CN108109148A - Image solid distribution method, mobile terminal - Google Patents
Image solid distribution method, mobile terminal Download PDFInfo
- Publication number
- CN108109148A CN108109148A CN201711319033.2A CN201711319033A CN108109148A CN 108109148 A CN108109148 A CN 108109148A CN 201711319033 A CN201711319033 A CN 201711319033A CN 108109148 A CN108109148 A CN 108109148A
- Authority
- CN
- China
- Prior art keywords
- image
- block
- point
- mrow
- matching
- 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
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
Abstract
The invention discloses a kind of image volume segmentation method, mobile terminal, which includes：Obtained the first image progress image segmentation will be shot, to obtain multiple images block；Respectively characteristic point is extracted from each described image block；It is found on obtained the second image of shooting and matched with the characteristic point extracted in described image block and the match point of Matching power flow value minimum using matching algorithm；Wherein, described first image and second image are the images shot with different view to Same Scene；According to matching result, the parallax value of all characteristic points in any image block is calculated；Parallax value based on all characteristic points in image block is to image district all pixels in the block point assignment, to generate disparity map.Using the technical program in the case where ensureing precision, calculation amount is dramatically saved, while also saves hardware, memory overhead, the extensive use for Stereo matching and depth algorithm is laid a good foundation.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of image solid distribution method, mobile terminals.
Background technology
With emerging for Dual-camera handset and making constant progress for other imaging devices, Stereo Matching Algorithm obtains
More and more applications.Shooting or single camera are carried out at the same time to Same Scene using two camera modules to same field
Scape is repeatedly shot in different angle, and sequence of pictures of the Same Scene in different visual angles is obtained with this.With in sequence of pictures
The picture of a certain viewing angles is reference chart, using the picture of another viewing angles as target figure.For any picture in reference chart
Vegetarian refreshments finds the pixel of Corresponding matching therewith on target image, the two pixels correspond to same object in real scene
Point, this is Stereo matching.
According to Epipolar geometry knowledge, for a certain object point in space, at least its can be just determined there are two picture point
Accurate location, this is also the basic goal for carrying out Stereo matching.According to the matching between pixel, any corresponding picture can be obtained
The parallax of vegetarian refreshments obtains a disparity map, and the depth information of object point is then obtained according to triangle telemetry.Stereo matching be into
The basis of the applications such as row depth calculation, ranging, background blurring, therefore the precision of Stereo matching also contributes to the accurate of these applications
Degree.
According to the application scenario of current Stereo matching and background, Stereo Matching Algorithm is to apply to count in real time in many cases
In calculation, this requires Stereo Matching Algorithm will not only have higher accuracy rate, while also there is the faster speed of service.At present,
The Stereo Matching Algorithm of mainstream is broadly divided into three categories, local (local) matching algorithm, the overall situation (global) matching algorithm and
Half global (semi-global) matching algorithm.
Local matching method is mainly to carry out cost (cost) using the pixel in current pixel point and its neighborhood window
It calculates, principle is simple, but accuracy is relatively low, is easy to cause depth calculation mistake.Global registration algorithm can take into account current picture
Relation on vegetarian refreshments and image between all pixels point, algorithm complexity is high, very time-consuming although matching accuracy is higher.Half
Global Algorithm is to be combined local algorithm and Global Algorithm, and algorithm arithmetic speed is also very slow, strictly limits algorithm
Practical application.In practical applications, it ensure that accuracy, in order to improve operational efficiency, higher hardware configuration can be utilized, by
This adds the cost of technology application again.
The content of the invention
The present invention addresses the above problem, provides a kind of image volume segmentation method, includes the following steps：
Obtained the first image progress image segmentation will be shot, to obtain multiple images block；
Respectively characteristic point is extracted from each described image block；
The characteristic point phase with being extracted in described image block is found on the second image obtained using matching algorithm in shooting
Matching and the match point of Matching power flow value minimum；Wherein, described first image and second image be to Same Scene never
With the image of viewing angles；
According to matching result, the parallax value of all characteristic points in any image block is calculated；
Parallax value based on all characteristic points in image block is to image district all pixels in the block point assignment, with generation
Disparity map.
Optionally, first image that shooting is obtained, which carries out image segmentation, includes：Utilize half-tone information, colouring information
And any one of space length information information or much information combination carry out image segmentation to described first image.
