CN105472308A - Multi-view naked eye 3D video conference system - Google Patents

Abstract

The invention provides a multi-view naked eye 3D video conference system, comprising a server terminal and a client, wherein the server terminal and the client establish communication, the client is provided with at least one camera, a single camera is used for shooting a client user, the client is further provided with an image uploading module and an image display module, and the image display module decodes and converts a multi-view image synthesis encoding result returned by the server terminal to display a 3D naked eye video; and the server terminal is provided with at least three cameras, a plurality of cameras are used for shooting server images, the server terminal is further provided with a camera input module, a face detection module and an image processing module, and the image processing module carries out synthesis encoding on a face of the client user detected by the face detection module and a multi-view image obtained by the camera input module and transmits the same to the client. By adopting the multi-view naked eye 3D video conference system provided by the invention, the usability and the application field of multi-view scenes are greatly expanded, and meanwhile, the user experience and the performance can be improved.

Description

Various visual angles bore hole 3D video conferencing system

Technical field

The present invention relates to video conferencing technology field, particularly one various visual angles bore hole 3D video conferencing system.

Background technology

Video conferencing system is in worldwide extensive use, and remote transmission high definition 3D video can make user obtain more how useful image information, thus meet the further demand such as Machine Design, remote diagnosis, but the series of technical such as the multiple data stream of 3D video conference is synchronous, Video coding, real-time Transmission and bore hole display are not yet fully solved, and the realization of 3D video conferencing system has many technical bottlenecks.

At present, watch 3D video and often need certain special glasses.No matter which prevent the communication between affiliate, because their camera lens, be colored filter, polarisation or liquid crystal display (LCD) shutter, and hide portrait, in the field of nonverbal communication, this is very important.Glasses are not used to be called automatic stereo screen, as the substitute of Nintendo 3DS control desk in addition.Unfortunately, if user change, innovation cost and 3D effect aspect still limited.Such as, through finding the retrieval of prior art document and patent, a kind of method realizing 3D video conference with camera of double vision frequency and the Optical devices of stereo-picture video picture is proposed in the 3D video conference patent of Koninklijke Philips Electronics N.V's application in 2004, but the imaging modality that this patent uses 3D right and left eyes image to switch, must support the use special 3D glasses, therefore system realizes complicated; And this patent defines emphatically 3D collecting device and display device, and the digital coding of 3D rendering, transmission and storage means are not clearly provided.

Summing up, mainly there is following problem in prior art or product:

1. user's face detection is undesirable, and the central point of face can change along with the pixel difference between successive frame.

2., because object is not in real-life situation, the effect reached by being separated layer does not meet reality.

3. final effect is enough untrue.

Therefore study effective 3D rendering coding, transmission, to store and bore hole 3D formation method has great importance for the realization of 3D video conferencing system.In further retrieving, not yet find any document of realizing bore hole 3D video-meeting method similar with the present invention and report.

Summary of the invention

Before this, the technical problem such as multiple data stream synchronous, Video coding, real-time Transmission and bore hole display is not yet fully solved.The object of the present invention is to provide a kind of bore hole 3D video conferencing system that can overcome above-mentioned technical problem.

Technical solution of the present invention provides a kind of various visual angles bore hole 3D video conferencing system, comprises server end and client, sets up communication between server end and client,

Described client arranges at least one camera, adopts single camera to take client user; And image uploading module and image display are set,

Described image uploading module, for taking gained client user image transmitting to server end by client camera;

Described image display, carries out decoding conversion for transmitting the multi-view image composite coding result returned to server end, display bore hole 3D video;

Described server end arranges at least three cameras, adopts multiple camera to take service end image; And shooting input module, face detection module and image processing module are set,

Described shooting input module, the adjustment for back-level server end camera is calibrated, and inputs server end camera shooting gained multi-view image after calibration;

Described face detection module, for detecting the face of client user in client user's image of receiving from client;

Described image processing module, carries out composite coding for the face and shooting input module gained multi-view image detecting gained client user according to face detection module, and is transferred to client.

And, server end arranges three cameras, client adopts the 3D display screen of exempting from glasses based on post mirror, described image display is transmitted after the multi-view image composite coding result returned decodes to server end, obtain the image of the 3 width service end users be separated, according to the viewpoint map number P of 3D display screen support display, conversion obtains viewpoint data, then according to synthesis post mirror format pixel arrangement requirement, show in the mode of bore hole 3D holography.

