CN108062765A - Binocular image processing method, imaging device and electronic equipment - Google Patents
Binocular image processing method, imaging device and electronic equipment Download PDFInfo
- Publication number
- CN108062765A CN108062765A CN201711375443.9A CN201711375443A CN108062765A CN 108062765 A CN108062765 A CN 108062765A CN 201711375443 A CN201711375443 A CN 201711375443A CN 108062765 A CN108062765 A CN 108062765A
- Authority
- CN
- China
- Prior art keywords
- image
- pixel
- image processing
- parallax
- processing method
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/32—Determination of transform parameters for the alignment of images, i.e. image registration using correlation-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
- G06T7/55—Depth or shape recovery from multiple images
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
Imaging device and electronic equipment the present invention provides a kind of binocular image processing method and using the binocular image processing method.The binocular image processing method includes：The imaging unit being made of a pair of of Image Acquisition subelement gathers the first image and the second image；First image and the second image are sent to image processing unit and carry out Stereo matching, obtains the initial parallax of the first image and the second image；First image is divided into several pieces of subgraphs；The initial parallax of pixel in each subgraph is counted, determines to stablize pixel and unstable pixel；The parallax of unstable pixel is optimized, obtains final parallax.The present invention not only ensure that arithmetic speed, while improve the precision of Stereo matching disparity computation, have stronger practical application meaning.
Description
Technical field
The present invention relates to image processing techniques, more particularly to a kind of solid matching method of image segmentation post-processing.
Background technology
With the continuous development of stereoscopic vision, especially double appearances for taking the photograph mobile phone, realization picture list reverse phase that can be very light
The fuzzy virtualization of the machine depth of field different like that.Double take the photograph with the principle of human eye is similar, when people is with one eye eyeball, it is difficult to accurate
It is difficult to obtain scene depth really to identify the distance of object namely singly take the photograph mobile phone, and when using two eyes simultaneously, brain utilizes
The image at two visual angles reconstructs real three-dimensional world, so as to form the most important information source of the mankind --- vision.It is double to take the photograph
Also it is to copy human eye to work.To a certain scene in real world, data are carried out to it simultaneously using dual camera and are adopted
Collection imaging, so as to which some object point in three-dimensional world has been recorded in two images simultaneously, i.e., same object point corresponding two
A picture point.And one of double important steps for taking the photograph reconstruct three-dimensional world are exactly Stereo matching.
Stereo matching is using the picture point on reference picture (such as " left figure "), is found on target image (such as " right figure ")
Corresponding match point.According to Epipolar geometry knowledge, using left and right match point, parallax can be obtained, with reference to triangulation,
The depth of the same object point in the corresponding three-dimensional world of two picture points can be calculated.With this, left images can be calculated
The depth information of all object points in public view field.Stereo matching is the basis for obtaining Object Depth, and depth is to realize that background is empty
Change, depth survey, the unmanned basis for waiting applications.The precision of Stereo matching also determines the accuracy of these applications.
But the precision of Stereo matching is affected by many factors, as having large stretch of low texture in the scene or there are objects
When blocking, these corresponding pixels are known as point of instability (unstable point), otherwise pixel is known as stable point
(stable point).It is easy to cause disparity computation mistake for unstable point, so as to cause deep errors.For
Also there is diversified solid matching method in these defects, such as absolute error and algorithm SAD (Sum of Absolute
Differences), error sum of squares algorithm SSD (Sum of Squared Differences) normalizes product correlation al gorithm
NCC (Normalized Cross Correlation), confidence spread BP (Belief Propagation), figure is cut
(graph-cut) the methods of.These methods cut both ways, and as SAD principles itself are simple, but it is easy to appear matching errors；And
Graph-cut is complicated compared with the former principle, and accuracy is higher, but computationally intensive, and algorithm operation time is longer, is extremely difficult to count in real time
Calculate the requirement of depth.
