes the parallax information of handoff scenario image.

Wherein, described main broadcaster's information is processed, specifically includes:

Main broadcaster's information is carried out FIG pull handle to obtain pending foreground image；

Described pending foreground image is carried out parallax process, to obtain described pending foreground image Initial parallax information；

Obtain the parallax letter of the reference parallax information of described pending foreground image, handoff scenario image Breath；The parallax information of described handoff scenario image includes: the background parallax free degree of depth；

Obtain the described pending foreground image benchmark parallax when being positioned at the shooting parallax free degree of depth, its Described in shoot the parallax free degree of depth equal to the background parallax free degree of depth of described handoff scenario；

According to described with reference to parallax information, the parallax information of described handoff scenario image, described benchmark The parallactic shift amount of pending foreground image described in parallax, described initial parallax acquisition of information；According to Described pending foreground image is adjusted by described parallactic shift amount.

In embodiments of the present invention, main broadcaster's information is carried out FIG pull handle to obtain pending foreground picture Picture, specifically includes:

1) obtaining the Background statistic information of main broadcaster's scene image, described Background statistic information includes institute State the first color area in main broadcaster's scene at the background mean value of color space and standard deviation；

Specifically, obtain the Background statistic information of main broadcaster's scene image, specifically include: transcoding service Device extracts the first color area in main broadcaster's scene image；Transcoding server determines the first color area Background mean value and standard deviation at color space.

Wherein, transcoding server extracts the first color area in main broadcaster's scene image, specifically includes: Transcoding server determines the first color-values of each pixel in main broadcaster's scene image；Judge the first color Whether color value exceedes default color threshold value；If the first color-values exceedes color threshold value, then transcoding Server determines that the pixel of the first color-values indicator belongs to the first color area, and to the first color Region is extracted.

Wherein, the first color-values meets following formula one:

$g^{\′}=\frac{g-r}{g+1}\·\frac{g-b}{g+1}$ Formula one

Wherein, described r, g, b be specially described pixel in described color space, red, green, The property value of blue channel.

It should be noted that the g operator used in above-mentioned calculating the first color-values is rgb color sky Between the product of difference of Green channel attributes value and other two channel attributes values, or to this product Positive correlation conversion.

Further, after the transcoding server the first color area in extracting main broadcaster's scene image, Determine that the first color area is at the background mean value of color space and standard deviation.In the embodiment of the present invention In, described determine that the first color area, at the background mean value of color space and standard deviation, that is to say First color area is united by transcoding server at background mean value and the standard deviation of color space Meter.

The Main Function of statistic processes is that the distribution of color to the first color area is added up, in order to Subsequent step carries out FIG pull handle.Statistical result is that the first color area is each logical in color space The background mean value in road and standard deviation.Wherein, non-limiting as example, multiple numerical value can be used The calculation statistics background mean value being averaged after Lei Jia；Use existing formula of variance statistical standard Difference.

Concrete for adding up the color space of background mean value and standard deviation in embodiments of the present invention It it is CIE LAB color space.The described CIE LAB colour space is can not the most both based on a kind of color Be indigo plant be again that yellow this is theoretical and set up.That is, single numerical value can be used for describing red green color and yellow blue Color characteristic.When a kind of color is with CIE L*a*b*, L* represents brightness value；A* represent red green and B* represents yellow blue value.

In embodiments of the present invention, green in described first color area specifically main broadcaster's scene image Territory, zone；The value of green of described first color-values specifically pixel；Described color threshold value is the greenest Chromatic threshold value.

It should be noted that transcoding server can be without obtaining in the former rate respectively of main broadcaster's scene image Take background mean value and standard deviation, after main broadcaster's scene image can be zoomed in and out by transcoding server, then Obtain background mean value and standard deviation.After main broadcaster's scene image is zoomed in and out, also can be further Save the amount of calculation of transcoding server.

In embodiments of the present invention, main broadcaster's scene image can be at least one pictures；Or for regarding A two field picture in frequency evidence or multiple image.

2) according to described background mean value and standard deviation, carry out described main broadcaster's scene image scratching figure Process, obtain the foreground image information of described main broadcaster's scene image；

Specifically, after transcoding server obtains background mean value and standard deviation, to main broadcaster's scene image Carry out FIG pull handle, obtain the foreground image information of main broadcaster's scene image.In embodiments of the present invention, The character image letter that the foreground image information that transcoding server obtains is specially in main broadcaster's scene image Breath.

Further, in embodiments of the present invention, during foreground image information includes main broadcaster's scene image The standardization prospect fractional value of pixel；

Main broadcaster's scene image, according to background mean value and standard deviation, is carried out scratching at figure by transcoding server Reason, obtains the foreground image information of main broadcaster's scene image, specifically includes: transcoding server determines master The color distance value broadcast between each pixel and the background mean value of scene image (is implemented in the present invention In example, described color is Gauss color distance value according to distance value)；Transcoding server to color away from Distance values carries out mapping process, obtains standardization prospect fractional value.

Further, transcoding server determines that each pixel of main broadcaster's scene image is equal with background Color distance value between value, specifically includes:

Described color distance value meets following formula two:

$H=c_l{e}^{\frac{l-\stackrel{\‾}{l}}{n\·\mathrm{\δ}l}}+c_a{e}^{\frac{a-\stackrel{\‾}{a}}{n\·\mathrm{\δ}a}}+c_b{e}^{\frac{b-\stackrel{\‾}{b}}{n\·\mathrm{\δ}b}}$ Formula two

Wherein, described c_l,c_a,c_bIt is respectively l, the weighter factor of a, b passage, c_l,c_a,c_b>=0 and c_l+c_a+c_b=1；Described n is fixed coefficient, n=1,2,3；It is described that (l, a are b) that described pixel is in institute State the l in color space, a, b property value；DescribedFor described background mean value；Described (δ_l,δ_a,δ_b) For described standard deviation.