And the implementation that described conversion obtains viewpoint data is as follows,

I if () P=3, then direct output by 3 width images is viewpoint data;

(ii) if P < 3, then by 3 width images with the 2nd width image for mid point, symmetrical reduce picture number by being uniformly distributed, until P is equal with picture number, now being exported by P width image is viewpoint data;

(iii) if P > 3, then by 3 width images with the 2nd width image for mid point, symmetrical increase picture number by being uniformly distributed, until P is equal with picture number, now being exported by P width image is viewpoint data.

And multiple cameras that described server end is arranged are in same level position.

And described image processing module detects the position of the face of gained client user according to face detection module, the image that each camera of change server end gets synthesizes scale in building-up process.

The invention has the advantages that:

1., without the need to using special glasses, Optical devices or picture screen system, stereo image effect can be experienced easily;

2. the main scope of application: mobile terminal (mobile phone, notebook, panel computer etc.);

3. provide the experience of better 3D, improve 3D rendering quality, improve 3D and watch comfort level;

4. can realize multiple views bore hole 3D to show, solve the problem that many people multi-angle is watched simultaneously;

5. device is simple and cost is lower, is a solution more geared to actual circumstances in bore hole 3D video conference field;

6. the availability of multi views scene and application are widened greatly, and Consumer's Experience and performance can be improved simultaneously;

7. at image transmitting terminal to the 3D vedio data collected, convert to be more suitable for 3D rendering synthesis storage and standardization encoding and decoding; 3D composograph is converted to receive image display converter the bore hole 3D data matched with display device at video receiver, broken away from the constraint of 3D glasses.

Accompanying drawing explanation

Fig. 1 is the 3D video conferencing system schematic diagram of the embodiment of the present invention.

Fig. 2 is the application schematic diagram of the hysteresis mentioned in the embodiment of the present invention.

Fig. 3 is the server end schematic diagram of the embodiment of the present invention.

Fig. 4 is the effect schematic diagram of the 3D video conferencing system of the embodiment of the present invention.

Embodiment

Technical scheme of the present invention is described in detail below in conjunction with drawings and Examples.

The present invention is proposed to be applied to 3 D video meeting multi views computation model by the biologically inspiration of motion parallax.Described motion parallax refers to, the eyes of people are in same plane substantially, but has certain spacing between two, therefore watch object time sight line can form an intersecting angle, angle is larger, third dimension and distance perspective stronger.Due to the existence of intersecting angle, the picture that eyes are seen is not identical, namely creates " parallax ", and the picture that two width have parallax just can obtain complete stereoscopic views after brain process.This living model object is to use neuron to provide the visual cues of perception except stereoscopic vision.Bore hole 3D video conference provided by the invention have employed " sight equation barrier " technology of similar human eye, uses multiple camera thus realizes the acquisition of the image of different angles, and then completing synthesis and the displaying of 3D rendering.

See Fig. 1, the present invention proposes one and uses multi-cam at remote computer (server end) and use the client/server of single camera pursuit movement local user's end (client).Therefore, a kind of various visual angles bore hole 3D video conferencing system of the present invention, comprises server end and client, sets up network service between server end and client.Described client arranges at least one camera, adopts single camera to take client user; Described server end arranges at least three cameras, adopts multiple camera to take client user.Consider 3D effect and equipment cost, suggestion client arranges a camera, and service end arranges three cameras, and embodiment realizes by this.Client can utilize the mobile terminals such as mobile phone to realize, and server can utilize the equipment such as computer to realize.In actual use, no matter be server end or client, the quantity of video camera can increase according to the demand of user and the cost that can bear, and the video camera meeting more than 3 and 3 just can provide the effect of 3D video conference.If such as client arranges three cameras, also can realize the function of server end, support reverse client/server.

Client camera image is for the treatment of head tracking.Server end camera image uses multi-cam to obtain the multi-view image of scene.Various rending model is introduced into a certain specific angle scene and presents.Processing chain of the present invention can comprise: multi-view image is caught, 3D scene display, compressed encoding, transmit, play up and show image.Client uses the computer (or mobile terminal) being provided with single camera, and server end is the computer of the camera using multiple stage arranged in parallel.Client obtains user images, and sends to server by network.Then, this server detects the face of user from the image that client receives, and according to the position of corresponding camera, produces the multi-cam composograph that focuses on user, thus forms the composograph of 3D effect.Repeat this process until video conference terminates.

