CN104301706A - Synthetic method for improving naked eye stereoscopic display effect - Google Patents

Synthetic method for improving naked eye stereoscopic display effect Download PDF

Info

Publication number
CN104301706A
CN104301706A CN201410533537.4A CN201410533537A CN104301706A CN 104301706 A CN104301706 A CN 104301706A CN 201410533537 A CN201410533537 A CN 201410533537A CN 104301706 A CN104301706 A CN 104301706A
Authority
CN
China
Prior art keywords
parallax
display effect
synthetic method
image
information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201410533537.4A
Other languages
Chinese (zh)
Other versions
CN104301706B (en
Inventor
曾超
廖巍巍
梁园
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Kunshan Hui Ren Corelle Intelligent Technology Co., Ltd.
Original Assignee
CHENGDU SEEFELD SCIENCE & TECHNOLOGY Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by CHENGDU SEEFELD SCIENCE & TECHNOLOGY Co Ltd filed Critical CHENGDU SEEFELD SCIENCE & TECHNOLOGY Co Ltd
Priority to CN201410533537.4A priority Critical patent/CN104301706B/en
Publication of CN104301706A publication Critical patent/CN104301706A/en
Application granted granted Critical
Publication of CN104301706B publication Critical patent/CN104301706B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Processing Or Creating Images (AREA)
  • Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)

Abstract

The invention relates to a synthetic method for improving the naked eye stereoscopic display effect. The synthetic method includes the steps that a, different depth information obtaining operations are carried out on different input data; b, if the input data are binocular data, an energy function is built through feature matching to obtain the optimal parallax; c, if the input data are images plus depth information, parallax information is extracted, filtered and de-noised to obtain the optimal parallax; d, the depth information is segmented, classified and marked as different areas; e, the optimal parallax of each area is pre-estimated again; f, the parallax deviation parameter is adjusted to obtain a proper depth space; g, a multi-view image is synthesized according to drawing rules and output to a naked eye television. The synthetic method for improving the naked eye stereoscopic display effect can effectively resolve the display problem that the naked eye television is poor in stereoscopic display effect, effectively synthesize a multi-view image high in depth of field and out-of-screen sense under various resolution ratios, resolve the problem of high calculation loads when synthesizing multiple view points, and relieve the problem that invalid view points are calculated repeatedly in the process of calculating the multi-view image.