Further, transcoding server carries out mapping process to color distance value, obtains standardization Prospect fractional value, specifically includes:

After transcoding server determines the color distance value of pixel, color distance value is mapped to mapping In interval, described mapping range is [0,1].In embodiments of the present invention, concrete mapping relations Can be Linear Mapping or nonlinear mapping.In one example, illustrate with Linear Mapping, Set color distance value two ends threshold value, i.e. max-thresholds th and minimum threshold tl, it is judged that color distance Whether value exceedes max-thresholds th, and whether less than minimum threshold tl；If color distance value is not Exceed minimum threshold tl, then this color distance value is mapped as 0 by transcoding server；If color distance Value exceedes max-thresholds th, then this color distance value is mapped as 1 by transcoding server；If color Distance value exceedes minimum threshold tl and less than max-thresholds th, then transcoding server by this color away from Distance values is mapped to the interval interior of [0,1].

It should be noted that in embodiments of the present invention, transcoding server determines that color distance value is Color space is carried out, the color space used in this step and step 1) transit code The color space that server is used when obtaining Background statistic information is identical, all uses CIE LAB color Color space.This color space is suitable for the difference stating between color by digitized mode. By the CIE calculated numerical value of LAB color space, its value is consistent with human eye sensation, and calculates Result is the most accurate.

Wherein, in actual applications, also the first color-values can be calculated in CIE LAB color space； Background mean value and standard deviation is added up in rgb color space；Mark is calculated in rgb color space Quasi-ization prospect fractional value.That is to say, in abovementioned steps 1) and step 2) in, by CIE LAB Color space and rgb color space are mutually carried out.It is understood that when color space is mutual When changing, its corresponding computing formula also needs in succession to be adjusted, and does not repeats them here.

In embodiments of the present invention, described pending foreground image is carried out parallax process, to obtain The initial parallax information of pending foreground image particularly as follows:

1) the first mask figure and the second mask figure of personage in current anchor scene image are obtained；

Specifically, in studio (current scene), photographic head (can be binocular camera, or Two common camera of person) gather main broadcaster audiovisual information.The video and audio letter that photographic head will gather Breath transmits to server (such as, transcoding server).

It is understood that main broadcaster is when recording audiovisual information, main broadcaster's background cloth behind is set to Single color curtain, wherein non-limiting as example, curtain color is green curtain or blue curtain, In the embodiment of the present invention, background cloth is set to green curtain, and for follow-up calculating accuracy, main broadcaster wears Non-green clothing.

The view (left view, right view) of camera collection is processed by transcoding server, profit Scratch diagram technology with existing green curtain, main broadcaster's image is plucked out from the view of left and right, obtain current anchor field The first mask figure of personage and the second mask figure in scape image.

In embodiments of the present invention, left view obtains the first mask figure after carrying out FIG pull handle, the right side regards Figure obtains the second mask figure after carrying out FIG pull handle, may also be right view in actual applications and scratch Figure obtains the first mask figure after processing, and left view obtains the second mask figure after carrying out FIG pull handle, This does not limits.

Described green curtain scratches diagram technology specifically, for each pixel in a certain view, according to this The color value of pixel, it is judged that whether this pixel is prospect, and calculates the color of this pixel Value accounts for the percentage ratio of foreground color value, and calculated percentage ratio that is to say the alpha of this pixel Value, by each pixel is calculated alpha value, obtains the alpha mask that a certain view is corresponding Figure.

2) described first mask figure and described second mask figure are carried out dimension-reduction treatment respectively, obtain The first row matrix corresponding with described first mask figure and corresponding with described second mask figure Two row matrixs；

Specifically, transcoding server carries out dimension-reduction treatment to the first mask figure and the second mask figure respectively, Obtain the first row matrix corresponding with the first mask figure and second row corresponding with the second mask figure Matrix.

Further, in one implementation, transcoding server carries out dimension-reduction treatment to mask figure, Obtain the row matrix corresponding with mask figure, specifically include: according to mask figure, transcoding server obtains The mask value (i.e. alpha value) of each pixel in mask figure；Utilize the mask of each pixel Value, transcoding server calculates the mask total value of each column pixel in mask figure, and (i.e. accumulated pixel point is covered Code value) or mask average (i.e. averaging after accumulated pixel point mask value) or mask median Value (i.e. after accumulated pixel point mask value, mask value is ranked up, takes mask intermediate value)；Turn Code server using mask total value or mask average or mask I d median as the respective column of row matrix.

In embodiments of the present invention, transcoding server can be chosen for each column and calculate different values, example As, transcoding server calculates the mask total value of first row pixel；Secondary series calculates covering of pixel Code average；The mask I d median of the 3rd column count pixel, or, the value that every column count is identical, Do not limit at this.

In another kind of implementation, transcoding server carries out dimension-reduction treatment to mask figure, obtain with The row matrix that mask figure is corresponding, specifically includes: according to mask figure, transcoding server obtains mask figure In the mask value (i.e. alpha value) of each pixel；Transcoding server judges to belong in mask figure Whether it is more than mask threshold value with the mask value of string pixel；If mask value is more than mask threshold value, Then transcoding server calculates and is more than the vertical coordinate total value belonging to same string pixel of mask threshold value (i.e. The ordinate value of accumulated pixel point), or calculate more than described mask threshold value belong to same string pixel Point vertical coordinate average (i.e. averaging after the ordinate value of accumulated pixel point), or calculate be more than Vertical coordinate I d median (the i.e. accumulated pixel point belonging to same string pixel of described mask threshold value After ordinate value, ordinate value is ranked up, takes vertical coordinate intermediate value)；Transcoding server will Vertical coordinate total value or vertical coordinate average or vertical coordinate I d median are as the respective column of row matrix.

In embodiments of the present invention, transcoding server can be chosen for each column and calculate different values, example As, transcoding server calculates the vertical coordinate total value of first row pixel；Secondary series calculates pixel Vertical coordinate average；The vertical coordinate I d median of the 3rd column count pixel, or, every column count phase Same value, does not limits at this.

It should be noted that transcoding server carries out dimension-reduction treatment to the first mask figure, obtain first The first row matrix that mask figure is corresponding, or transcoding server carries out dimension-reduction treatment to the second mask figure, Obtaining the second row matrix that the second mask figure is corresponding, all may select in above two mode is any one Kind, do not limit at this.