The content of the invention
According to embodiments of the present invention, a kind of binocular image processing method is provided, for an electronic equipment, electronic equipment bag
The imaging unit being made of a pair of of Image Acquisition subelement is included, binocular image processing method includes：It is single by a pair of of Image Acquisition
The imaging unit that member is formed gathers the first image and the second image；First image and the second image are sent to image processing unit
Stereo matching is carried out, obtains the initial parallax of the first image and the second image；First image is divided into several pieces of subgraphs；It is right
The initial parallax of pixel in each subgraph is counted, and determines to stablize pixel and unstable pixel；To unstable
The parallax of pixel optimizes, and obtains final parallax.
Further, binocular image processing method according to embodiments of the present invention is calculated described vertical using global registration algorithm
Body matching obtains initial parallax.
Further, binocular image processing method according to embodiments of the present invention, global registration algorithm use confidence spread
Algorithm.
Further, binocular image processing method according to embodiments of the present invention, the Matching power flow bag of belief propagation algorithm
Contain but be not limited to：Pixel, the neighborhood territory pixel point information of pixel and neighborhood territory pixel neighborhood of a point point pass to neighborhood territory pixel point
Information.
Further, binocular image processing method according to embodiments of the present invention, using K-means algorithms to the first image into
Row image is split, and in K-means algorithms, closest cluster center is found in the range of certain distance for each pixel.
Further, binocular image processing method according to embodiments of the present invention finds the calculating at closest cluster center
Parameter includes：Distance and color intensity.
Further, binocular image processing method according to embodiments of the present invention is measured by confidence level to each subgraph
Interior pixel is judged, determines to stablize pixel and unstable pixel.
Further, binocular image processing method according to embodiments of the present invention, using plane fitting in each subgraph
The parallax of unstable pixel optimize.
According to another embodiment of the present invention, a kind of electronic equipment is provided, including：By a pair of of Image Acquisition subelement structure
Into imaging unit, image processing unit and memory.Wherein, imaging unit, for gathering the first image and the second image；
Image processing unit, for carrying out Stereo matching to the first image and the second image, obtain the first image and the second image just
Beginning parallax；First image is divided into several pieces of subgraphs；The initial parallax of pixel in each subgraph is counted,
It determines to stablize pixel and unstable pixel；The parallax of unstable pixel is optimized, obtains final parallax；Storage
Device, for storing the program and the required data of operation program of image processing unit operation.
According to another embodiment of the present invention, a kind of binocular image processing method is provided, for an imaging device, imaging
Device includes the imaging unit being made of a pair of of Image Acquisition subelement, and binocular image processing method includes：It is adopted by a pair of of image
Collect the imaging unit that subelement is formed and gather the first image and the second image；First image and the second image are sent at image
It manages unit and carries out Stereo matching, obtain the initial parallax of the first image and the second image；First image is divided into several pieces of sons
Image；The initial parallax of pixel in each subgraph is counted, determines to stablize pixel and unstable pixel；It is right
The parallax of unstable pixel optimizes, and obtains final parallax.
According to another embodiment of the present invention, a kind of imaging device is provided, including：Imaging unit, by a pair of of image
It gathers subelement to form, for gathering the first image and the second image；Image processing unit, for the first image and the second figure
As carrying out Stereo matching, the initial parallax of the first image and the second image is obtained；First image is divided into several pieces of subgraphs；
The initial parallax of pixel in each subgraph is counted, determines to stablize pixel and unstable pixel；To non-steady
The parallax of fixation vegetarian refreshments optimizes, and obtains final parallax.
Binocular image processing method according to embodiments of the present invention and the imaging dress using the binocular image processing method
It puts and electronic equipment, not only ensure that arithmetic speed, while improve the precision of Stereo matching disparity computation, have stronger
Practical application meaning.
It is to be understood that foregoing general description and following detailed description are both illustrative, and it is intended to
In the further explanation for providing claimed technology.