Description

A kind of synthetic method strengthening bore hole stereo display effect
Technical field
The present invention relates to disparity computation, Iamge Segmentation and multiple views synthesis technical field, particularly relate to a kind of synthetic method strengthening bore hole stereo display effect.
Background technology
Along with the development of television industries correlation technique, stereoscopic TV progresses into the visual field of people, the realization of stereoscopic TV will be the great revolution in another field after Digital Television, realizing completely of it will be the major progress from shooting to video picture, and stereoscopic TV is the over all Integration system be based upon completely in digital technology, computer technology and network technology.
Stereoscopic TV is a kind of television system can simulating the real space relationship of actual scenery.Also known as three-dimensional television (Threedimensionaltelevision, 3DTV).Three-dimensional television image can not only give spectators a kind of depth preception, and can be stretched in the sensation seeming within reach outside screen to a kind of scenery of spectators.The difference of stereoscopic TV and existing common TV is, existing TV only transmits the two-dimensional image information of a viewpoint, and three-dimensional transmission of television is the image information of multiple views.People watch the image of an object simultaneously with two eyes when watching object, the object of different depth forms the parallax difference of image in two, left and right, brain then melts picture according to the difference of these parallaxes and goes out the different degree of depth and distance, thus produces three-dimensional three-dimensional sense organ.
Bore hole 3D TV is a kind of leading-edge products of stereoscopic TV, be different from traditional hyperphoria with fixed eyeballs mirror stereoscopic TV, bore hole 3D TV utilizes people two to have the characteristic of parallax, when without any need for auxiliary equipment (as 3D eyes, the helmet etc.), can obtain the lifelike stereoscopic image with space and the degree of depth, be three-dimensional stereo display technique truly.But the general stereoeffect of existing bore hole 3D TV is not strong, and particularly at high resolutions, spectators' stereoeffect sense organ is not good.
Summary of the invention
The object of the invention is to overcome the deficiencies in the prior art, a kind of synthetic method strengthening bore hole stereo display effect is provided.The method significantly can promote the stereo display effect of bore hole TV, and can both bring better stereo image effect to spectators under various resolution.
Technical scheme of the present invention is: a kind of synthetic method strengthening bore hole stereo display effect, and comprising parallax acquiring unit, deep space synthesis unit, multi-view image synthesis unit, method flow diagram is shown in Fig. 1.Wherein parallax acquiring unit can receive two kinds of data formats: binocular image format and " video+depth " (video+depth, VD) form.
Described parallax acquiring unit can obtain corresponding parallax information respectively according to different input picture forms.Parallax acquiring unit comprises binocular parallax acquiring unit and VD parallax acquiring unit.
Described binocular parallax acquiring unit: the view data according to input carries out component, is divided into left view and right view, and calculates left and right view in conjunction with Stereo matching, obtain corresponding parallax information.
Described VD parallax acquiring unit: the view data according to input carries out component, isolates the depth information in initial data, obtains corresponding parallax information in conjunction with preliminary treatment.
Described deep space synthesis unit: reconstruct the deep space corresponding to view data, marks off different regions according to Iamge Segmentation, repartitions the depth information of zones of different and is normalized.Spectators can oneself regulation depth space, is subject to the third dimension reaching the better spectators of meeting.
Described multi-view image synthesis unit: according to virtual viewpoint rendering method synthesis multi-view image.
According to technical scheme provided by the invention, the synthetic method of described enhancing bore hole stereo display effect comprises the steps: 1) parallax acquiring unit acquisition initial data image information, select corresponding unit for different pattern of the inputs.If input data are binocular images, then carry out binocular parallax acquiring unit, if video+depth during input data, then carry out VD parallax acquiring unit.
2) concrete, binocular parallax acquiring unit is separated left and right road view ImgL and ImgR, after to ImgL and ImgR filtering convergent-divergent in the disparity range arranged [-α, α] calculate Matching power flow and obtain profile gradients and color gradient respectively, α is maximum permission parallax, wherein when pixel parallax angle , set up energy function according to color gradient, , setting up energy function according to profile gradients, obtaining optimum parallax information by calculating minimum global energy function .
3) concrete, VD parallax acquiring unit extracts the depth information that initial data itself comprises, and needing the depth information to extracting to carry out further filtering and noise reduction, having obtained depth information accurately at this unit.According to meeting virtual view synthesis specification, depth information is converted to the parallax information of original image .
4) further, described deep space synthesis unit processes the parallax information that parallax acquiring unit obtains.In this unit, need parallax information to be converted to spatial depth information , map initial data.
5) concrete, at conversion level spatial depth information in, need to carry out edge image segmentation to parallax information, determine the best disparity plane parameter ω in each block comminute region.For needing the deep space generating N viewpoint, , wherein scale(N) and be the relativity shift coefficient of N number of parallax, spectators can by adjustment scale(N) obtain better bore hole 3D impression.
6) further, multi-view image synthesis unit synthesizes final multi-view image according to deep space according to virtual view composite formula and outputs on bore hole tv display screen.
Technical problem to be solved by this invention: a kind of synthetic method strengthening bore hole stereo display effect is provided.The method effectively can solve the not good display problem of bore hole TV stereoeffect.And effectively can both synthesize the depth of field and go out all very strong multi-view image of screen sense under various resolution.Solve the problem that the multiple viewpoint amount of calculation of synthesis is huge simultaneously, decrease the problem calculating the invalid viewpoint of double counting in multi-view image.
Accompanying drawing explanation
Below in conjunction with accompanying drawing and object lesson, the present invention is described in further detail.
Accompanying drawing 1 is the flow chart of the synthetic method strengthening bore hole stereo display effect.
Accompanying drawing 2 is the system block diagram of binocular parallax acquiring unit.
Accompanying drawing 3 is the system block diagram of VD parallax acquiring unit.
Accompanying drawing 4 is the system block diagram of deep space synthesis unit.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in further detail:
This method comprises 101 parallax acquiring units as shown in Figure 1,102 deep space synthesis units and 103 multi-view image synthesis units.
In Fig. 2,201 modules receive binocular image data, binocular image are divided into left and right two-way image ImgL and ImgR.Left and right two-way image will keep resolution consistent, needs to carry out image scaling to ImgL and ImgR as inconsistent, the same to ensure left images resolution sizes.Figure 20 2 modular character coupling will carry out characteristic matching detection to left and right two-way image, permission the parallax [-α arranged, α] interior execution SIFT algorithm (Scale Invariant Feature Transform, Scale invariant features transform), generate the characteristic vector based on SIFT.In characteristic matching module by setting up the metric space information to image of gaussian pyramid , , wherein: Img(x, y) and be the point on image ImgL or ImgR, be to Img (x, y) gaussian filtering after information, n is the number of plies of gaussian pyramid, and σ is the metric space factor, and σ value is more large more can the general picture feature of representative image, and σ is less more can the minutia of representative image.Establish metric space after, utilize the difference of on same single order two adjacent two-layer metric space functions to obtain DOG(Difference Of Gaussian) one deck of gaussian pyramid. , wherein k is the dimension scale in gaussian pyramid between two adjacent layers, and preferred this method is set to 0.5.The set f that one is estimated parallax β is obtained, if in the module 203 by the Local Extremum detecting DOG space , set up energy function according to color gradient, , set up energy function according to profile gradients.Based on the energy letter of parallax , wherein, p and q is the position of pixel, and N represents the p in image, q coordinate relation, and V weighs the correlation function of inconsistent degree, D is the function weighing f and correct parallax error in classification.In Figure 20 4 color gradient module, V is by the R of left and right two-way image, and G, B pixel difference information obtains.In Figure 20 5 profile gradients module, V is obtained by the shade of gray information gap of left and right two-way image.In Figure 20 6 module, E(f) minimum global energy distribution is exactly the parallax distribution of original image optimum.
The view data of 301 module receiver, video+depth format in Fig. 3,302 modules extract the depth information of VD formatted data.Owing to can produce various noise such as blocking effect etc. in video coding process, so 303 modules need to carry out filtering and noise reduction to original depth information, this method adopts windowing medium filtering, the situation of effective place to go and gaussian noise mixing.By the depth information smoothing processing after denoising in 304 modules, retain complete depth profile and obtain optimum parallax information.
Deep space synthesis unit 102 needs to obtain deep space information aggregate to optimum parallax process .First split optimum parallax, segmentation adopts the method for edge segmentation, marks off different parallax regions by the edge detecting parallax information.The parallax value in each region is sorted, selects intermediate value as best disparity plane parameter ω, the best disparity plane parameter ω of each cut zone is estimated again to parallaxes all in cut zone.For needing the deep space generating N viewpoint, , wherein Zone is each region sequence number, and scale (N) is the relativity shift coefficient of N number of parallax.In Fig. 4, a-quadrant and B region are the depth areas after two different segmentations, and the best disparity plane of deep space in a-quadrant and B region is respectively DV (a) and DV (b), the space length in a-quadrant and B region .By changing scale(N), the space length in a-quadrant and B region can proportionally change.In Fig. 4 A and B Dis can the final original image Scene of effective influence far and near, user is by changing scale(N), suitable parameters can be selected to arrive better stereoeffect.
Multi-view image synthesis unit 103, will with original image according to multiple views composition rule synthesis multi-view image, and synthetic multi-view image is outputted on bore hole TV, the video image of the stereoeffect that presents to audience.