Optionally, the method for extracting characteristic point from each described image block respectively is to described image area using operator
Block carries out feature point extraction.
Optionally, found on second image obtained using matching algorithm in shooting with being extracted in described image block
Characteristic point match and the match point of Matching power flow value minimum includes：
Set disparity range；
In the disparity range, found using matching algorithm on second image with being extracted in described image block
Characteristic point match and the match point of Matching power flow value minimum.
The parallax value of all characteristic points in any image block is calculated using equation below：
Wherein, n represents feature point number in present image block, disparityiFor the parallax value of ith feature point.
Optionally, the parallax value based on all characteristic points in image block assigns image district all pixels in the block point
Value, is included with generating disparity map：
1) use equation below calculate the parallax value of each pixel in each image block for：
Wherein, N is present image area sum of all pixels in the block；
2) parallax value based on each image district pixel in the block is to generate disparity map.
The embodiment of the present invention additionally provides a kind of mobile terminal with dual camera, including：
Image processor；
Dual camera；And memory, the memory storage have program instruction, described in the execution of described image processor
During program instruction, proceed as follows：
Obtained the first image progress image segmentation will be shot, to obtain multiple images block；
Respectively characteristic point is extracted from each described image block；
The characteristic point phase with being extracted in described image block is found on the second image obtained using matching algorithm in shooting
Matching and the match point of Matching power flow value minimum；Wherein, described first image and second image be to Same Scene never
With the image of viewing angles；
According to matching result, the parallax value of all characteristic points in any image block is calculated；
Based on the parallax value of identified image block to image district all pixels in the block point assignment to generate parallax
Figure.
Compared with prior art, technical solution of the present invention at least has the advantages that：
An embodiment of the present invention provides a kind of image volume segmentation method, this method first to the first image of shooting (as
Reference picture) piecemeal is carried out, and feature point extraction is carried out to the image block after piecemeal；Then the second image from shooting then,
The match point of Matching power flow value minimum is found on (target image).The parallax of characteristic point is recycled to count regarding for each image block
Difference, and generate disparity map.Compared with the parallax calculation method of the prior art, the parallax of all pixels point on image is no longer calculated,
It only selects representational characteristic point and carries out disparity computation.So as in the case where ensureing precision, dramatically save calculation amount,
Hardware, memory overhead are also saved simultaneously, the extensive use for Stereo matching and depth algorithm is laid a good foundation.
Description of the drawings
Fig. 1 is a kind of flow diagram of image volume segmentation method of the embodiment of the present invention.
Specific embodiment
It is understandable for the above objects, features and advantages of the present invention is enable to become apparent, below in conjunction with the accompanying drawings to the present invention
Specific embodiment be described in detail.
The present invention provides a kind of image volume segmentation methods.As shown in Figure 1, this method comprises the following steps：
Step S1：Obtained the first image progress image segmentation will be shot, to obtain multiple images block；
Step S2：Respectively characteristic point is extracted from each described image block；
Step S3：The spy with being extracted in described image block is found on the second image obtained using matching algorithm in shooting
Sign point matches and the match point of Matching power flow value minimum；Wherein, described first image and second image are to same field
The image that scape is shot with different view；
Step S4：According to matching result, the parallax value of all characteristic points in any image block is calculated；
Step S5：Based on the parallax value of identified image block to image district all pixels in the block point assignment, with
Generate disparity map.
It will be appreciated by those skilled in the art that the mobile terminal that described image Stereo matching refers to there is dual camera to configure is to same
One scene is carried out at the same time shooting or wherein single camera repeatedly shoots Same Scene in different angle, is obtained with this
Same Scene is in the image sequence of different visual angles.Using the picture of a certain viewing angles in image sequence as reference chart, with another
The picture of viewing angles is target figure.For any pixel point in reference chart, Corresponding matching therewith is found on target image
Pixel, the two picture points correspond to same object point in real scene, this is Stereo matching.In the present embodiment, by institute
It states the first image to be used as with reference to image, second image is as target image.
As described in step S1, the first image that shooting is obtained carries out image segmentation, to obtain multiple images block.Specifically
Ground can utilize different characteristic informations (for example, any one of half-tone information, colouring information and space length information letter
Breath or much information combination) image segmentation is carried out to described first image, so as to which described first image is divided into multiple images
Block.Described first image can be divided into different image blocks according to certain characteristic according to different characteristic informations.