Description of the drawings
Fig. 1 is the block diagram for illustrating electronic equipment according to the present invention；
Fig. 2 is the flow chart for illustrating the binocular image processing method of electronic equipment according to the present invention；
Fig. 3 is the block diagram for illustrating imaging device according to the present invention；
Fig. 4 is the flow chart for illustrating the binocular image processing method of imaging device according to the present invention.
Specific embodiment
Below with reference to attached drawing, detailed description of the present invention preferred embodiment is further elaborated the present invention.
First, electronic equipment according to embodiments of the present invention, electronic equipment of the invention preferably example will be described with reference to Fig. 1
In this way：Any one and other in smart mobile phone, tablet computer, digital camera, laptop etc. are real with dual camera
Now gather the electronic equipment of image.
Fig. 1 is the block diagram of diagram electronic equipment according to embodiments of the present invention.As shown in Figure 1, the electricity of the embodiment of the present invention
Sub- equipment 1 has imaging unit 11, image processing unit 12 and storage unit 13, it is to be understood that for simplification in Fig. 1
Description illustrate only the component being closely related with the present invention, and electronic equipment 1 according to embodiments of the present invention can also include all
As central processing unit, communication unit, I/O units other assemblies.
Specifically, imaging unit 11 includes a pair of of Image Acquisition subelement, and in the present embodiment, a pair of of Image Acquisition is single
Member is a pair of of camera 111 and 112, and camera 111 and 112 shown in FIG. 1 arranges that the arrangement is only example for left and right
Property, camera 111 and 112 can also be arranged up and down.In the present embodiment, camera 111 and 112 be by calibration, it is each
From the image of acquisition physical world, the binocular image so obtained meets to epipolar-line constraint, i.e. corresponding pixel points are in image
With in a line, a large amount of operands are saved for subsequent algorithm.
Specifically, image processing unit 12 be used for a pair of of Image Acquisition subelement, that is, camera 111 and 112 form into
As unit 11 obtains left mesh image and right mesh image.Image processing unit 12 can be by such as image processing unit GPU, number
Any configuration in signal processor DSP, application-specific integrated circuit ASIC.
Since using left mesh head picture as reference picture, in the present embodiment, image processing unit 12 performs left mesh image complete
The processing such as office's matching, image segmentation, confidence level measurement, plane fitting, to obtain accurate final parallax.
Specifically, storage unit 13 can be used for being stored in image processing unit 12 handle and control program and
Permanently or temporarily store image data and input or output data.Storage unit 13 can select flash-type storage medium, hard
The miniature storage medium of dish-type storage medium, multimedia card, random access memory, read-only memory etc..In addition, electronic equipment 1
Network storage medium such as cloud platform can be operated, is passed to the function that network transmission performs storage unit 13.In the present invention
One embodiment in, left mesh image, right mesh image and other related ephemeral datas are stored in by image processing unit 12
In storage unit 13, and image processing unit 12 reads various data from storage unit 13.
As described above, in electronic equipment 1 according to embodiments of the present invention, by double to being arranged by such as camera or so
The raw image data that camera 111 and 112 obtains, and perform global registration, image segmentation, confidence level measurement, plane fitting
One or more of, so as to obtain the final parallax of binocular view, to ensure arithmetic speed, while improve Stereo matching
The precision of disparity computation has stronger practical application meaning.
Electronic equipment 1 according to embodiments of the present invention is described above in association with Fig. 1, hereinafter with reference to Fig. 2 descriptions according to this
The binocular image processing method of inventive embodiments.
Fig. 2 is the binocular image process flow figure of diagram electronic equipment according to embodiments of the present invention.Such as Fig. 2 institutes
Show, the binocular image processing method of electronic equipment according to embodiments of the present invention includes the following steps：
In step s 11, left mesh image and right mesh image are obtained by camera 111,112, and by left mesh image and right mesh
Image rectification is complete；Hereafter, step process is entered in step S12.