Claims (6)

1. strengthen a synthetic method for bore hole stereo display effect, its feature specifically comprises:
A. obtain initial data image information, select corresponding unit for different pattern of the inputs;
If b. inputting data is binocular images, then select binocular parallax acquiring unit, binocular image is divided into left and right two-way image and performs in permission parallax and characteristic matching detection is carried out to left and right two-way image, build energy function by color gradient or profile gradients again, obtain the distribution of minimum global energy;
If c. inputting data is video+depth, then selects VD parallax acquiring unit, extract the depth information of VD formatted data and filtering and noise reduction and smoothing processing are carried out to original depth information;
D. parallax information is split, and each cut zone is marked;
E. the parallax value in each region is sorted, to the best disparity plane parameter of each cut zone parallaxes all in cut zone are estimated again;
F. for needing the deep space generating N viewpoint, the relativity shift coefficient effective influence final original image Scene distance of adjustment parallax generates suitable deep space;
G. deep space information and original image are synthesized multi-view image according to multiple views composition rule, and synthetic multi-view image is outputted on bore hole TV.
2. a kind of synthetic method strengthening bore hole stereo display effect as claimed in claim 1, it is characterized in that: in step b, characteristic matching detects and adopts SIFT algorithm, maximum permission parallax is set, the metric space factor, the number of plies of gaussian pyramid, build energy function threshold values, color gradient is by the R of left and right two-way image, and G, B pixel difference information obtains, in profile gradients module, the shade of gray information gap of left and right two-way image obtains.
3. a kind of synthetic method strengthening bore hole stereo display effect as claimed in claim 1, is characterized in that: in step c, filtering and noise reduction adopts windowing medium filtering, arranges Filtering Template and frame size.
4. a kind of synthetic method strengthening bore hole stereo display effect as claimed in claim 1, is characterized in that: in steps d, segmentation adopts the method for edge segmentation, marks off different parallax regions by the edge detecting parallax information.
5. a kind of synthetic method strengthening bore hole stereo display effect as claimed in claim 1, is characterized in that: in step e, parallax is estimated again is select intermediate value as best disparity plane parameter .
6. a kind of synthetic method strengthening bore hole stereo display effect as claimed in claim 1, is characterized in that: the deep space composite formula of corresponding N number of viewpoint in step f , by arranging scale (N), regulate the space length between each region.
CN201410533537.4A 2014-10-11 2014-10-11 A kind of synthetic method for strengthening bore hole stereoscopic display effect Active CN104301706B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410533537.4A CN104301706B (en) 2014-10-11 2014-10-11 A kind of synthetic method for strengthening bore hole stereoscopic display effect