Selection that there are many image partition methods specifically used, may be employed the marginal information based on image and is split,
Such as utilize differential operator, boundary curve fitting etc.；Method based on region can also be used to carry out image segmentation, as threshold method,
The methods of watershed；Can also the image partition method based on cluster, such as K averages (K-means), average drifting (mean
Shift) etc.；The image partition method of graph theory is also based on, as normalized cut (Normalized Cut), figure cut (graph
Cut) etc..
As described in step S2, characteristic point is extracted from each described image block respectively.
In the present embodiment, operator (such as canny operators, sobel operators, Harris operators etc.) can be utilized to the figure
As block carries out feature point extraction.Wherein, the characteristic point be concentrated mainly on described image sub-block edge or details more
Abundant region, used operator are the operators of Image Edge-Detection.
As described in step S3, found using matching algorithm on the second obtained image is shot with being carried in described image block
The characteristic point taken matches and the match point of Matching power flow value minimum.
It is being extracted after characteristic point from each image block, it is necessary to be found and each feature on second image
The match point that point matches.
Specifically, disparity range is set first.Parallax refers to (have on mobile terminal from two points for having certain distance
Dual camera at regular intervals) on to observe direction caused by same target (pixel on the image of Same Scene) poor
It is different.
Then, in the disparity range, found and described image block on second image using matching algorithm
The characteristic point of middle extraction matches and the match point of Matching power flow value minimum.
As described in step S4, according to matching result, the parallax value of all characteristic points in any image block is calculated.
The parallax value of all characteristic points in any image block is specifically calculated using equation below：
Wherein, n represents feature point number in present image block, disparityiFor the parallax value of ith feature point.
As described in step S5, the parallax value based on all characteristic points in image block is to image district all pixels in the block
Point assignment, to generate disparity map.
Specifically, for each image block, according to the parallax value of all characteristic points in the image block to the image
Area's all pixels point assignment in the block, may be employed equation below and is calculated：
Wherein, N is present image area sum of all pixels in the block.
Then, the disparity map of the parallax value generation entire image based on each image district pixel in the block.
Image solid distribution method provided in an embodiment of the present invention is described with reference to specific example.
First, (Normalized Cut, Ncut) is split to the piece image taken (as with reference to figure using normalization
Picture) it is split, the formula of image segmentation is as follows：
Cut (A, B)=∑u∈A,v∈BW (u, v),
Wherein, A, B are any two image block that separates of piece image, and u is any pixel point in image block A, v
It is weight for any pixel point, w in image block B.
The formula of normalized is as follows：
Wherein, assoc (A, V)=∑u∈A,t∈VW (u, t), u are any pixel point in image block A, and V represents view picture figure
As upper all pixels point, t is any point in image block, and w represents weight.
Then, the characteristic point of each image block is obtained using the gradient information of image, can specifically passes through kernel and figure
As carrying out convolution, the gradient information of image can be preferably obtained：
As shown in above-mentioned formula, GxAnd GyRespectively image I is in the gradient in x directions and y directions, the approximation ladder that G is image I
Degree.Using approximate gradient G's as a result, with reference to threshold value set, the characteristic point in image block can be obtained.
Further, be the precision that improves Stereo matching, avoid by large stretch of low texture region, block and left images brightness not
Matching error caused by consistent, may be employed layering confidence spread (hierarchical Belief Propagation,
HBP) global registration algorithm carries out Stereo matching and disparity computation, by finding the distance d made corresponding to Matching power flow value minimum
As required parallax.
Find dx, the value for making above-mentioned formula is minimum.Wherein, X is a certain characteristic point calculated by gradient, and Y is the neighbour of X
In domain N (X) a bit.The d to be calculatedxThe parallax of as X points.
ED,X(dx) it is that the costs of pixel X in itself calculate,For all X
Neighborhood point pass to the cost of X.
Then, for any image block, according to feature extraction as a result, calculating the parallax of all characteristic points：
Wherein, n represents the number of characteristic point in present image block, disparityiFor the parallax value of ith feature point.
Assuming that present image area sum of all pixels in the block is N, then the parallax value of each pixel is in image block：
After parallax filtering is finally completed to all pixels point in entire image, an accurate disparity map is obtained.