The display device of embodiment of the present invention client can adopt the equipment such as existing bore hole 3D module.Bore hole 3D module directly can input double vision point signal and solve the problem of content to equipment, compatible universal video signal.Current bore hole 3D module generally comprises display floater, grating and viewpoint translation circuit; Display floater is made up of flat-panel display devices such as LCD, PDP usually, and grating is made up of the device such as slit grating or column mirror grating usually.Display floater comprises multiple pixel, and each pixel is made up of the multiple sub-pixels for showing different colours component; Grating comprises multiple raster unit, is configured in before display floater by grating by certain way, the output of different sub-pixels can be directed to different locus; Viewpoint translation circuit is connected with display floater, and display floater and grating are configured by setting means; The input signal (such as double vision point input signal) received is converted to multiple views composite signal and exports display floater to by viewpoint translation circuit; Viewpoint translation circuit comprises disparity compensation unit and multiple views synthesis unit; Disparity compensation unit calculates the parallax of double vision point input signal, and carries out parallax compensation, exports different points of view image; Multiple views synthesis unit synthesizes different points of view image according to viewpoint matrix, exports multiple views composite signal to described display floater.The principle of bore hole 3D module is the characteristic utilizing people two to have parallax, when without any need for auxiliary equipment (as 3D glasses, the helmet etc.), can obtain the lifelike stereoscopic image with space, the degree of depth.If multi-view image is carried out synthesis according to the dichroism of grating and shows on a display panel, divide picture through grating, the eyes of people can be made to see different visual point images, produce stereo perception.Like this, without the need to additionally adopting special 3D screen, when only needing image to change, adopt the technology such as corresponding encoded, can be cost-saving.

See Fig. 3, the server-side system of the embodiment of the present invention comprises with lower part:

(1) the camera group device for acquisition server end multi-angle image information of server end is located at, this camera group device comprises 3 cameras of server end, catch for multi-view image, to show the 3D image of server-side user to client user, generally comprise speaker and meeting background context, the such as PPT etc. of server end.Can arrange shooting input module during concrete enforcement, for providing corresponding camera to adjust calibration function, and input provides the multi-view image of server end after calibration.

(2) that is located at server end identifies the face detection module of catching with face-image, for determining face position in the user images captured by client for realizing character face.So that the follow-up change according to its face's the unknown thus change the image that server end three cameras get synthesize scale in building-up process.Client is relative set image uploading module then, for client camera being taken gained client user image transmitting to server end.

(3) image processing module of server end is located at, the multi-angle image information of the collection of camera group device described in step (1) is converted in real time to the image of applicable Image Coding and image transmitting, generally comprise compressed encoding, image transmitting, image rendering, 3D rendering synthesis, during concrete enforcement, corresponding units can be embodied as respectively.Client is relative set image display modular converter then, transmits the multi-view image composite coding result returned carry out decoding conversion to server end, display bore hole 3D video; Also after image display conversion, effect better can be reached by playing up.

Be used for, according to 3D rendering synthesis unit gained 3D composograph, carrying out coupling conversion from different display devices in the image display converting unit of server end, enable image include but not limited to bore hole 3D equipment correctly shows.

During concrete enforcement, more multimode can also be set in server-side extensions, such as delay phenomenon process etc.The each module of server end can adopt computer software to realize, or adopts software modularity mode to realize.

Further, being implemented as follows of camera group device in embodiment:

1) described camera group device comprises three acquisition camera, in the face of the user of server end carries out multi-angle image information gathering;

2) method of the camera distribution of described camera group device is to be horizontally disposed with and towards the user being in server end;

3) image information of described camera group device collection can comprise: group shot camera image model, full screen image pattern adapt to the ratio of width to height image model Three models with adjustment.

Group shot camera image model: whole image with the inherent parameters of video camera for benchmark;

Full screen image pattern: image is paved with whole computer picture display box;

Adjustment adapts to the ratio of width to height image model: with applicable user's viewing for starting point suitably adjust picture Aspect Ratio after the pattern that Shows Picture.

Further, in embodiment, the 3 width images in the same moment of shooting head group collection merge by compressed encoding unit, and gained 3D rendering adopts the macroblock arrangement mode being applicable to coding and image storage; Specifically comprise the following steps:

Step one: be 1,2,3 according to camera position adjacent sequential number consecutively by the 3 width images in same moment, such as from left to right or from right to left number consecutively;

Step 2: the macro block 3 width images in same moment being divided into respectively a × a pixel, during concrete enforcement, those skilled in the art can preset the value of a voluntarily, a=16 in embodiment, and namely 3 width images are divided into the macro block of 16x16 pixel respectively; Obtain the 1st width image be set of macroblocks MB11, MB12 ..., MB1k}, the 2nd width image is set of macroblocks { MB21, MB22, ..., MB2k} ... ..., obtaining the 3rd width image is set of macroblocks { MB31, MB32 ..., MB3k}, wherein k=M/ (16 × 16), M are the number of pixels that every width image comprises;