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410533537.4A CN104301706B (en) 2014-10-11 2014-10-11 A kind of synthetic method for strengthening bore hole stereoscopic display effect

Publications (2)

Publication Number Publication Date
CN104301706A true CN104301706A (en) 2015-01-21
CN104301706B CN104301706B (en) 2017-03-15

Family

ID=52321241

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410533537.4A Active CN104301706B (en) 2014-10-11 2014-10-11 A kind of synthetic method for strengthening bore hole stereoscopic display effect

Country Status (1)

Country Link
CN (1) CN104301706B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104820981A (en) * 2015-04-22 2015-08-05 上海交通大学 Image three-dimensional representation method based on parallax hierarchical segmentation, and system
CN105812767A (en) * 2016-03-17 2016-07-27 北京邮电大学 Three-dimensional display depth adjusting method and device based on multi-frequency decomposition
CN108195563A (en) * 2017-12-29 2018-06-22 深圳超多维科技有限公司 Display effect assessment method, device and the test and appraisal terminal of a kind of 3 d display device
CN109522866A (en) * 2018-11-29 2019-03-26 宁波视睿迪光电有限公司 Naked eye 3D rendering processing method, device and equipment
CN114095717A (en) * 2021-09-24 2022-02-25 锋芒科技南京有限公司 Optical field film source synthesis method

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050185048A1 (en) * 2004-02-20 2005-08-25 Samsung Electronics Co., Ltd. 3-D display system, apparatus, and method for reconstructing intermediate-view video
CN101322155A (en) * 2005-12-02 2008-12-10 皇家飞利浦电子股份有限公司 Stereoscopic image display method and apparatus, method for generating 3D image data from a 2D image data input and an apparatus for generating 3D image data from a 2D image data input
CN103269435A (en) * 2013-04-19 2013-08-28 四川长虹电器股份有限公司 Binocular to multi-view virtual viewpoint synthetic method
CN103384337A (en) * 2012-05-02 2013-11-06 索尼公司 Image processing apparatus and image processing method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050185048A1 (en) * 2004-02-20 2005-08-25 Samsung Electronics Co., Ltd. 3-D display system, apparatus, and method for reconstructing intermediate-view video
CN101322155A (en) * 2005-12-02 2008-12-10 皇家飞利浦电子股份有限公司 Stereoscopic image display method and apparatus, method for generating 3D image data from a 2D image data input and an apparatus for generating 3D image data from a 2D image data input
CN103384337A (en) * 2012-05-02 2013-11-06 索尼公司 Image processing apparatus and image processing method
CN103269435A (en) * 2013-04-19 2013-08-28 四川长虹电器股份有限公司 Binocular to multi-view virtual viewpoint synthetic method