In embodiments of the present invention, described row matrix is specially the matrix of 1 row n row.

3) in described second row matrix, determine and each sample point pair in described the first row matrix The similitude answered；

Specifically, transcoding server determines sample point in the first row matrix, meanwhile, according to determining Sample point, transcoding server is determined at the similitude that sample point is corresponding in the second row matrix.

Further, in the second row matrix, determine corresponding with each sample point in the first row matrix Similitude, specifically include:

According to default sampling interval pattern or random model, in the first row matrix, determine One sample point；In embodiments of the present invention, the sampling interval is specially the positive integer more than or equal to 0 (n >=0), that is to say that the sampling interval can be at equal intervals every n point 1 sample point of extraction Or unequal interval.After transcoding server determines the first sample point, obtain the abscissa of the first sample point Value.Meanwhile, in the second row matrix, transcoding server determines and the abscissa value of the first sample point Identical primary importance point；Transcoding server calculates the whole pixels in the first sample point contiguous range The affinity score of the whole pixels in the range of the affinity score value of point and primary importance vertex neighborhood Value；Utilize the maximum in the abscissa value of the first sample point and affinity score value, determine and the The first similitude that one sample point is corresponding.

In embodiments of the present invention, can formerly preset territory, described contiguous range specifically refers to Centered by the abscissa of the first sample point, fluctuate in the range of whole pixels.Such as, The abscissa of the first sample point is x, and default contiguous range is upper and lower 20.Namely it is selected at x-20 Whole pixels to x+20.

Further, the whole pixels in transcoding server calculates the first sample point contiguous range Affinity score value and primary importance vertex neighborhood in the range of the affinity score value of whole pixels, Specifically include: transcoding server obtains the whole pixels in the first sample point contiguous range respectively First abscissa value；And the second horizontal stroke of the whole pixels in the range of acquisition primary importance vertex neighborhood Coordinate figure；According to the first abscissa value and the second abscissa value, transcoding server calls similar point Number calculates functions, calculate the affinity score value of whole pixels in the first sample point contiguous range with And the affinity score value of the whole pixels in the range of primary importance vertex neighborhood.

Further, transcoding server utilizes abscissa value and the affinity score of the first sample point Maximum in value, determines first similitude corresponding with the first sample point, particularly as follows:

The abscissa of the first similitude need to meet following formula:

$x^{\′}=x+\arg \underset{i\∈\[p_1,p_2\]}{\max }\left(similarity\right(N_l\left(x\right),N_r\left(x+i\right)\left)\right);$

Wherein, described x is the abscissa value of described first sample point；Described similarity (P, Q) is institute State affinity score and calculate function；Described N₁X () is the whole pictures in described first sample point contiguous range First abscissa value of vegetarian refreshments；Described N_r(x+i) it is whole in the range of described primary importance vertex neighborhood Second abscissa value of pixel；Described p₁,p₂Bound for described contiguous range；Described i is Constant.

Previously described similarity (P, Q) be affinity score calculate function, its act on set P with And set Q.In embodiments of the present invention, set P can be by N₁X () replaces, set Q can be by N_r(x+i) Replace.General, this function can be zero-mean normalization intersection similarity (Zero-mean Normalized Cross-Correlation), the most following formula of specific formula for calculation:

$similarity(P,Q)=\frac{{\mathrm{\Σ}}_i(P_i-\stackrel{\‾}{P})(Q_i-\stackrel{\‾}{Q})}{\sqrt[2]{{\mathrm{\Σ}}_i{(P_i-\stackrel{\‾}{P})}^2{(Q_i-\stackrel{\‾}{Q})}^2}}$

It is understood that be not limited in embodiments of the present invention use affinity score to calculate function Calculate affinity score, it is possible to use antipode sum (Sum of Absolute Differences), Or square differences sum (Sum of Squared Differences), or normalization intersection is similar The functions such as degree (Normalized Cross-Correlation) calculate affinity score.Above-mentioned calculating Formula is well-known formula, in the embodiment of the present invention, is only to use this formula to calculate affinity score value.

4) utilize the parallax of described sample point and described similitude, determine described current anchor scene The initial parallax information of personage in image.

Specifically, according to step 3) sample point and the corresponding similitude that determine, obtain sample point With the abscissa value of similitude, by the difference of the abscissa value of similitude Yu the abscissa value of sample point As the parallax between 2 o'clock, determine the parallax of personage in current anchor scene image by this parallax Scope.Such as, the abscissa value of sample point is x, and the abscissa value of similitude is x', then parallax L'=x'-x.

It is understood that extraction sample point be multiple, accordingly, it is determined that similitude the most equal For multiple.Each parallax in this step, is the corresponding similitude of each sample point Difference, determines the disparity range of personage in current anchor scene image by multiple parallaxes.

Further, utilizing the parallax of sample point and similitude, transcoding server determines current anchor The disparity range of personage in scene image, specifically includes: whether transcoding server first determines whether parallax More than the first parallax threshold value preset；Transcoding server is by regarding more than the first parallax threshold value in parallax Difference is as the first sub-parallax, and judges that whether the first sub-parallax is more than the second parallax threshold value preset； Transcoding server will be more than the first sub-parallax of the second parallax threshold value as second in the first sub-parallax Sub-parallax, and using the maximum disparity of the second sub-parallax as the upper limit of disparity range, the second son is regarded The minimum parallax of difference is as the lower limit of disparity range.

In embodiments of the present invention, the second parallax threshold value is more than the first parallax threshold value；Described first regards Difference limen value is for selecting the similarity of multiple parallaxes；Described second parallax threshold value is used for screening out Go out the discrete error in multiple first sub-parallax.

Wherein, when determining disparity range, also can only judge once, that is to say that transcoding server is sentenced Whether disconnected parallax is more than the second parallax threshold value preset, if parallax is more than the second parallax threshold preset Value, then will be greater than the second parallax threshold value, and the parallax maximum upper limit the most；Will be greater than second to regard Difference limen value, and parallax minima lower limit the most.