In step s 12, left mesh image and right mesh image are sent to image processing unit 12 and carry out Stereo matching, obtained
The initial parallax of left mesh image and right mesh image；Specifically, in the present embodiment, using global Stereo Matching Algorithm：Confidence level
Propagation algorithm (Belief Propagation) carries out disparity computation：The formula is
Matching power flow (or matching error) cost calculation formula of any pixel point q, wherein Dq(fq) it is the cost meters of pixel q in itself
It calculates,The Matching power flow of q is passed to for the vicinity points of the T times iterative calculation time point q.Hereafter, at step
Reason is entered in step S13.
In step s 13, left mesh image is divided into several pieces of subgraphs；Specifically, using K-means to left mesh image
It is split, the segmentation of K-means images is to minimize criterion based on error sum of squaresWherein, k
For the block number of subgraph, uiFor the central pixel point of i-th piece of subgraph, ciFor any one piece of subgraph.According toClosest central pixel point is found for each pixel, and is grouped into respective cluster
(cluster), k cluster is formed.ciRepresent i-th of cluster, uiFor the central pixel point of corresponding cluster, x is arbitrary point.And in this implementation
In example, when carrying out image segmentation using K-means, when finding closest cluster central pixel point for each pixel, no
It is to calculate any pixel point to the distance of all cluster central pixel points again, is then judged, but in the range of certain distance
It finds, as window size can beN is sum of all pixels, and K is the number of blocks of segmentation.Further, counting
Calculate pixel to each center apart from when, can consider distance and color intensity two parts：Wherein, Hereafter, step process enters step
In rapid S14.
In step S14, the initial parallax of the pixel in each subgraph is counted, determine stablize pixel and
Unstable pixel；Specifically, in the present embodiment, (confidence measurement) is measured to subgraph using confidence level
Pixel as in is judged, is for stable point or unstable point, formula：Wherein, C1And C2
The minimum value and sub-minimum of Matching power flow cost when respectively finding match point, when ratio is more than predetermined threshold value, the pixel
It can be determined that as stable point, conversely, being then point of instability.Hereafter, step process is entered in step S15.
In step S15, the parallax of unstable pixel is optimized, obtains final parallax.Specifically, in this reality
It applies in example, the parallax of the unstable point in each subgraph is optimized using plane fitting (plane fitting), is obtained
One accurate final parallax.
More than, describe electronic equipment according to embodiments of the present invention and its binocular image processing side referring to figs. 1 to Fig. 2
Method.Further, it is also possible to apply the invention to imaging devices 2.
As shown in figure 3, imaging device 2 according to embodiments of the present invention includes imaging unit 21 and image processing unit 22.
Specifically, imaging unit 21 is similar to the imaging unit 11 with reference to Fig. 1 descriptions, and including a pair of of camera 211,212, a pair is taken the photograph
As first 211,212 or so arrangements, and by calibrating, the image of physical world is each gathered, the binocular image so obtained is expired
To epipolar-line constraint, i.e. corresponding pixel points are in same a line of image foot, and a large amount of operands are saved for subsequent algorithm.Camera can
To be colour imagery shot or black and white camera.Specifically, image processing unit 22 is used for single to a pair of of Image Acquisition
Member is that the imaging unit 21 that camera 211 and 212 is formed obtains left mesh image and right mesh image.Image processing unit 22 can be
By any configuration in such as image processing unit GPU, digital signal processor DSP, application-specific integrated circuit ASIC.
Since using left mesh head picture as reference picture, in the present embodiment, image processing unit 22 performs left mesh image complete
One or more of processing such as office's matching, image segmentation, confidence level measurement, plane fitting, to obtain accurate final parallax
Figure.
As described above, in imaging device 2 according to embodiments of the present invention, by double to being arranged by such as camera or so
The raw image data that camera 211,212 obtains, and perform global registration, image segmentation, confidence level measurement, plane fitting etc.
One or more of processing, to obtain accurate final parallax, to ensure arithmetic speed, while improves Stereo matching and regards
The precision that difference calculates has stronger practical application meaning.