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104820981A (en) * 2015-04-22 2015-08-05 上海交通大学 Image three-dimensional representation method based on parallax hierarchical segmentation, and system
CN104820981B (en) * 2015-04-22 2017-10-31 上海交通大学 A kind of image stereo representation method and system that segmentation is layered based on parallax
CN105812767A (en) * 2016-03-17 2016-07-27 北京邮电大学 Three-dimensional display depth adjusting method and device based on multi-frequency decomposition
CN108195563A (en) * 2017-12-29 2018-06-22 深圳超多维科技有限公司 Display effect assessment method, device and the test and appraisal terminal of a kind of 3 d display device
CN108195563B (en) * 2017-12-29 2020-04-28 深圳超多维科技有限公司 Display effect evaluation method and device of three-dimensional display device and evaluation terminal
CN109522866A (en) * 2018-11-29 2019-03-26 宁波视睿迪光电有限公司 Naked eye 3D rendering processing method, device and equipment
CN114095717A (en) * 2021-09-24 2022-02-25 锋芒科技南京有限公司 Optical field film source synthesis method

Also Published As

Publication number Publication date
CN104301706B (en) 2017-03-15

Similar Documents

Publication Publication Date Title
CN103236082B (en) Towards the accurate three-dimensional rebuilding method of two-dimensional video of catching static scene
Tam et al. 3D-TV content generation: 2D-to-3D conversion
CN109242954B (en) Multi-view three-dimensional human body reconstruction method based on template deformation
CN102223556B (en) Multi-view stereoscopic image parallax free correction method
CN101443817B (en) Method and device for determining correspondence, preferably for the three-dimensional reconstruction of a scene
CN103971408B (en) Three-dimensional facial model generating system and method
CN104301706B (en) A kind of synthetic method for strengthening bore hole stereoscopic display effect
CN104680496A (en) Kinect deep image remediation method based on colorful image segmentation
CN106709948A (en) Quick binocular stereo matching method based on superpixel segmentation
CN101866497A (en) Binocular stereo vision based intelligent three-dimensional human face rebuilding method and system
CN103248906B (en) Method and system for acquiring depth map of binocular stereo video sequence
CN110853151A (en) Three-dimensional point set recovery method based on video
CN102609950B (en) Two-dimensional video depth map generation process
CN111988593B (en) Three-dimensional image color correction method and system based on depth residual optimization
CN101996399A (en) Device and method for estimating parallax between left image and right image
CN104850847B (en) Image optimization system and method with automatic thin face function
CN103854301A (en) 3D reconstruction method of visible shell in complex background
CN106530336B (en) Stereo matching method based on color information and graph cut theory
CN106447718B (en) A kind of 2D turns 3D depth estimation method
CN103702103A (en) Optical grating three-dimensional printing image synthetic method based on binocular camera
CN105704476B (en) A kind of virtual visual point image frequency domain fast acquiring method based on edge reparation
Fan et al. Vivid-DIBR based 2D–3D image conversion system for 3D display
CN103955886A (en) 2D-3D image conversion method based on graph theory and vanishing point detection
CN106231290B (en) 3D Tree image construction methods based on depth template
CN103945206A (en) Three-dimensional picture synthesis system based on comparison between similar frames

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
CB02 Change of applicant information
CB02 Change of applicant information

Address after: Jinjiang District Cherry Street Chengdu city Sichuan province 610000 No. 999 green 468 two 4 Tung Mansion

Applicant after: CHENGDU SEEFELD SCIENCE & TECHNOLOGY CO., LTD.

Address before: 610041 Sichuan province Chengdu Tianfu four Street No. 199 Changhong Technology Building 35 floor

Applicant before: CHENGDU SEEFELD SCIENCE & TECHNOLOGY CO., LTD.

C14 Grant of patent or utility model
GR01 Patent grant
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20190116

Address after: Room 109, Room No. 1, 268 Dengyun Road, Yushan Town, Kunshan City, Suzhou City, Jiangsu Province

Patentee after: Kunshan Hui Ren Corelle Intelligent Technology Co., Ltd.

Address before: 610000 Chengdu City, Sichuan Province, Jinjiang District Cherry Blossom Street 999 Green Space 468 Mansion Phase II 4

Patentee before: CHENGDU SEEFELD SCIENCE & TECHNOLOGY CO., LTD.

CP03 Change of name, title or address
CP03 Change of name, title or address

Address after: 215300 room 105, 202, 203, No. 268, Dengyun Road, Yushan Town, Kunshan City, Suzhou City, Jiangsu Province

Patentee after: Huirenkangning (Suzhou) Health Technology Co., Ltd

Address before: Room 109, Room No. 1, 268 Dengyun Road, Yushan Town, Kunshan City, Suzhou City, Jiangsu Province

Patentee before: Kunshan Hui Ren Corelle Intelligent Technology Co., Ltd.