Therefore, by above-mentioned parallax calculation method, transcoding server obtains current anchor scene image The first mask figure of middle personage and the second mask figure, enter respectively to the first mask figure and the second mask figure Row dimension-reduction treatment, obtains the first row matrix and the second row matrix；Transcoding server divides in row matrix Do not determine sample point and similitude；Utilize the parallax of sample point and similitude, determine current anchor field The disparity range of personage in scape image.Comparing more existing parallax calculation method, the present invention is in basis When sample point determines similitude, only calculate in row matrix picture in sample point and similitude contiguous range The affinity score value of vegetarian refreshments, and then determine similitude, and no longer the pixel of whole image is carried out Mating after calculating, the computation complexity solving parallax calculation method in prior art is high, and meter again Calculate the problem that accuracy rate is low；Greatly reduce disparity computation complexity, and improve calculating accuracy rate.

It should be noted that in hereinbefore described parallax calculation method, current anchor scene image In can be 1 or multiple.

Preferably, in embodiments of the present invention, if personage is unique in current anchor scene image, In the second row matrix, determine the similitude corresponding with each sample point in the first row matrix, specifically Including:

Transcoding server, respectively in the first row matrix neutralizes the second row matrix, determines the first row matrix The first focus point and the second focus point of the second row matrix.

Utilize the parallax of sample point and similitude, determine the parallax of personage in current anchor scene image Scope, specifically includes:

Transcoding server using the difference of the first focus point and the second focus point as current anchor scene The parallax of personage in image.

It should be noted that the method for the first focus point and the second focus point that determines is that this area is known altogether General knowledge, no longer repeats at this.When the aforementioned parallax utilizing focus point is as personage's parallax, it is possible to should Being used in personage in current anchor scene image is multiple situations, but now needs to ensure that each personage protects Hold in same depth bounds.

In embodiments of the present invention, described pending foreground image includes maximum with reference to regarding with reference to parallax Difference (Dis_{cam_max}) and minimum with reference to parallax (Dis_{cam_min}), the parallax of described handoff scenario image Information includes the maximum disparity (Dis of handoff scenario image_{bg_max}), minimum parallax (Dis_{bg_min})、 Parallax free.

In embodiments of the present invention, according to described with reference to parallax information, described handoff scenario image Pending foreground image described in parallax information, described benchmark parallax, described initial parallax acquisition of information Parallactic shift amount, specifically include:

1) according to described with reference to parallax information, the parallax information of described handoff scenario image, described Benchmark parallax obtains the parallax regulation coefficient of described pending foreground image.

In embodiments of the present invention, in order to obtain the parallax regulation coefficient of described pending foreground image, Need first to obtain the reference parallax information of pending foreground image, the parallax information of handoff scenario image And benchmark parallax.

Specifically, as before, in embodiments of the present invention, described reference parallax information includes maximum ginseng Examine parallax (Dis_{cam_max}) and minimum with reference to parallax (Dis_{cam_min})；Described handoff scenario image Parallax information includes: the handoff scenario image parallax free degree of depth, handoff scenario image maximum disparity (Dis_{bg_max}) and handoff scenario image minimum parallax (Dis_{bg_min}).Wherein, due to handoff scenario Image is the most selected, and the parallax information of the most described handoff scenario image can be by described handoff scenario figure The relevant parameter of picture obtains.Pending foreground image is to be obtained by stingy nomography, and therefore this is initial Parallax (Dis_int) can also be obtained by the result of FIG pull handle, treat place specifically by aforementioned Reason foreground image parallax processes step and obtains.

So, for maximum with reference to parallax and minimum with reference to parallax, can lead in embodiments of the present invention Cross following manner to obtain.

When utilizing stereoscopic camera to shoot, can obtain by reading the configuration information of stereoscopic camera The acquisition parameters of described stereoscopic camera, wherein said acquisition parameters includes: regarding of described stereoscopic camera Wild angle, level interval and horizontal resolution.Assume that subject apart from the distance of stereoscopic camera is D, then, can determine that subject exists according to stereoscopic camera acquisition parameters etc. according to equation below The parallax when degree of depth is d:

${\mathrm{Dis}}_{cam}=\frac{2\·D}{2\·d\·tg\frac{f}{2}+D}\·\mathrm{Res}$

Wherein, Dis_camRepresent subject parallax (unit is pixel), D table when the degree of depth is d Showing that level interval, f represent the angle of visibility of stereoscopic camera, Res represents horizontal resolution.

Set stereoscopic camera ultimate range d apart from the scene at pending foreground image place₁And minimum Distance d₂.So, according to described angle of visibility, described level interval, described horizontal resolution, institute State described in distance maximum and described distance minima (1) according to the following equation and (2) acquisition Restricted publication of international news and commentary entitled parallax and minimum reference parallax.

${\mathrm{Dis}}_{cam\_max}=\frac{2\·D}{2\·d_1\·tg\frac{f}{2}+D}\·\mathrm{Res}---\left(1\right)$

${\mathrm{Dis}}_{cam\_min}=\frac{2\·D}{2\·d_2\·tg\frac{f}{2}+D}\·\mathrm{Res}---\left(2\right)$

Wherein, Dis_{cam_max}Represent maximum with reference to parallax, Dis_{cam_min}Represent minimum with reference to parallax, D table Showing that level interval, f represent the angle of visibility of stereoscopic camera, Res represents horizontal resolution, d₁And d₂Point Biao Shi distance maximum and distance minima.

It follows that obtain described pending foreground image further when being positioned at the shooting parallax free degree of depth Benchmark parallax.

Specifically, acquisition parameters and the shooting parallax free degree of depth according to above-mentioned acquisition are according to following public affairs Formula (3) obtains described benchmark parallax:

${\mathrm{Dis}}_{cam\_0}=\frac{2\·D}{2\·d_0\·tg\frac{f}{2}+D}\·\mathrm{Res}---\left(3\right)$

Wherein, Dis_{cam_0}Representing benchmark parallax, D represents that level interval, f represent stereoscopic camera Angle of visibility, Res represents horizontal resolution, d₀Represent the shooting parallax free degree of depth.