As shown in figure 4, the binocular image processing method of imaging device according to embodiments of the present invention includes the following steps：
S21：Left mesh image and right mesh image are obtained by camera 211,212, and by left mesh image and right mesh image rectification
Completely；
S22：Left mesh image and right mesh image are sent to image processing unit 22 and carry out Stereo matching, obtains left mesh image
With the initial parallax of right mesh image；
S23：Left mesh image is divided into several pieces of subgraphs；
S24：The initial parallax of pixel in each subgraph is counted, determines to stablize pixel and unstable picture
Vegetarian refreshments；
S25：The parallax of unstable pixel is optimized, obtains final parallax.
The binocular image processing method one of specific image processing method and electronic equipment for the embodiment of the present invention
It causes, details are not described herein again.
More than, it describes image processing method according to embodiments of the present invention with reference to Fig. 1~4 and uses the image procossing
The imaging device and electronic equipment of method, by handling the raw image data obtained by such as dual camera, not only
Ensure arithmetic speed, also improve the precision of Stereo matching disparity computation, there is stronger practical application meaning.
It should be noted that in the present specification, term " comprising ", "comprising" or its any other variant are intended to
Non-exclusive inclusion, so that process, method, article or equipment including a series of elements not only will including those
Element, but also including other elements that are not explicitly listed or further include as this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
Also there are other identical elements in process, method, article or equipment including the element.
Finally, it is to be noted that, a series of above-mentioned processing are not only included with order described here in temporal sequence
The processing of execution, and the processing including performing parallel or respectively rather than in chronological order.
Through the above description of the embodiments, those skilled in the art can be understood that the present invention can be by
Software adds the mode of required hardware platform to realize, naturally it is also possible to all be implemented by hardware.Based on such understanding,
Technical scheme can be embodied in the form of software product in whole or in part to what background technology contributed,
The computer software product can be stored in storage medium, such as ROM/RAM, magnetic disc, CD, including some instructions making
It obtains a computer equipment (can be personal computer, server or network equipment etc.) and performs each embodiment of the present invention
Or the method described in some parts of embodiment.
Claims (11)
1. a kind of binocular image processing method, for an electronic equipment, the electronic equipment includes single by a pair of of Image Acquisition
The imaging unit that member is formed, it is characterised in that：
The imaging unit being made of the pair of Image Acquisition subelement gathers the first image and the second image；
By described first image and the second image be sent to image processing unit carry out Stereo matching, obtain described first image and
The initial parallax of second image；
Described first image is divided into several pieces of subgraphs；
The initial parallax of pixel in each subgraph is counted, determines to stablize pixel and unstable pixel
Point；
The parallax of the unstable pixel is optimized, obtains final parallax.
2. binocular image processing method as described in claim 1, which is characterized in that calculated using global registration algorithm described vertical
Body matching obtains initial parallax.
3. binocular image processing method as claimed in claim 2, which is characterized in that the global registration algorithm uses confidence level
Propagation algorithm.
4. binocular image processing method as claimed in claim 3, which is characterized in that the matching generation of the belief propagation algorithm
Valency including but not limited to：Pixel, the neighborhood territory pixel point information of the pixel and the neighborhood territory pixel neighborhood of a point point transfer
To the information of the neighborhood territory pixel point.
5. binocular image processing method as claimed in claim 1 or 2, which is characterized in that using K-means algorithms to described
One image carries out image segmentation, in the K-means algorithms, is found for each pixel in the range of certain distance closest
Cluster center.
6. binocular image processing method as claimed in claim 5, which is characterized in that the closest cluster center of the searching
Calculating parameter includes：Distance and color intensity.
7. binocular image processing method as described in claim 1, which is characterized in that measured by confidence level to each subgraph
Interior pixel is judged, determines to stablize pixel and unstable pixel.
8. the binocular image processing method as described in claim 1 or 7, which is characterized in that using plane fitting to each subgraph
The parallax of unstable pixel as in optimizes.