Step 3: the mode that the macro block of 3 width images stores by applicable 3D rendering coding and image is arranged, the set of macroblocks such as 3 width images comprised by MB11, MB21, MB31, MB12, MB22, MB32 ..., the interleaved order of MB1k, MB2k, MB3k} arranges;

The set of macroblocks of 3 width images can be regarded as the two-dimensional array that 3 row k arrange, then the element in this two-dimensional array is staggered and form 3D rendering form.I.e. following two-dimensional array

MB11MB12......MB1k

MB21MB22......MB2k

MB31MB32......MB3k

Be converted to { MB11, MB21, MB31, MB12, MB22, MB32 ..., MB1k, MB2k, MB3k}, then can carry out coding transmission, such as, H.264 Image Coding be carried out to 3D rendering formatted data, together with the realaudio data of G.729 encoding, by Internet Transmission to client.

What the image after client receives synthesis in embodiment also showed is implemented as follows:

1), after client obtains data, carry out image and H.264 decompress, regain 3D composograph;

2) data macro fractionation is carried out to 3D composograph, regain the image of 3 width service end users of separation;

3) 3 width images are mated according to display device, such as, for bore hole 3D module, namely exempt from the 3D display screen of glasses based on post mirror, provide processing mode as follows,

Step one: detect the 3D display screen equipment of client, obtains the viewpoint map number P of this equipment support display, and general value is 1,2,3,4,5,7,9 etc.;

Step 2: the viewpoint figure according to 3D display screen equipment requirement produces viewpoint data;

I if () P=3, then direct output by 3 width images is viewpoint data;

(ii) if P < 3, then by 3 width graphic sequences with the 2nd width image for mid point (supposing that the 2nd width image is the direct picture of service end user), by being uniformly distributed symmetrical minimizing picture number, until P is equal with picture number, now P width image is exported as viewpoint data.For P=1, format picture data separation obtains 3 width images, and { 1,2,3}, finally exports viewpoint data according to 3D display screen equipment requirement and comprises 1 width image { 2}; As P=2, be a width by the 1st and the 2nd width Images uniting, the 2nd width and the 3rd width Images uniting are a width simultaneously, then are a width by this two width Images uniting;

(iii) if P > 3, then by 3 width graphic sequences with the 2nd width image for mid point (supposing that the 2nd width image is the direct picture of service end user), by being uniformly distributed symmetrical increase picture number, until P is equal with picture number, now P width image is exported as viewpoint data.For P=5, according to 3D display screen equipment requirement, use adjacent two width image spatial domain averaging methods, first the first and second width figure are synthesized a width figure, be set to (1+2)/2, again by second and the 3rd width figure synthesize a width figure, be set to (2+3)/2, the final viewpoint data that export comprise 5 width images { 1, (1+2)/2,2, (2+3)/2,3};

Distinguishingly, as p=4, the first and second width image processing and tracking unit can be become a width, become a width by second with the 3rd width image processing and tracking unit, then this two width image and the first width and the 3rd width image processing and tracking unit.

4), after obtaining respective viewpoints data, according to synthesis post mirror format pixel arrangement requirement, show in the mode of bore hole 3D holography.

For the sake of ease of implementation, according to the embodiment of the present invention a kind of three-dimensional bore hole video conferencing system based on multi-cam equipment is provided, the step realizing various visual angles bore hole 3D transmission of video imaging process is as follows:

1) install camera respectively at client and server end: in client, single camera is used for observing client user's action continuously, such as, adopt model to be the camera of LogitechQuickCamProSphereAF.At server end, three same model cameras adopt consistent resolution, contrast, illumination, contribute to the integrated of gained image and adjustment, and produce consistent final effect.

2) server end camera adjustment.

At server end, camera suggestion be set to all towards the object of observation (server-side user) and lay respectively at distance the object of observation central point be about on the point of 15-20 centimetre.

Before communication, the camera group device calibration of server can be completed in advance, guarantee the image that synthesis client user can identify.Camera adjustment mainly realizes the calibration of set camera by the field of mating its image overlap.During concrete enforcement, can by the communication between server end and client, select server end camera to go forward side by side Row sum-equal matrix by user's long-range adjustment from client computer, such as, adjust space size screen showing image and Aspect Ratio makes it to adapt with camera quantity.When server end and client communication, the image sent is the combination of the image coming from server end.