Regard for the shooting parallax free degree of depth of the pending foreground image of guarantee and the zero of handoff scenario image Difference the degree of depth match, need at this by pending foreground image shoot the parallax free degree of depth time parallax, Also zero will be adjusted to by benchmark parallax.

In embodiments of the present invention, parallax regulation coefficient can be obtained by following either type.

Mode one, in a particular application, owing to pending foreground image is likely located at the parallax free degree of depth Before, it is also possible to after being positioned at the parallax free degree of depth, therefore, in this step can be according to different Situation obtains this parallax regulation coefficient in different ways.

Therefore, in the manner, first described initial parallax and described benchmark parallax are compared.

When described initial parallax, more than or equal to described benchmark parallax, (pending foreground image is positioned at zero and regards Before the difference degree of depth) time, the maximum disparity of described handoff scenario image will be utilized divided by described maximum ginseng Examine the result of difference of parallax and described benchmark parallax as described parallax regulation coefficient.

Concrete, (4) calculating parallax regulation coefficient now according to the following equation:

${\mathrm{co}}_{pos}=\frac{{\mathrm{Dis}}_{bg\_max}}{{\mathrm{Dis}}_{cam\_\max }-{\mathrm{Dis}}_{cam\_0}}---\left(4\right)$

When described initial parallax, less than described benchmark parallax, (it is deep that pending foreground image is positioned at parallax free After degree) time, by utilizing the minimum parallax of described handoff scenario image divided by described benchmark parallax and The result of the minimum difference with reference to parallax is as described parallax regulation coefficient.

Concrete, (5) calculating parallax regulation coefficient now according to the following equation:

${\mathrm{co}}_{neg}=\frac{{\mathrm{Dis}}_{bg\_min}}{{\mathrm{Dis}}_{cam\_0}-{\mathrm{Dis}}_{cam\_\min }}---\left(5\right)$

Wherein, co_posRepresent that described initial parallax adjusts more than or equal to parallax during described benchmark parallax Coefficient, co_negRepresent described initial parallax less than described benchmark parallax time parallax regulation coefficient, Dis_{cam_max}Represent maximum with reference to parallax, Dis_{cam_min}Represent minimum with reference to parallax, Dis_{cam_0}Represent base Quasi-parallax, Dis_{bg_max}Represent background image maximum disparity, Dis_{bg_min}Represent that background image is neglected most Difference.

Mode two, by the maximum disparity of described handoff scenario image and described handoff scenario image The difference of little parallax as the first parameter, by described maximum with reference to parallax and described minimum with reference to parallax it Difference is as the second parameter.Then, using the business of described first parameter and described second parameter as described Parallax regulation coefficient.

It is to say, in the case, pending foreground image and the parallax free degree of depth can not be considered Relation, and (6) calculate parallax regulation coefficient the most according to the following equation.

$co=\frac{{\mathrm{Dis}}_{bg\_max}-{\mathrm{Dis}}_{bg\_min}}{{\mathrm{Dis}}_{cam\_max}-{\mathrm{Dis}}_{cam\_min}}---\left(6\right)$

Wherein, co represents parallax regulation coefficient, Dis_{cam_max}Represent maximum with reference to parallax, Dis_{cam_min} Represent minimum with reference to parallax, Dis_{bg_max}Represent background image maximum disparity, Dis_{bg_min}Represent background Image minimum parallax, Dis_{bg_max}-Dis_{bg_min}Represent the first parameter, Dis_{cam_max}-Dis_{cam_min}Represent Second parameter.

Mode one is compared with mode two, owing to considering to wait in a particular application to locate in mode one The physical location of reason foreground image and the relation of the parallax free degree of depth, hence with regarding that mode one calculates Difference regulation coefficient more can make the parallax of pending foreground image and handoff scenario image match, and keeps away Exempt from that the phenomenon that distortion occurs in the image of adjusted rear display occurs, thus improve display effect further.

2) according to described initial parallax, described parallax regulation coefficient, described benchmark parallax and described The parallactic shift amount of described pending foreground image is determined with reference to parallax information.

In this step, including following process:

Step 21), calculate target parallax.Concrete, by described benchmark parallax with described initially regard The difference of difference regards as the target of described pending foreground image with the product of described parallax regulation coefficient Difference.

Concrete, when using the parallax regulation coefficient of aforesaid way one, according to the following equation (7) (8) target parallax is calculated.

When described initial parallax is more than or equal to described benchmark parallax:

Dis_obj=co_pos·(Dis_int-Dis_{cam_0}) (7)

When described initial parallax is less than described benchmark parallax,

Dis_obj=co_neg·(Dis_int-Dis_{cam_0}) (8)

Wherein, Dis_objRepresent target parallax, Dis_intRepresent initial parallax, Dis_{cam_0}Represent that benchmark regards Difference, co_posRepresent described initial parallax more than or equal to described benchmark parallax time parallax regulation coefficient, co_negRepresent that described initial parallax is less than parallax regulation coefficient during described benchmark parallax；

Concrete, when using the parallax regulation coefficient of aforesaid way two, according to the following equation (9) Calculate target parallax.

Dis_obj=co (Dis_int-Dis_{cam_0}) (9)

Wherein, Dis_objRepresent target parallax, Dis_intRepresent initial parallax, Dis_{cam_0}Represent that benchmark regards Difference, co represents parallax regulation coefficient.

Step 22), calculate parallactic shift amount.Concrete, by described initial parallax and described target The difference of parallax is as described parallactic shift amount.

Concrete, when using the parallax regulation coefficient of aforesaid way one, according to equation below (10) Obtain described parallactic shift amount:

$Val={\mathrm{Dis}}_{\mathrm{int}}-{\mathrm{Dis}}_{obj}=\left\{\begin{array}{c}(1-{\mathrm{co}}_{pos}){\mathrm{Dis}}_{\mathrm{int}}+{\mathrm{co}}_{pos}\&CenterDot;{\mathrm{Dis}}_{cam\_0}...{\mathrm{Dis}}_{\mathrm{int}}\&GreaterEqual;{\mathrm{Dis}}_{cam\_0}\\ (1-{\mathrm{co}}_{neg}){\mathrm{Dis}}_{\mathrm{int}}+{\mathrm{co}}_{neg}\&CenterDot;{\mathrm{Dis}}_{cam\_0}...{\mathrm{Dis}}_{\mathrm{int}}<{\mathrm{Dis}}_{cam\_0}\end{array}\right\}---\left(10\right)$

Wherein, Val represents parallactic shift amount, Dis_objRepresent target parallax, Dis_intRepresent and initially regard Difference, Dis_{cam_0}Represent benchmark parallax, co_posRepresent that described initial parallax regards more than or equal to described benchmark Parallax regulation coefficient during difference, co_negRepresent that described initial parallax is less than regarding during described benchmark parallax Difference regulation coefficient, Dis_{cam_0}Represent benchmark parallax.

