CN106097289B - A kind of stereo-picture synthetic method based on MapReduce model - Google Patents
A kind of stereo-picture synthetic method based on MapReduce model Download PDFInfo
- Publication number
- CN106097289B CN106097289B CN201610377079.9A CN201610377079A CN106097289B CN 106097289 B CN106097289 B CN 106097289B CN 201610377079 A CN201610377079 A CN 201610377079A CN 106097289 B CN106097289 B CN 106097289B
- Authority
- CN
- China
- Prior art keywords
- stereo
- picture
- original image
- pixel difference
- average level
- 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.)
- Expired - Fee Related
Links
- 238000010189 synthetic method Methods 0.000 title claims abstract description 8
- 238000000034 method Methods 0.000 claims abstract description 30
- 230000015572 biosynthetic process Effects 0.000 claims abstract description 13
- 238000003786 synthesis reaction Methods 0.000 claims abstract description 13
- 230000000694 effects Effects 0.000 claims abstract description 7
- 238000002360 preparation method Methods 0.000 claims abstract description 4
- 238000006073 displacement reaction Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 208000003164 Diplopia Diseases 0.000 claims description 3
- 208000029444 double vision Diseases 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- YREOLPGEVLLKMB-UHFFFAOYSA-N 3-methylpyridin-1-ium-2-amine bromide hydrate Chemical compound O.[Br-].Cc1ccc[nH+]c1N YREOLPGEVLLKMB-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000003707 image sharpening Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
Abstract
The stereo-picture synthetic method based on MapReduce model that the present invention relates to a kind of, includes the following steps:(1) original image calculates average level pixel difference to data preparation (2);(3) map process;(4) reduce process;By the average information image in (3) and the original image in (1) to together as input, pass through the synthesis of the overlapping realization stereo-picture of pixel column, meanwhile using average level pixel difference as overlapping benchmark, the synthetic effect of stereo-picture is adjusted;(5) edge sharpening.The present invention has the advantages that efficient and practicability.
Description
Technical field
The invention belongs to technical field of image processing, are related to a kind of stereo-picture synthetic method.
Background technique
Stereo display technique is the third generation after the white and black displays technology of the first generation and the color display technology of the second generation
Multimedia display technology can bring great visual enjoyment for user.Stereo-picture synthesis refers to the figure by computer
As Processing Algorithm handles two dimensional image, the process with 3 D visual image is obtained.Stereo-picture synthesis at present is
Become a research hotspot of computer nowadays research field.
Cloud computing is a kind of emerging calculating mode, it provides several times by distribution, parallelization, modular mode and is
To the calculated performance of decades of times, it can be used for solving the bottleneck problem of single cpu mode image procossing.Hadoop is by Google
The open source distributed system architecture that three distributed computing papers derive.It is current most popular big data technological frame,
By the mode of open source, numerous manufacturers and developer is attracted to participate, keeps its ecosphere mature rapidly.Hadoop system frame
It is made of HDFS (Hadoop distributed file system) and MapReduce (parallel computational model) two parts.Wherein, HDFS is big
Data volume file cocurrent storage provide may, the data access performance of high-throughput can be provided, at the same have high reliability and
Stable fault tolerance;MapReduce is a kind of programming model, the characteristic with functional expression programming with vector programming, magnanimity
The treatment process parallelization of data is simplified, and the data processing work for being bad at distributed programmed personnel is greatly facilitated.
Hadoop frame is used widely in big data processing field.
With the development of internet and smart machine, multimedia messages enter the epoch of an explosive increase.And image
As the important form of multimedia messages, more become the research hotspot of computer field.And it is past research be often confined to it is small
The image procossing of data volume, but with the continuous growth of image data amount, the image procossing of small data quantity has been under single cpu mode
It is far from satisfying requirement.The treatment process of large nuber of images must be a time-consuming process, how improve processing speed, realize
The optimization for the treatment of effeciency becomes as the task of top priority.