9. a kind of electronic equipment, including the imaging unit being made of a pair of of Image Acquisition subelement, image processing unit and storage
Device, which is characterized in that
The imaging unit, for gathering the first image and the second image；
Described image processing unit for carrying out Stereo matching to described first image and the second image, obtains first figure
The initial parallax of picture and the second image；Described first image is divided into several pieces of subgraphs；To the pixel in each subgraph
The initial parallax of point is counted, and determines to stablize pixel and unstable pixel；To the parallax of the unstable pixel into
Row optimization, obtains final parallax；
The memory, for storing the program and the required data of operation program of the operation of described image processing unit.
10. a kind of binocular image processing method, for an imaging device, the imaging device includes single by a pair of of Image Acquisition
The imaging unit that member is formed, which is characterized in that the binocular image processing method includes：
The imaging unit being made of the pair of Image Acquisition subelement gathers the first image and the second image；
By described first image and the second image be sent to image processing unit carry out Stereo matching, obtain described first image and
The initial parallax of second image；
Described first image is divided into several pieces of subgraphs；
The initial parallax of pixel in each subgraph is counted, determines to stablize pixel and unstable pixel；
The parallax of the unstable pixel is optimized, obtains final parallax.
11. a kind of imaging device, which is characterized in that including：
Imaging unit is made of a pair of of Image Acquisition subelement, for gathering the first image and the second image；
Image processing unit, for carrying out Stereo matching to described first image and the second image, obtain described first image and
The initial parallax of second image；Described first image is divided into several pieces of subgraphs；To the pixel in each subgraph
Initial parallax is counted, and determines to stablize pixel and unstable pixel；The parallax of the unstable pixel is carried out excellent
Change, obtain final parallax.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711375443.9A CN108062765A (en) | 2017-12-19 | 2017-12-19 | Binocular image processing method, imaging device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711375443.9A CN108062765A (en) | 2017-12-19 | 2017-12-19 | Binocular image processing method, imaging device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108062765A true CN108062765A (en) | 2018-05-22 |
Family
ID=62139557
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711375443.9A Pending CN108062765A (en) | 2017-12-19 | 2017-12-19 | Binocular image processing method, imaging device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108062765A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110021043A (en) * | 2019-02-28 | 2019-07-16 | 浙江大学 | A kind of scene depth acquisition methods based on Stereo matching and confidence spread |
| CN111382591A (en) * | 2018-12-27 | 2020-07-07 | 海信集团有限公司 | Binocular camera ranging correction method and vehicle-mounted equipment |
| CN113129400A (en) * | 2021-03-17 | 2021-07-16 | 维沃移动通信有限公司 | Image processing method, image processing device, electronic equipment and readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101976455A (en) * | 2010-10-08 | 2011-02-16 | 东南大学 | Color image three-dimensional reconstruction method based on three-dimensional matching |
| KR101505360B1 (en) * | 2014-01-22 | 2015-03-30 | 경북대학교 산학협력단 | A disparity search range estimation apparatus and method |
| CN104966290A (en) * | 2015-06-12 | 2015-10-07 | 天津大学 | Self-adaptive weight three-dimensional matching method based on SIFT descriptor |
| CN105513064A (en) * | 2015-12-03 | 2016-04-20 | 浙江万里学院 | Image segmentation and adaptive weighting-based stereo matching method |
| CN107301664A (en) * | 2017-05-25 | 2017-10-27 | 天津大学 | Improvement sectional perspective matching process based on similarity measure function |
-
2017
- 2017-12-19 CN CN201711375443.9A patent/CN108062765A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101976455A (en) * | 2010-10-08 | 2011-02-16 | 东南大学 | Color image three-dimensional reconstruction method based on three-dimensional matching |