Concrete, when using the parallax regulation coefficient of aforesaid way two, according to equation below (11) Obtain described parallactic shift amount:

Val=Dis_int-Dis_obj=(1-co) Dis_int+co·Dis_{cam_0} (11)

Wherein, Val represents parallactic shift amount, Dis_objRepresent target parallax, Dis_intRepresent and initially regard Difference, Dis_{cam_0}Representing benchmark parallax, co represents parallax regulation coefficient, Dis_{cam_0}Represent benchmark parallax.

Further, in embodiments of the present invention, according to described parallactic shift amount to described pending Foreground image is adjusted, and specifically includes:

When described pending foreground image is the right camera shooting utilizing described stereoscopic camera, if Parallactic shift amount is more than 0, is put down to the right by described pending foreground image according to described parallactic shift amount Move；If parallactic shift amount is less than 0, according to described parallactic shift amount by described pending foreground picture As to left.

When described pending foreground image is the left camera shooting utilizing described stereoscopic camera, if Parallactic shift amount is more than 0, is put down to the left by described pending foreground image according to described parallactic shift amount Move；If parallactic shift amount is less than 0, according to described parallactic shift amount by described pending foreground picture As to right translation.

By above it can be seen that due to relevant according to handoff scenario image in embodiments of the present invention The parallax of pending foreground image is adjusted by parallax information, hence with the solution of the present invention So that pending foreground image and the disparity correspondence of handoff scenario image, so that the most at last After pending foreground image and the synthesis of handoff scenario image, after the image stereo display after this synthesis The parallax of image is the most reasonable.And the most also by pending foreground picture Picture and handoff scenario image determine as Consideration relative to position relationship during parallax free and regard Difference regulation coefficient so that the technical scheme described by those embodiments more can make pending foreground picture Picture and the disparity correspondence of handoff scenario image, it is to avoid the figure that adjusted and synthesis shows after processing occurs As distortion occurs, thus further improve display effect, also more conform to wanting of stereo display Ask.

Alternatively, in embodiments of the present invention, in main broadcaster's end method for changing scenes, it is also possible to including: Creating three-dimensional scenic, described establishment three-dimensional scenic specifically includes: synthesized three-dimensional scenic by two-dimension picture, Directly shooting obtains three-dimensional scenic；Or import model creation three-dimensional scenic.

Wherein, create model of place and support the conventional 3D softwares such as Maya, 3Dmax, model file Form can be the most common 3D form.The creation method of model of place includes multiple enforcement Example, illustrates the creation method of model of place below as a example by picture, serial picture, model, but not It is confined to this three kinds of methods.

Picture

The model of place using picture to create, has scene and takies that memory space is little, is prone to transmission Advantage.

Numerous owing to creating the picture shooting method of model of place, with left and right image pickup method be the most only Example, uses picture to create the method for model of place including, but not limited to following two:

Double camera shoots

The camera lens of two cameras of adjustment, between left and right away to recommendation, enters with main broadcaster's image for ease of the later stage Row synthesis, herein recommend take main broadcaster 3D video recording time camera lens between left and right away from, i.e. value is 40 millimeters, It is adjusted also dependent on practical situation；

Use two cameras to shoot same object simultaneously, obtain two 2D pictures；

Use the 3D software such as Maya, 3D max by two 2D pictures by during correcting captured Camera move, rotate and focal length is inconsistent etc. that mode synthesizes 3D picture；

Synthesized 3D picture is uploaded onto the server by scene management backstage.

3D camera plug-in unit shoots

Use 3D camera plug-in unit shooting 3D picture；

Captured 3D picture is uploaded onto the server by scene management backstage.

(2) serial picture

Owing to using picture to create model of place, there is scene angle and fix, can not become according to content The shortcomings such as change, it is possible to use serial picture creates model of place.

The establishment model of place flow process of serial picture is as follows:

Use camera one 2D picture of shooting, and record the depth of field；

Repeatedly adjusting camera and subject spacing, shooting picture also records the corresponding depth of field；

Serial picture and the corresponding depth of field are imported the softwares such as Maya, 3D max, synthesizes 3D picture；

Synthesized 3D picture is uploaded onto the server by scene management backstage.

(3) model

Owing to using serial picture to create model of place, there is the shortcomings such as photographic head can not rotate, also Can be by importing model creation model of place.

Model imports and includes but not limited to following three kinds of methods: 3D engine build (include Cocos, The products such as Unity), 3D deapth UI builds, is introduced directly into 3D model.

Alternatively, in embodiments of the present invention, in client scene switching mode or loading method, Can also include: uploaded onto the server by new scene, new scene is write in scene library by server, And scene library is updated.

Wherein, scene library updates and includes multiple implementation method, below with manager's renewal, supplier Illustrate as a example by renewal, but scene library update method is not limited to this.

(1) manager updates

The idiographic flow that manager updates scene library method is as follows:

Manager logs in scene management backstage, chooses the new scene needing to upload, after scene management New scene is uploaded onto the server by platform；

After server process new scene, in write scene library, for functions such as scene switchings；

After scene library has write, more new state is back to server；

More new state is transferred to scene management backstage by server, and is shown to manager.

Manager updates the flow chart of scene library as shown in Figure 4.

(2) supplier updates

The idiographic flow that supplier updates scene library method is as follows:

Supplier logs in scene management backstage, chooses the new scene needing to upload, after scene management New scene is uploaded onto the server by platform；

After server process new scene, in write scene library, for follow-up correlation function；

After scene library has write, more new state is back to server；

More new state is transferred to scene management backstage by server, and is shown to supplier.

