CN106846249A - A kind of panoramic video joining method - Google Patents
A kind of panoramic video joining method Download PDFInfo
- Publication number
- CN106846249A CN106846249A CN201710047613.4A CN201710047613A CN106846249A CN 106846249 A CN106846249 A CN 106846249A CN 201710047613 A CN201710047613 A CN 201710047613A CN 106846249 A CN106846249 A CN 106846249A
- Authority
- CN
- China
- Prior art keywords
- image
- video
- template
- carried out
- panoramic
- 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
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000013507 mapping Methods 0.000 claims abstract description 27
- 238000012937 correction Methods 0.000 claims abstract description 24
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 16
- 238000000605 extraction Methods 0.000 claims abstract description 13
- 230000004927 fusion Effects 0.000 claims description 12
- 239000013598 vector Substances 0.000 claims description 6
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 241000251468 Actinopterygii Species 0.000 abstract description 5
- 230000001172 regenerating effect Effects 0.000 abstract 1
- 230000001629 suppression Effects 0.000 description 2
- HPTJABJPZMULFH-UHFFFAOYSA-N 12-[(Cyclohexylcarbamoyl)amino]dodecanoic acid Chemical compound OC(=O)CCCCCCCCCCCNC(=O)NC1CCCCC1 HPTJABJPZMULFH-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Studio Devices (AREA)
Abstract
The invention discloses a kind of panoramic video joining method, obtain one group of image and generate template；The each frame of video is carried out into space reflection using template；Frame of video after mapping is rendered to panoramic video, wherein, generation template is comprised the following steps：Distortion correction is carried out to image；Feature point extraction and Feature Points Matching are carried out to the image after correction；Using optimized algorithm and Feature Points Matching result estimation space mapping parameters, mapping parameters are preserved into template.In order that the characteristic point extracted is more, more accurate, the present invention eliminates the distortion phenomenon of panorama fish eye camera by distortion correction；Subsequent video images are directly used after generation template, without regenerating, lifting splicing efficiency.
Description
Technical field
The invention belongs to technical field of video processing, it is related to a kind of panoramic video joining method.
Background technology
Video-splicing technology refer to by several camera acquisitions to video image be spliced to form a width panoramic picture
Technology, conventional video-splicing technology is mostly based on the coincidence part found out in adjacent video image in merging algorithm for images at present
Point carry out conversion splicing, but this kind of method changed by shooting scene, the shadow of the factor such as the different and stitching algorithm of shooting angle
Ring.And panoramic camera is fish eye lens, the image for acquiring is fish eye images, there is distortion, if directly carrying out spy
An extraction is levied, then the Feature Points Matching degree extracted is very low, and error hiding rate is higher.
The content of the invention
To solve the above problems, it is an object of the invention to provide a kind of panoramic video joining method.
To achieve the above object, the invention provides a kind of panoramic video joining method, comprise the following steps：
Obtain one group of image and generate template；
The each frame of video is carried out into space reflection using template；
Frame of video after mapping is rendered to panoramic video,
Wherein, generation template is comprised the following steps：
Distortion correction is carried out to image；
Feature point extraction and Feature Points Matching are carried out to the image after correction；
Using optimized algorithm and Feature Points Matching result estimation space mapping parameters, mapping parameters are preserved into template.
Preferably, it is described that distortion correction is carried out to image, using longitude and latitude correction method, on a longitude for determination, figure
As upper elliptic equation is：
Coordinate after longitude and latitude correction is (x1, y1), and correction relationship is：
Preferably, the feature point extraction is comprised the following steps：
Metric space is built with gaussian pyramid method, convolution several times and down-sampling is carried out to the image after correction,
Obtain scale space images；
Scale space images to building are scanned for, and find Local modulus maxima as preliminary key point, Ran Houtong
Maximum suppression is crossed, it is determined that final key point, i.e. characteristic point；
By to image-region piecemeal around characteristic point, calculating every piece of inside gradient histogram, then generated with these histograms
One 128 dimensional vector, that is, carry out feature point description.
Preferably, the convolution, image after convolution is obtained using equation below：
L (x, y, σ)=I (x, y) * G (x, y, σ),
Wherein I (x, y) is original image, and G (x, y, σ) is Gaussian function, and L (x, y, σ) is image after convolution.
Preferably, the Feature Points Matching, selects a characteristic point first from an image, is retouched according to its characteristic point
State, searched out from another image and describe most like feature point description with this feature point, that is, form a pair of features of matching
Point.