Summary of the invention
The present invention provides a kind of stereo-picture synthetic method based on MapReduce model.The present invention is by by Hadoop
Frame application synthesizes field in stereo-picture, provides a kind of efficient, practical stereo-picture synthetic method.Technical side of the invention
Case is as follows:
A kind of stereo-picture synthetic method based on MapReduce model, includes the following steps:
(1) original image is to data preparation
By what is be acquired by double vision point camera, the original image pair comprising stereoscopic depth information;It is stored in Hadoop
In HDFS distributed file system on platform, MapReduce frame parallelization is facilitated to read;
(2) average level pixel difference is calculated
By SAD algorithm to original image to matching, and then average level pixel difference is obtained, method is as follows:With a left side
Centered on the source match point of view, the sum that the window that a size is (2N+1) * (2N+1) counts its gray value is constructed, is then existed
In right view an equal amount of window is moved as displacement vector using d in the horizontal direction and count the sum of gray value, works as gray value
And it is minimum when, current displacement vector d is that horizontal pixel is poor;The all pixels point of entire original image pair is used above
Method, obtain horizontal pixel difference matrix, average value carried out on entire horizontal pixel difference matrix, obtain average level pixel
Difference;
(3) map process
It, can by left images to be added except 2 by the map process of MapReduce frame for original image pair
To obtain average information image;
(4) reduce process
By the average information image in (3) and the original image in (1) to together as input, pass through the overlapping of pixel column
Realize the synthesis of stereo-picture, meanwhile, using average level pixel difference as overlapping benchmark, adjust the synthesis effect of stereo-picture
Fruit;
(5) edge sharpening
For the composograph exported after (4) processing, it is sharpened processing, the stereo-picture after output sharpening.
Beneficial effects of the present invention are as follows:
1. the algorithm provided by the invention based on MapReduce model synthetic stereo image can utilize MapReduce frame
Frame realizes the distributed computing synthesized to stereo-picture, improves processing speed and efficiency.
2. the present invention can adjust increment, the convenient display for adjusting stereo-picture according to the horizontal pixel difference calculated
Effect has self-optimization ability.
3. scalability of the present invention is strong, three-dimensional video-frequency synthesis field, very convenient progress parallelization processing can be applied to.
Detailed description of the invention
Fig. 1 scheme block diagram.
Fig. 2 edge sharpening.
Specific embodiment
The basic thought of method proposed by the present invention is:The present invention uses the Hadoop platform handled dedicated for big data
To carry out image processing process.It is distributed programmed using MapReduce model progress, by average information figure composition algorithm and Map
It is combined with Reduce process, parallelization, which calculates, obtains stereo-picture.In order to optimize the display effect of stereo-picture, it is added horizontal
Pixel difference automatic adjusument, constantly feedback parameter optimizes three-dimensional synthesis process during stereo-picture synthesis.Final defeated
Before out, edge sharpening is carried out, solves the problems, such as that average information nomography introduces edge blurry.Scheme overall schematic such as Fig. 1, tool
Steps are as follows for body:
1 original image is to data preparation
It is acquired by double vision point camera, obtains image pair largely comprising stereoscopic depth information.By image log
It, while can according in the HDFS distributed file system being stored in Hadoop platform, facilitating MapReduce frame parallelization to read
To simplify input process, while safety with higher.
2 calculate average level pixel difference
To original image pair, we can by SAD algorithm to left images to matching, and then obtain average level
Pixel difference.The specific method is as follows:Centered on the source match point of left view, the window that a size is (2N+1) * (2N+1) is constructed
Mouth counts the sum of its gray value, then moves an equal amount of window in the horizontal direction as displacement vector using d in right view
And count the sum of gray value.When sum of the grayscale values minimum, current displacement vector d is that horizontal pixel is poor.To entire picture
All pixels point uses above method, can obtain horizontal pixel difference matrix.It is averagely taken on entire pixel difference matrix
Value, so that it may obtain average level pixel difference.