Supplier updates the flow chart of scene library as it is shown in figure 5, stand after supplier and scene management Structure relation is as shown in Figure 6.

Fig. 7 is embodiment of the present invention one main broadcaster's class interaction platform main broadcaster's end device for changing scenes structure Schematic diagram, as it is shown in fig. 7, described device includes:

First transmitting element 71, for asking scene list to scene library under current scene, and Handoff scenario information is asked to scene library,

Receiving unit 72, for receiving the scene list that scene library returns, and reception scene library is returned The handoff scenario information returned,

Select unit 73, for selecting handoff scenario information from described scene list,

Integrated unit 74, for the handoff scenario information of main broadcaster's information with return is merged, and / or,

Second transmitting element 75, for sending out after the handoff scenario information fusion of main broadcaster's information and return Deliver to client, or, the handoff scenario information of main broadcaster's information with return is sent to client.

In embodiments of the present invention, the first transmitting element 71 specifically for:

Send scene list request information to scene library, described scene list request information comprises master Broadcast end mark, and,

Send handoff scenario solicited message to scene library, described handoff scenario solicited message comprises field Scape display position, scene display size and scene identity so that described scene library according to described in cut Carry over scape solicited message returns handoff scenario information.

It should be noted that in embodiments of the present invention, the first transmitting element 71 and second sends Unit 75 is for the convenience on describing, it practice, the first transmitting element 71 and the second transmitting element 75 may be collectively referred to as transmitting element, may be used for realizing the first transmitting element 71 and second and send list The function of unit 75.

In embodiments of the present invention, described device, also include:

Main broadcaster's information process unit 76, pending to obtain for main broadcaster's information being carried out FIG pull handle Foreground image, carries out parallax process to described pending foreground image, with obtain described pending before The initial parallax information of scape image；Obtain the reference parallax information of described pending foreground image, cut Change the parallax information of scene image；The parallax information of described handoff scenario image includes: background zero regards The difference degree of depth；Obtain the described pending foreground image benchmark parallax when being positioned at the shooting parallax free degree of depth, The wherein said shooting parallax free degree of depth is equal to the background parallax free degree of depth of described handoff scenario；According to institute State with reference to parallax information, the parallax information of described handoff scenario image, described benchmark parallax, described The parallactic shift amount of pending foreground image described in initial parallax acquisition of information, and according to described parallax Described pending foreground image is adjusted by side-play amount.

It should be noted that main broadcaster's class interaction platform client scene switching dress in the embodiment of the present invention Put the contents such as mutual, the execution process of the information between unit, due to real with the inventive method Executing example based on same design, the particular content in embodiment of the method is equally applicable.The embodiment of the present invention In unit can realize as single hardware or software, and can make as required The combination of the function of unit is realized with single hardware or software.

In embodiments of the present invention, by asking scene information to scene library under current scene, and The handoff scenario information of request is sent to client after main broadcaster's information fusion, or by main broadcaster's information Send to client with the handoff scenario information returned, it is achieved that main broadcaster's class interaction platform main broadcaster's end field Scape handoff functionality, and further, the realization of scene handoff functionality is to complete at main broadcaster's end, alleviates The data processing pressure of other equipment, especially alleviates the data processing pressure of client, can So that user also can obtain the experience of smooth switching while random handoff scenario.And one A little embodiments can also realize the switching of the scene to clients all under main broadcaster's platform, improve use Family is experienced.

Fig. 8 is embodiment of the present invention one main broadcaster's class interaction platform main broadcaster's end structure schematic diagram, including: Central processing unit 81, wherein:

Central processing unit 81, under current scene, asks scene list to scene library, receives The scene list that described scene library returns, and from described scene list, select handoff scenario information, Ask described handoff scenario information to described scene library, receive the handoff scenario that described scene library returns Information, is carried out the handoff scenario information of main broadcaster's information Yu return merging and send to client, with Make the content that presents after client acquisition handoff scenario, or, by main broadcaster's information and handoff scenario information Send to client, to present content after obtaining handoff scenario after client merges.

In the embodiment of the present invention, the aforesaid operations performed by processor can be with the form of one section of program This section of program in memory, when needs perform aforesaid operations, is called in processor and is performed by storage Process.

In embodiments of the present invention, main broadcaster's end is personal computer.

In embodiments of the present invention, by asking scene information to scene library under current scene, and The handoff scenario information of request is sent to client after main broadcaster's information fusion, or by main broadcaster's information Send to client with the handoff scenario information returned, and further, the reality of scene handoff functionality It is now to complete at main broadcaster's end, alleviates the data processing pressure of other equipment, especially alleviate visitor The data processing pressure of family end, so that user also can obtain stream while random handoff scenario The experience of smooth switching.And can also realize in certain embodiments clients all under main broadcaster's platform The scene switching of end, improves Consumer's Experience.

The foregoing is only embodiments of the present invention, not thereby limit the scope of the claims of the present invention, Every equivalent structure utilizing description of the invention and accompanying drawing content to be made or equivalence flow process conversion, or Directly or indirectly being used in other relevant technical fields, the patent being the most in like manner included in the present invention is protected In the range of protecting.

Claims (14)

1. main broadcaster's class interaction platform main broadcaster's end method for changing scenes, it is characterised in that described Method is applied to main broadcaster's end of main broadcaster's interaction platform, and described main broadcaster's end is connected with scene library, described side Method includes:

Under current scene, ask scene list to scene library；

Receive the scene list that described scene library returns, and from described scene list, select switching field Scape information；

Described handoff scenario information is asked to described scene library；

Receive the handoff scenario information that described scene library returns；

Send to client after the handoff scenario information of main broadcaster's information Yu return is merged, so that Client obtains the content that presents after handoff scenario, or, main broadcaster's information and handoff scenario information are sent out Deliver to client, present content to obtain after client merges after handoff scenario.

Method the most according to claim 1, it is characterised in that described to scene library request Scene list, particularly as follows:

Main broadcaster's end sends scene list request information to scene library, in described scene list request information Comprise main broadcaster's end mark.