3) face of client user is detected at server end: server end uses the cvHaarDetectObject function in OpenCV (OpenSourceComputerVisionLibrary), a CvSeq structure is returned from the image being used to identification face, and determine the face position in image thus, thus realize the face detecting client user at server end.During concrete enforcement, the position of the face of a square box mark client user can be had on server end display screen.

4) at server end, according to the pattern set, (group shot camera image model, full screen image pattern adapt to the ratio of width to height image model with adjustment, camera is configured accordingly, then from the visible shared region of three width images that video camera obtains, determine to select to determine main camera according to the position of the face of client user thus the image of synthesis 3D effect.

During concrete enforcement, group shot camera image model can being defaulted as, without the need to adjusting video camera, being configured when changing two other camera pattern into.

5) server end by the 3D rendering of synthesis according to the display device with client in addition coded combination etc. be sent to client after processing.

6) 3 are returned), within the next unit interval, repeat said process (i.e. step 3 ~ 5), until video conference terminates.Realize naked eye three-dimensional video conference, comprise the change according to the unknown in gained image of client user's face recognition point, and change the relevant parameter (such as determining main camera, Images uniting accounting) of the acquisition of server-side user face image and composograph, and obtain the image of the face of the same time point different angles of same user with this, be sent to client decoding display after being encoded by certain sequence by gained image.Only need two ends camera and follow the tracks of the effect that namely head angle can realize naked eye three-dimensional video conference.

The present invention proposes further: the height of final combination image is consistent with picture altitude in single camera, and the peak width that the final width of final combination image is total with the image that three video cameras obtain is consistent.(final combination image is that three images are by forming concrete 3D rendering display result after service end and client process.) top of some part of image and below in single camera, can abandon after the space allocation of final combination image.

The Optimal action between adjustment image, vertical direction is consistent---the change between video camera should be reduced as far as possible.Therefore, for ensureing that the height of final combination image is consistent with picture altitude in single camera, video camera can be placed in advance and be in same level position, in combination picture process, adjustment process be carried out to picture to reduce.

As with reference to passing through using nose manually adjust and combine gained image, some real space regions can be shared by two video cameras.In this region, selected by the position of client user's face in combination gained image and determined main camera.When user from left to right moving-head time, just have selected one of them video camera as main camera.The i.e. process of a camera seizure image, main camera is the video camera that face is corresponding, or is referred to as center camera.When the head of client user moves to side, the aspect ratio of image shows the part of needs pro rata.Such as, client user A and server-side user B bis-people carry out face-to-face meeting negotiation (now by video camera and computer screen abstract in air).When the head of client user offsets as the left side, the screen of server end can be observed and capture this change, and immediately according to offset distance and hardware device with data are set etc. and increase the left side of curstomer-oriented end subscriber definitely video camera accessed by the parameter logistic of image in 3D rendering synthesizes.

The present invention proposes further, and some point applies hysteresis, and when no matter the region, center of user face exceedes these from the left or right, master image will become another.

When continuous print (or closely) frame, when the mid point of center of face exactly captured by adjacent camera between image, picture there will be flicker.Introduce sluggishness and can avoid this phenomenon, thus make only when sluggishness point exceedes a certain amount of of change between video camera.Between camera, switching server solves image flicker problem.When user keeps static, image can be a video camera or other very little time slots, thus quick switching between cameras.Switch between server end different camera gained image and introduce sluggish, namely to detect and previous frame image performs the introducing of sluggishness to correct interference relatively current face, solve scintillation effect problem.In image transmitting of the present invention and processing procedure, when sluggishness occurs, directly abandon the image transmitted in this period, and require that server end resends, to reduce system processing load, improve the video effect that system transfers efficiency and user obtain more smooth, improve Consumer's Experience.

Like this, determine main camera according to each site that the face of user may be in, determine some points by the sluggishness point restriction between video camera central authorities and estimate the width of client image with this pro rata.During combination gained image, some real space regions shared by two video cameras.When selecting main camera in this region, should be determined by the position of the face of client user in combination gained image.So, when user moves his/her head from left to right, just have selected one of them video camera as main camera.For selecting a camera, need each site that the face determining user may be in.First, partly/cross section is determined by the mid point restriction between video camera central authorities.These points can estimate the width of client image pro rata.Some point is applied in hysteresis (with reference to Fig. 2).When no matter exceeding these from the left or right, master image will become another, and this depends on that the center of user face is in certain or another region.