3 map processes
Average information image can be obtained by left images to addition is carried out except 2 to original image pair.Average information figure
More different informations are contained as in, less identical information can increase composograph stereo display effect.By this mistake
Journey is realized by the map process of MapReduce frame, and the operational efficiency of algorithm can be improved, and the same time can handle
More pictures.
4 reduce processes
During reduce, by the left images in the average information image and 1 in 3 to together as input, pass through
The synthesis of the overlapping realization stereo-picture of pixel column.Meanwhile using the average level pixel difference in 5.2 as overlapping benchmark, adjust
The synthetic effect of whole stereo-picture.
5 edge sharpenings
For the synthetic stereo image of 4 outputs, edge blurry is solved the problems, such as by the method for image sharpening.Input is three-dimensional
Image is calculated with sobel operator and obtains image border, stereo-picture and edge are summed up, the perspective view after output sharpening
Picture.The schematic diagram of detailed process such as Fig. 2.
The three-dimensional media resource that the present invention is suitable for MultiMedia Field produces business.Traditional solid based on single cpu mode
Image synthesis system can not cope with the processing and synthesis task of big data era large nuber of images.The present invention uses
MapReduce model solves the parallel computation problem of large-scale image, comes the hind computation time of reduction image synthesis system, main
The big data for having used Hadoop reads the ability of characteristic and the parallel computation of MapReduce.
Claims (1)
1. a kind of stereo-picture synthetic method based on MapReduce model, includes the following steps:
(1) original image is to data preparation
By what is be acquired by double vision point camera, the original image pair comprising stereoscopic depth information;It is stored in Hadoop platform
On HDFS distributed file system in, facilitate MapReduce frame parallelization read;
(2) average level pixel difference is calculated
By SAD algorithm to original image to matching, and then average level pixel difference is obtained, method is as follows:With left view
Source match point centered on, construct the window that a size is (2N+1) * (2N+1) and count the sum of its gray value, then in right view
In figure an equal amount of window is moved as displacement vector using d in the horizontal direction and count the sum of gray value, when the sum of gray value
When minimum, current displacement vector d is that horizontal pixel is poor;To the side above all pixels point use of entire original image pair
Method obtains horizontal pixel difference matrix, and average value is carried out on entire horizontal pixel difference matrix, obtains average level pixel difference;
(3) map process
For original image pair, can be obtained by the map process of MapReduce frame by left images to addition is carried out except 2
Obtain average information image;
(4) reduce process
By the average information image in (3) and the original image in (1) to together as input, pass through the overlapping realization of pixel column
The synthesis of stereo-picture, meanwhile, using average level pixel difference as overlapping benchmark, adjust the synthetic effect of stereo-picture;
(5) edge sharpening
For the composograph exported after (4) processing, it is sharpened processing, the stereo-picture after output sharpening.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610377079.9A CN106097289B (en) | 2016-05-30 | 2016-05-30 | A kind of stereo-picture synthetic method based on MapReduce model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610377079.9A CN106097289B (en) | 2016-05-30 | 2016-05-30 | A kind of stereo-picture synthetic method based on MapReduce model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106097289A CN106097289A (en) | 2016-11-09 |
CN106097289B true CN106097289B (en) | 2018-11-27 |
Family
ID=57229685