| KR101505360B1 (en) * | 2014-01-22 | 2015-03-30 | 경북대학교 산학협력단 | A disparity search range estimation apparatus and method |
| CN104966290A (en) * | 2015-06-12 | 2015-10-07 | 天津大学 | Self-adaptive weight three-dimensional matching method based on SIFT descriptor |
| CN105513064A (en) * | 2015-12-03 | 2016-04-20 | 浙江万里学院 | Image segmentation and adaptive weighting-based stereo matching method |
| CN107301664A (en) * | 2017-05-25 | 2017-10-27 | 天津大学 | Improvement sectional perspective matching process based on similarity measure function |
Non-Patent Citations (5)
| Title |
|---|
| KUK-JIN YOON ET AL: ""Locally Adaptive Support-Weight Approach for Visual Correspondence Search"", 《IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
| 曹云云等: ""一种具有选择性的置信度传播立体匹配算法"", 《东南大学学报》 * |
| 祁敏等: ""基于Tao 立体匹配框架的全局优化算法"", 《红外与激光工程》 * |
| 郁理等: ""基于分级置信度传播的立体匹配新方法"", 《中国图象图形学报》 * |
| 高博: ""基于彩色图像分割的立体匹配算法研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111382591A (en) * | 2018-12-27 | 2020-07-07 | 海信集团有限公司 | Binocular camera ranging correction method and vehicle-mounted equipment |
| CN111382591B (en) * | 2018-12-27 | 2023-09-29 | 海信集团有限公司 | Binocular camera ranging correction method and vehicle-mounted equipment |
| CN110021043A (en) * | 2019-02-28 | 2019-07-16 | 浙江大学 | A kind of scene depth acquisition methods based on Stereo matching and confidence spread |
| CN113129400A (en) * | 2021-03-17 | 2021-07-16 | 维沃移动通信有限公司 | Image processing method, image processing device, electronic equipment and readable storage medium |
| CN113129400B (en) * | 2021-03-17 | 2023-02-24 | 维沃移动通信有限公司 | Image processing method, image processing device, electronic equipment and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11687151B2 (en) | Methods and apparatuses for determining and/or evaluating localizing maps of image display devices | |
| Yang | Dealing with textureless regions and specular highlights-a progressive space carving scheme using a novel photo-consistency measure | |
| Čech et al. | Scene flow estimation by growing correspondence seeds | |
| CN104867135B (en) | A kind of High Precision Stereo matching process guided based on guide image | |
| CN110148181A (en) | A kind of general binocular solid matching process | |
| US20240048932A1 (en) | Personalized hrtfs via optical capture | |
| CN108734776A (en) | A kind of three-dimensional facial reconstruction method and equipment based on speckle | |
| Perri et al. | Adaptive Census Transform: A novel hardware-oriented stereovision algorithm | |
| CN109640066B (en) | Method and device for generating high-precision dense depth image | |
| WO2008029345A1 (en) | Method for determining a depth map from images, device for determining a depth map | |
| CN105336005B (en) | A kind of method, apparatus and terminal obtaining target object sign data | |
| CN109448036A (en) | A kind of method and device determining disparity map based on binocular image | |
| CN108062765A (en) | Binocular image processing method, imaging device and electronic equipment | |
| WO2020010561A1 (en) | Method and apparatus for measuring object parameters | |
| US20240029301A1 (en) | Efficient localization based on multiple feature types | |
| Gadasin et al. | Reconstruction of a Three-Dimensional Scene from its Projections in Computer Vision Systems | |
| CN109644280A (en) | The method for generating the depth of seam division data of scene | |
| CN109842811A (en) | A kind of method, apparatus and electronic equipment being implanted into pushed information in video | |
| CN110443228B (en) | Pedestrian matching method and device, electronic equipment and storage medium | |
| CN109544622A (en) | A kind of binocular vision solid matching method and system based on MSER | |
| CN109544619A (en) | A kind of binocular vision solid matching method cut based on figure and system | |
| CN116704111A (en) | Image processing method and apparatus | |
| Leimkühler et al. | Perceptual real-time 2D-to-3D conversion using cue fusion | |
| Liu et al. | Improved template matching based stereo vision sparse 3D reconstruction algorithm | |
| CN108062741A (en) | Binocular image processing method, imaging device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180522 |