Preferably, the use optimized algorithm and Feature Points Matching result estimation space mapping parameters, it is existing some right
The characteristic point of matching, by optimized algorithm estimation space mapping parameters, including rotation amount, translational movement and amount of distortion, these is joined
Number preserves into template.
Preferably, the optimized algorithm is least square method or LevenBerg-Marquardt algorithms.
Preferably, each frame of video is carried out space reflection by the use template, is the sky in the template for being generated
Between mapping parameters, each two field picture of video is projected on panoramic plane.
Preferably, the frame of video by after mapping is rendered to panoramic video, is that, by the linear fusion of image, generation is complete
Scape video.
Preferably, the linear fusion, is to carry out linear weighted function to some images, is comprised the following steps：
When mapping an image to panoramic picture, the weight coefficient w of each effective pixel points, the weight coefficient of dead space are generated
It is 0；
There is overlapping region between different camera lenses, that is, there is the same point during different width images are mapped to panoramic picture, entirely
The pixel value of scape image is obtained by equation below：
Wherein, N is the same point in having N number of image to be mapped to panoramic picture, In(x, y) is the pixel of n-th image
Value, wn(x, y) is weight of the n-th image in point (x, y).
Preferably, the frame of video by after mapping is rendered to panoramic video, is by the multiple fusion or seamless of image
Fusion or graph cut, generate panoramic video.
Beneficial effects of the present invention are as follows：Because panoramic camera is fish eye lens, the image for acquiring is fish-eye image
, there is distortion in picture, if directly carrying out feature point extraction, the Feature Points Matching degree extracted is very low, and error hiding rate compared with
It is high.The present invention obtains more accurately Feature Points Matching pair to extract more characteristic points, and fish eye images first are entered into line distortion school
Just, the distortion phenomenon of flake is eliminated；Then extracting and matching feature points are carried out；Template is generated with estimation space mapping parameters.
Subsequent video images directly can be spliced using the template kept, lifting splicing efficiency.
Brief description of the drawings
The step of Fig. 1 is the panoramic video joining method of a specific embodiment of the invention flow chart；
The step of Fig. 2 is the panoramic video joining method of still another embodiment of the present invention flow chart.
Specific embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Conversely, the present invention covers any replacement done in spirit and scope of the invention being defined by the claims, repaiies
Change, equivalent method and scheme.Further, in order that the public has a better understanding to the present invention, below to of the invention thin
It is detailed to describe some specific detail sections in section description.Part without these details for a person skilled in the art
Description can also completely understand the present invention.
Embodiment 1
Referring to Fig. 1, flow chart the step of be the panoramic video joining method of a specific embodiment of the invention, including following step
Suddenly：
S10, obtains one group of image and generates template；
S20, space reflection is carried out using template by each frame of video；
S30, panoramic video is rendered to by the frame of video after mapping.
By above-mentioned steps, one group of image to be spliced can be generated template, the template of generation is one group of parameter, is circumference
Fish eye images are mapped to the mapping parameters of final equidistant cylinder image (i.e. panoramic picture).Can be by each of video using template
Two field picture carries out space reflection, is then rendered to panoramic video.The video that the template of generation can shoot to follow-up panoramic camera
Image is directly spliced, and substantially increases efficiency.
Embodiment 2
Referring to Fig. 2, flow chart the step of be the panoramic video joining method of a specific embodiment of the invention, including following step
Suddenly：
S101, distortion correction is carried out to image；
S102, feature point extraction and Feature Points Matching are carried out to the image after correction；
S103, using optimized algorithm and Feature Points Matching result estimation space mapping parameters, mapping parameters is preserved into
Template；
S20, space reflection is carried out using template by each frame of video；
S30, panoramic video is rendered to by the frame of video after mapping.