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610377079.9A Expired - Fee Related CN106097289B (en) | 2016-05-30 | 2016-05-30 | A kind of stereo-picture synthetic method based on MapReduce model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106097289B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111770273B (en) * | 2020-06-29 | 2021-12-07 | 维沃移动通信有限公司 | Image shooting method and device, electronic equipment and readable storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101175223A (en) * | 2007-07-10 | 2008-05-07 | 天津大学 | Multi-view point stereoscopic picture synthesizing method for LCD free stereoscopic display device based on optical grating |
CN102026013A (en) * | 2010-12-18 | 2011-04-20 | 浙江大学 | Stereo video matching method based on affine transformation |
CN102143321A (en) * | 2010-02-01 | 2011-08-03 | 卡西欧计算机株式会社 | Image capture apparatus and control method |
CN103702103A (en) * | 2014-01-10 | 2014-04-02 | 武汉大学 | Optical grating three-dimensional printing image synthetic method based on binocular camera |
CN105430368A (en) * | 2014-09-22 | 2016-03-23 | 中兴通讯股份有限公司 | Two-viewpoint stereo image synthesizing method and system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101200490B1 (en) * | 2008-12-10 | 2012-11-12 | 한국전자통신연구원 | Apparatus and Method for Matching Image |
-
2016
- 2016-05-30 CN CN201610377079.9A patent/CN106097289B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101175223A (en) * | 2007-07-10 | 2008-05-07 | 天津大学 | Multi-view point stereoscopic picture synthesizing method for LCD free stereoscopic display device based on optical grating |
CN102143321A (en) * | 2010-02-01 | 2011-08-03 | 卡西欧计算机株式会社 | Image capture apparatus and control method |
CN102026013A (en) * | 2010-12-18 | 2011-04-20 | 浙江大学 | Stereo video matching method based on affine transformation |
CN103702103A (en) * | 2014-01-10 | 2014-04-02 | 武汉大学 | Optical grating three-dimensional printing image synthetic method based on binocular camera |
CN105430368A (en) * | 2014-09-22 | 2016-03-23 | 中兴通讯股份有限公司 | Two-viewpoint stereo image synthesizing method and system |
Also Published As
Publication number | Publication date |
---|---|
CN106097289A (en) | 2016-11-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Casser et al. | Unsupervised monocular depth and ego-motion learning with structure and semantics | |
GB2544596B (en) | Style transfer for headshot portraits | |
CN104008569B (en) | A kind of 3D scene generating method based on deep video | |
CN108961327A (en) | A kind of monocular depth estimation method and its device, equipment and storage medium | |
Mu et al. | 3d photo stylization: Learning to generate stylized novel views from a single image | |
CN103236082A (en) | Quasi-three dimensional reconstruction method for acquiring two-dimensional videos of static scenes | |
CN103606186A (en) | Virtual hair style modeling method of images and videos | |
CN108124489B (en) | Information processing method, apparatus, cloud processing device and computer program product | |
US9736449B1 (en) | Conversion of 2D image to 3D video | |
CN104506872B (en) | A kind of method and device of converting plane video into stereoscopic video | |
CN112884682A (en) | Stereo image color correction method and system based on matching and fusion | |
CN111612878B (en) | Method and device for making static photo into three-dimensional effect video | |
US10922852B2 (en) | Oil painting stroke simulation using neural network | |
CN103927787A (en) | Method and device for improving three-dimensional reconstruction precision based on matrix recovery | |
CN106651835A (en) | Entropy-based double-viewpoint reference-free objective stereo-image quality evaluation method | |
BR102020027013A2 (en) | METHOD TO GENERATE AN ADAPTIVE MULTIPLANE IMAGE FROM A SINGLE HIGH RESOLUTION IMAGE | |
CN106097289B (en) | A kind of stereo-picture synthetic method based on MapReduce model | |
Xiong et al. | An efficient underwater image enhancement model with extensive Beer-Lambert law | |
US10832469B2 (en) | Optimizing images for three-dimensional model construction | |
Lu et al. | 3D real-time human reconstruction with a single RGBD camera | |
Chen et al. | Panic-3d: Stylized single-view 3d reconstruction from portraits of anime characters | |
WO2021245326A1 (en) | A method, an apparatus and a computer program product for video encoding and video decoding | |
KR20160081841A (en) | Automatic extracting system for 3d digital image object based on 2d digital image and extracting method using thereof | |
CN117115398A (en) | Virtual-real fusion digital twin fluid phenomenon simulation method | |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181127 |