The embodiment of the present invention additionally provides a kind of mobile terminal with dual camera, which includes double camera shootings
Head, image processor and memory, the memory storage have program instruction, and described program is performed in described image processor
The step of embodiment as described in Figure 1 is performed during instruction, with specific reference to above-described embodiment, details are not described herein.
Although the invention has been described by way of example and in terms of the preferred embodiments, but it is not for limiting the present invention, any this field
Technical staff without departing from the spirit and scope of the present invention, may be by the methods and technical content of the disclosure above to this hair
Bright technical solution makes possible variation and modification, therefore, every content without departing from technical solution of the present invention, according to the present invention
Any simple modifications, equivalents, and modifications made to above example of technical spirit, belong to technical solution of the present invention
Protection domain.
Claims (12)
1. a kind of image solid matching method, which is characterized in that include the following steps：
Obtained the first image progress image segmentation will be shot, to obtain multiple images block；
Respectively characteristic point is extracted from each described image block；
It finds on the second image obtained using matching algorithm in shooting and matches with the characteristic point extracted in described image block
And the match point of Matching power flow value minimum；Wherein, described first image and second image are that Same Scene is regarded from difference
The image of angle shooting；
According to matching result, the parallax value of all characteristic points in any image block is calculated；
Parallax value based on all characteristic points in image block is to image district all pixels in the block point assignment, to generate parallax
Figure.
2. image solid matching method as described in claim 1, which is characterized in that it is described will shoot obtained the first image into
The segmentation of row image includes：Utilize any one of half-tone information, colouring information and space length information information or much information
Combination carries out image segmentation to described first image.
3. image solid matching method as described in claim 1, which is characterized in that carried respectively from each described image block
The method for taking characteristic point is to carry out feature point extraction to described image block using operator.
4. image solid matching method as described in claim 1, which is characterized in that described to be obtained using matching algorithm in shooting
The second image on find and match with the characteristic point extracted in described image block and the match point bag of Matching power flow value minimum
It includes：
Set disparity range；
In the disparity range, the spy with being extracted in described image block is found on second image using matching algorithm
Sign point matches and the match point of Matching power flow value minimum.
5. image solid matching method as described in claim 1, which is characterized in that any image area is calculated using equation below
The parallax value of all characteristic points in block：
<mrow>
<msub>
<mi>D</mi>
<mrow>
<mi>s</mi>
<mi>u</mi>
<mi>m</mi>
</mrow>
</msub>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msub>
<mi>disparity</mi>
<mi>i</mi>
</msub>
</mrow>
Wherein, n represents feature point number in present image block, disparityiFor the parallax value of ith feature point.
6. image solid matching method as claimed in claim 5, which is characterized in that based on all characteristic points in image block
Parallax value includes image district all pixels in the block point assignment to generate disparity map：
1) use equation below calculate the parallax value of each pixel in each image block for：
<mrow>
<mi>D</mi>
<mi>i</mi>
<mi>s</mi>
<mi>p</mi>
<mi>a</mi>
<mi>r</mi>
<mi>i</mi>
<mi>t</mi>
<mi>y</mi>
<mo>=</mo>
<mfrac>
<msub>
<mi>D</mi>
<mrow>
<mi>s</mi>
<mi>u</mi>
<mi>m</mi>
</mrow>
</msub>
<mi>N</mi>
</mfrac>
</mrow>
Wherein, N is present image area sum of all pixels in the block；
2) parallax value based on each image district pixel in the block is to generate disparity map.
7. a kind of mobile terminal with dual camera, which is characterized in that including：
Image processor；
Dual camera；And memory, the memory storage have program instruction, and described program is performed in described image processor
During instruction, proceed as follows：
Obtained the first image progress image segmentation will be shot, to obtain multiple images block；
Respectively characteristic point is extracted from each described image block；
It finds on the second image obtained using matching algorithm in shooting and matches with the characteristic point extracted in described image block
And the match point of Matching power flow value minimum；Wherein, described first image and second image are that Same Scene is regarded from difference
The image of angle shooting；
According to matching result, the parallax value of all characteristic points in any image block is calculated；
Based on the parallax value of identified image block to image district all pixels in the block point assignment to generate disparity map.