In embodiment, server end adopts a square marks to detect the position of face, may have problems like this, i.e. the unsteadiness of Face datection, and in elected time " adjustment adapts to the ratio of width to height image model ", this problem shows more remarkable in client.The present invention proposes further in Face datection, compares present frame and previous frame to correct this interference.If the antipode frame of four angle sums (Face datection square) lower than the filter set up in advance, then abandons new position, retain a upper position.

During concrete enforcement, the face tracking relevant configuration information of server can be preset: the setting of Show Options, reversion X-axis, startup face tracking, face tracking information displaying, filter and hysteresis size on the screen of server end.Setting reversion X-axis, filter and hysteresis are to solve the instable problem of Face datection.

After adopting above technical scheme:

At server end, user has a complete free space visual angle.If there are many video cameras in server, for its height, the image of synthesis is by non-constant width.

Flicker effect controls and introduces sluggishness in filter, effectively prevent the instability of gained image and fuzzy.

Control right and left eyes and see different images, then utilize the visual imaging theory of human eye, the 3D depth effect of formation.

The viewpoint map number that image imaging effect and receiver equipment that recipient accepts are supported is relevant, and the different shown image of the viewpoint map number that each equipment is supported also can be different.

Further,

Video camera array can be used for record scene and multiple vision signal, enables user check a remote 3D world, and makes scene more lively, accurately.

In group shot camera image model, final image size and single camera image size.The change of the result images that this pattern obtains and between difference more easily discovered.

From the angle of user, virtual box moves according to his/her position, forms the effect of long-distance user in screen-picture.This technique avoids separate picture prospect and this difficult task of background.

The result of experiment shows, participant can the 3D effect that provides of perceive motion parallax and the video conference mode liking this new.See Fig. 4, after client and service end set up network service, client user can see service end 3D video, and the image of oneself also can show by little frame simultaneously, and service end user can see client user, and does 3D displaying.

Above-described embodiment is the preferred implementation of this patent, is not used for limiting practical range of the present invention, does not allly deviate from any improvement that principle of the present invention does, retouching and combination etc., within the protection range all belonging to the claims in the present invention.

Claims (5)

1. a various visual angles bore hole 3D video conferencing system, is characterized in that: comprise server end and client, sets up communication between server end and client,

Described client arranges at least one camera, adopts single camera to take client user; And image uploading module and image display are set,

Described image uploading module, for taking gained client user image transmitting to server end by client camera;

Described image display, carries out decoding conversion for transmitting the multi-view image composite coding result returned to server end, display bore hole 3D video;

Described server end arranges at least three cameras, adopts multiple camera to take service end image; And shooting input module, face detection module and image processing module are set,

Described shooting input module, the adjustment for back-level server end camera is calibrated, and inputs server end camera shooting gained multi-view image after calibration;

Described face detection module, for detecting the face of client user in client user's image of receiving from client;

Described image processing module, carries out composite coding for the face and shooting input module gained multi-view image detecting gained client user according to face detection module, and is transferred to client.

2. various visual angles bore hole 3D video conferencing system according to claim 1, it is characterized in that: server end arranges three cameras, client adopts the 3D display screen of exempting from glasses based on post mirror, described image display is transmitted after the multi-view image composite coding result returned decodes to server end, obtain the image of the 3 width service end users be separated, according to the viewpoint map number P of 3D display screen support display, conversion obtains viewpoint data, then according to synthesis post mirror format pixel arrangement requirement, show in the mode of bore hole 3D holography.

3. various visual angles bore hole 3D video conferencing system according to claim 2, is characterized in that: the implementation that described conversion obtains viewpoint data is as follows,

I if () P=3, then direct output by 3 width images is viewpoint data;

(ii) if P < 3, then by 3 width images with the 2nd width image for mid point, symmetrical reduce picture number by being uniformly distributed, until P is equal with picture number, now being exported by P width image is viewpoint data;

(iii) if P > 3, then by 3 width images with the 2nd width image for mid point, symmetrical increase picture number by being uniformly distributed, until P is equal with picture number, now being exported by P width image is viewpoint data.

4. various visual angles bore hole 3D video conferencing system according to claim 1 or 2 or 3, is characterized in that: multiple cameras that described server end is arranged are in same level position.

5. various visual angles bore hole 3D video conferencing system according to claim 4, it is characterized in that: described image processing module detects the position of the face of gained client user according to face detection module, the image that each camera of change server end gets synthesizes scale in building-up process.