3. according to the method described in described claim 2, it is characterised in that described to described field Jing Ku asks described handoff scenario information, specifically includes:

Main broadcaster's end sends handoff scenario solicited message to scene library, in described handoff scenario solicited message Comprise scene display position, scene display size and scene identity so that described scene library according to Described handoff scenario solicited message returns handoff scenario information.

4. according to the method described in any one of claim 1-3, it is characterised in that also include:

Main broadcaster's information is processed,

The described handoff scenario information by main broadcaster's information Yu return sends to client after merging, Particularly as follows: be sent to visitor after being merged with main broadcaster's information after treatment by scene information after switching Family end,

Described main broadcaster's information and handoff scenario information are sent to client, particularly as follows: field will be switched Scape information and main broadcaster's information after treatment send to client, and wherein, described scene information includes The parallax information of scene image.

Method the most according to claim 4, it is characterised in that described method also includes:

Described main broadcaster's information is processed, specifically includes:

Main broadcaster's information is carried out FIG pull handle to obtain pending foreground image；

Described pending foreground image is carried out parallax process, to obtain described pending foreground image Initial parallax information；

Obtain the parallax letter of the reference parallax information of described pending foreground image, handoff scenario image Breath；The parallax information of described handoff scenario image includes: the handoff scenario image parallax free degree of depth；

Obtain the described pending foreground image benchmark parallax when being positioned at the shooting parallax free degree of depth, its Described in shoot the parallax free degree of depth equal to the described handoff scenario image parallax free degree of depth；

According to described with reference to parallax information, the parallax information of described handoff scenario image, described benchmark The parallactic shift amount of pending foreground image described in parallax, described initial parallax acquisition of information；According to Described pending foreground image is adjusted by described parallactic shift amount.

Method the most according to claim 5, described main broadcaster's information is carried out FIG pull handle with Obtain pending foreground image, specifically include:

Obtaining the Background statistic information of main broadcaster's scene image, described Background statistic information includes described master Broadcast the first color area in scene at the background mean value of color space and standard deviation；

According to described background mean value and standard deviation, described main broadcaster's scene image is carried out FIG pull handle, Obtain the pending foreground image information of described main broadcaster's scene image.

Method the most according to claim 5, it is characterised in that described to pending prospect Image carries out parallax process, to obtain the initial parallax information of pending foreground image, particularly as follows:

Obtain the first mask figure and the second mask figure of personage in current anchor scene image；

Described first mask figure and described second mask figure are carried out dimension-reduction treatment respectively, obtains and institute State the first row matrix corresponding to the first mask figure and second row corresponding with described second mask figure Matrix；

In described second row matrix, determine corresponding with each sample point in described the first row matrix Similitude；

Utilize the parallax of described sample point and described similitude, determine people in described current scene image The initial parallax information of thing.

Method the most according to claim 5, it is characterised in that described pending foreground picture The reference parallax of picture includes maximum with reference to parallax and minimum with reference to parallax, described handoff scenario image Parallax information includes the maximum disparity of handoff scenario image, minimum parallax and parallax free.

Method the most according to claim 8, it is characterised in that described according to described reference Parallax information, the parallax information of described handoff scenario image, described benchmark parallax, described initially regard Described in difference acquisition of information, the parallactic shift amount of pending foreground image specifically includes:

According to described with reference to parallax information, the parallax information of described handoff scenario image, described benchmark Parallax obtains the parallax regulation coefficient of described pending foreground image；

According to described initial parallax, described parallax regulation coefficient, described benchmark parallax and described reference Parallax information determines the parallactic shift amount of described pending foreground image.

Method the most according to claim 9, it is characterised in that described according to described reference Parallax information, the parallax information of described handoff scenario image, described benchmark parallax are waited to locate described in obtaining The parallax regulation coefficient of reason image includes:

Described initial parallax and described benchmark parallax are compared；

When described initial parallax is more than or equal to described benchmark parallax, described handoff scenario figure will be utilized The maximum disparity of the parallax information of picture is divided by described maximum reference parallax and the difference of described benchmark parallax Result as described parallax regulation coefficient；

When described initial parallax is less than described benchmark parallax, described handoff scenario image will be utilized The minimum parallax of parallax information is made divided by the result of described benchmark parallax and the minimum difference with reference to parallax For described parallax regulation coefficient.

11. methods according to claim 10, it is characterised in that also include: create three Dimension scene, described establishment three-dimensional scenic specifically includes:

Three-dimensional scenic is synthesized by two-dimension picture,

Directly shooting obtains three-dimensional scenic；Or

Import model and create three-dimensional scenic.

12. methods according to claim 11, it is characterised in that also include: by new field Scape uploads onto the server, and new scene is write in scene library by server, and is updated scene library.

13. 1 kinds of main broadcaster's class interaction platform main broadcaster's end device for changing scenes, it is characterised in that described Device is applied to main broadcaster's end of main broadcaster's interaction platform, and described main broadcaster's end is connected with scene library, described dress Put and include:

First transmitting element, for asking scene list, Yi Jixiang to scene library under current scene Scene library request handoff scenario information,

Receive unit, for receiving the scene list that scene library returns, and receive scene library return Handoff scenario information,

Select unit, for selecting handoff scenario information from described scene list,

Integrated unit, for the handoff scenario information of main broadcaster's information with return is merged, and/ Or,

Second transmitting element, after merging the handoff scenario information of main broadcaster's information Yu return Send to client, or, the handoff scenario information of main broadcaster's information with return is sent to client.

14. 1 kinds of main broadcaster's class interaction platform main broadcaster's ends, it is characterised in that including:

Central processing unit, under current scene, asks scene list to scene library, receives institute State the scene list that scene library returns, and from described scene list, select handoff scenario information, to Described scene library asks described handoff scenario information, receives the handoff scenario letter that described scene library returns Breath, carries out merging and sending to client by the handoff scenario information of main broadcaster's information Yu return, so that Client obtains the content that presents after handoff scenario, or, main broadcaster's information and handoff scenario information are sent out Deliver to client, present content to obtain after client merges after handoff scenario.