8. there is the mobile terminal of dual camera as claimed in claim 7, which is characterized in that described image processor performs institute
When stating program instruction, proceed as follows：First image that shooting is obtained, which carries out image segmentation, to be included：Believed using gray scale
Any one of breath, colouring information and space length information information or much information combination carry out image to described first image
Segmentation.
9. there is the mobile terminal of dual camera as claimed in claim 7, which is characterized in that described image processor performs institute
When stating program instruction, proceed as follows：The method for extracting characteristic point from each described image block respectively is to utilize operator
Feature point extraction is carried out to described image block.
10. there is the mobile terminal of dual camera as claimed in claim 7, which is characterized in that described image processor performs
When described program instructs, proceed as follows：It is described found using matching algorithm on obtained the second image of shooting with it is described
The characteristic point extracted in image block matches and the match point of Matching power flow value minimum includes：
Set disparity range；
In the disparity range, the spy with being extracted in described image block is found on second image using matching algorithm
Sign point matches and the match point of Matching power flow value minimum.
11. there is the mobile terminal of dual camera as claimed in claim 7, which is characterized in that described image processor performs
When described program instructs, proceed as follows：The parallax value of all characteristic points in any image block is calculated using equation below：
<mrow>
<msub>
<mi>D</mi>
<mrow>
<mi>s</mi>
<mi>u</mi>
<mi>m</mi>
</mrow>
</msub>
<mo>=</mo>
<munderover>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msub>
<mi>disparity</mi>
<mi>i</mi>
</msub>
</mrow>
Wherein, n represents feature point number in present image block, disparityiFor the parallax value of ith feature point.
12. there is the mobile terminal of dual camera as claimed in claim 11, which is characterized in that described image processor performs
When described program instructs, proceed as follows：Parallax value based on all characteristic points in image block is in the block to the image district
All pixels point assignment, is included with generating disparity map：
1) use equation below calculate the parallax value of each pixel in each image block for：
<mrow>
<mi>D</mi>
<mi>i</mi>
<mi>s</mi>
<mi>p</mi>
<mi>a</mi>
<mi>r</mi>
<mi>i</mi>
<mi>t</mi>
<mi>y</mi>
<mo>=</mo>
<mfrac>
<msub>
<mi>D</mi>
<mrow>
<mi>s</mi>
<mi>u</mi>
<mi>m</mi>
</mrow>
</msub>
<mi>N</mi>
</mfrac>
</mrow>
Wherein, N is present image area sum of all pixels in the block；
2) parallax value based on each image district pixel in the block is to generate disparity map.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711319033.2A CN108109148A (en) | 2017-12-12 | 2017-12-12 | Image solid distribution method, mobile terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711319033.2A CN108109148A (en) | 2017-12-12 | 2017-12-12 | Image solid distribution method, mobile terminal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108109148A true CN108109148A (en) | 2018-06-01 |
Family
ID=62215715
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711319033.2A Pending CN108109148A (en) | 2017-12-12 | 2017-12-12 | Image solid distribution method, mobile terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108109148A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993781A (en) * | 2019-03-28 | 2019-07-09 | 北京清微智能科技有限公司 | Based on the matched anaglyph generation method of binocular stereo vision and system |
| CN111383254A (en) * | 2018-12-29 | 2020-07-07 | Tcl集团股份有限公司 | Depth information acquisition method and system and terminal equipment |
| CN111476231A (en) * | 2020-06-22 | 2020-07-31 | 努比亚技术有限公司 | Image area identification method and device and computer readable storage medium |
| WO2020181506A1 (en) * | 2019-03-12 | 2020-09-17 | 深圳市大疆创新科技有限公司 | Image processing method, apparatus and system |
| CN113345001A (en) * | 2021-05-19 | 2021-09-03 | 智车优行科技（北京）有限公司 | Disparity map determination method and device, computer-readable storage medium and electronic device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739684B (en) * | 2009-12-17 | 2011-08-31 | 上海交通大学 | Color segmentation and pixel significance estimation-based parallax estimation method |
| CN102333234A (en) * | 2011-10-28 | 2012-01-25 | 清华大学 | Binocular stereo video state information monitoring method and device |
| CN105513064A (en) * | 2015-12-03 | 2016-04-20 | 浙江万里学院 | Image segmentation and adaptive weighting-based stereo matching method |
| CN107220997A (en) * | 2017-05-22 | 2017-09-29 | 成都通甲优博科技有限责任公司 | A kind of solid matching method and system |
-
2017
- 2017-12-12 CN CN201711319033.2A patent/CN108109148A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739684B (en) * | 2009-12-17 | 2011-08-31 | 上海交通大学 | Color segmentation and pixel significance estimation-based parallax estimation method |
| CN102333234A (en) * | 2011-10-28 | 2012-01-25 | 清华大学 | Binocular stereo video state information monitoring method and device |
| CN105513064A (en) * | 2015-12-03 | 2016-04-20 | 浙江万里学院 | Image segmentation and adaptive weighting-based stereo matching method |
| CN107220997A (en) * | 2017-05-22 | 2017-09-29 | 成都通甲优博科技有限责任公司 | A kind of solid matching method and system |
Non-Patent Citations (4)
| Title |
|---|
| P.F. FELZENSZWALB等: "Efficient belief propagation for early vision", 《PROCEEDINGS OF THE 2004 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, 2004》 * |
| 董慧颖: "《典型目标识别与图像除雾技术》", 31 October 2016, 国防工业出版社 * |
| 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》 * |
| 马帅: "一种改进的置信度传输提取视差图算法的研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111383254A (en) * | 2018-12-29 | 2020-07-07 | Tcl集团股份有限公司 | Depth information acquisition method and system and terminal equipment |
| WO2020181506A1 (en) * | 2019-03-12 | 2020-09-17 | 深圳市大疆创新科技有限公司 | Image processing method, apparatus and system |
| CN109993781A (en) * | 2019-03-28 | 2019-07-09 | 北京清微智能科技有限公司 | Based on the matched anaglyph generation method of binocular stereo vision and system |
| CN111476231A (en) * | 2020-06-22 | 2020-07-31 | 努比亚技术有限公司 | Image area identification method and device and computer readable storage medium |
| CN111476231B (en) * | 2020-06-22 | 2024-01-12 | 努比亚技术有限公司 | Image area identification method, device and computer readable storage medium |
| CN113345001A (en) * | 2021-05-19 | 2021-09-03 | 智车优行科技（北京）有限公司 | Disparity map determination method and device, computer-readable storage medium and electronic device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11562498B2 (en) | Systems and methods for hybrid depth regularization | |
| CN108109148A (en) | Image solid distribution method, mobile terminal | |
| CN106940704B (en) | Positioning method and device based on grid map | |
| US10552973B2 (en) | Image vision processing method, device and equipment | |
| KR101923845B1 (en) | Image processing method and apparatus | |
| CN112085844B (en) | Unmanned aerial vehicle image rapid three-dimensional reconstruction method for field unknown environment | |
| JP5306652B2 (en) | Integrated image processor | |
| CN105005964B (en) | Geographic scenes panorama sketch rapid generation based on video sequence image | |
| CN110276768B (en) | Image segmentation method, image segmentation device, image segmentation apparatus, and medium | |
| CN105277169B (en) | Binocular distance-finding method based on image segmentation | |
| JP2009139995A (en) | Unit and program for real time pixel matching in stereo image pair | |
| US9406140B2 (en) | Method and apparatus for generating depth information | |
| CN108460792B (en) | Efficient focusing stereo matching method based on image segmentation | |
| CN107341804B (en) | Method and device for determining plane in point cloud data, and method and equipment for image superposition | |
| CN107560592A (en) | A kind of precision ranging method for optronic tracker linkage target | |
| CN115035235A (en) | Three-dimensional reconstruction method and device | |
| KR100943635B1 (en) | Method and apparatus for generating disparity map using digital camera image | |
| CN107133977A (en) | A kind of quick stereo matching process that model is produced based on probability | |
| CN106412441A (en) | Video anti-shake control method and terminal | |
| CN110567441A (en) | Particle filter-based positioning method, positioning device, mapping and positioning method | |
| CN108874269B (en) | Target tracking method, device and system | |
| CN112102504A (en) | Three-dimensional scene and two-dimensional image mixing method based on mixed reality | |
| CN113225484B (en) | Method and device for rapidly acquiring high-definition picture shielding non-target foreground | |
| Fu et al. | Image Stitching Techniques Applied to Plane or 3D Models: A Review | |
| US10430971B2 (en) | Parallax calculating apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| 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: 20180601 |