In above-mentioned steps, S101 carries out distortion correction to image, using longitude and latitude correction method, in a longitude for determination
On, the elliptic equation of any is on image：
Coordinate after longitude and latitude correction is (x1, y1), and correction relationship is：
After image distortion correction, the feature point extraction in S102 is carried out.The method of feature point extraction has a lot, such as edge
Point, harris angle points, SURF characteristic points, ORB features etc., the present invention uses Scale invariant features transform, comprises the following steps：
Metric space is built with gaussian pyramid method, convolution several times and down-sampling is carried out to the image after correction,
Obtain scale space images；
Image after convolution is obtained using equation below：
L (x, y, σ)=I (x, y) * G (x, y, σ),
Wherein I (x, y) is original image, and G (x, y, σ) is Gaussian function, and L (x, y, σ) is image after convolution；
Scale space images to building are scanned for, and find Local modulus maxima as preliminary key point, Ran Houtong
Maximum suppression is crossed, it is determined that final key point, i.e. characteristic point；
By to image-region piecemeal around characteristic point, calculating every piece of inside gradient histogram, then generated with these histograms
One 128 dimensional vector, that is, carry out feature point description, and this vector is a kind of abstract of the regional image information, with uniqueness.
Features described above point is extracted and ensure that in image rotation, image scaling, brightness of image even variation all without shadow
The result of feature point extraction is rung, is a kind of sufficiently stable feature point extraction algorithm.
Complete image feature point extraction after, carry out the Feature Points Matching in S102, so as to set up different images between
Corresponding relation.A characteristic point is selected first from an image, according to its feature point description, is searched out from another image
Most like feature point description is described with this feature point, that is, forms a pair of characteristic points of matching.In a particular embodiment, such as one
One feature point coordinates of image is (x1,y1), its description value is vectorWith vector in another imageMost like
Feature point description isThis feature point describes corresponding feature point coordinates for (x2,y2), then (x1,y1) and (x2,y2) it is exactly one
To Feature Points Matching pair, ((x can be designated as1,y1),(x2,y2)).The characteristic point of a pair of matchings is considered same in a scene
One point, simply the scene is by different shot by camera.
S103 is then carried out, with some characteristic points to matching, by optimized algorithm estimation space mapping parameters, space is reflected
Penetrating parameter includes rotation amount, translational movement and amount of distortion, and these parameters are preserved into template.
Optimized algorithm is least square method or LevenBerg-Marquardt algorithms in specific embodiment, using a most young waiter in a wineshop or an inn
During multiplication, p=(x, y) is made, represent a coordinate points, then matching pair can again be expressed as (p1,p2), space reflection pattern function
Y=f (p, β) can be expressed as.Wherein β=(β1,β2,...βm) be space reflection parameter, including rotation parameter, translation parameters with
And distortion parameter.The optimisation strategy of least square method is to make the quadratic sum of error minimum, that is, finding a solution β makes following formula minimum：
Its optimal solution is：
Wherein
After obtaining space reflection parameter using optimized algorithm, these parameters are preserved into template, follow-up video camera shooting figure
As after, it is not necessary to carry out the generation template procedure of complexity again, directly spliced using the template kept, lifting splicing effect
Rate.
After generation template, S20 is carried out, each frame of video is carried out into space reflection using template, be according to the mould for being generated
Space reflection parameter in plate, each two field picture of video is projected on panoramic plane.If a video camera has 4 mirrors
Head, i.e., will carry out panoramic mosaic, then can obtain 4 Zhang Quanjing images by space reflection, then carry out S30 to 4 images, lead to
The linear fusion of image is crossed, these panoramic pictures are rendered, be fused into the final panoramic picture of a frame.
The linear fusion, is to carry out linear weighted function to some images, is comprised the following steps：
When mapping an image to panoramic picture, the weight coefficient w of each effective pixel points, the weight coefficient of dead space are generated
It is 0；
There is overlapping region between different camera lenses, that is, there is the same point during different width images are mapped to panoramic picture, entirely
The pixel value of scape image is obtained by equation below：
Wherein, N is the same point in having N number of image to be mapped to panoramic picture, In(x, y) is the pixel of n-th image
Value, wn(x, y) is weight of the n-th image in point (x, y).
In specific embodiment, aphorama can also be generated by the multiple fusion of image or seamless fusion or graph cut
Frequently.It should be noted that when being rendered to video, data volume is very big, the especially video of high resolution.The now meter of CPU
Calculation ability far can not meet demand.In order to improve efficiency, it is necessary to which, using the strong platform of computing capability, the present invention is used
GPU is rendered to video, and GPU has very powerful computation capability, it is adaptable to image rendering, greatly promotes efficiency,
GPU is rendered can be realized using CUDA or OpenGL.
Presently preferred embodiments of the present invention is the foregoing is only, is not intended to limit the invention, it is all in essence of the invention
Any modification, equivalent and improvement made within god and principle etc., should be included within the scope of the present invention.
Claims (11)
1. a kind of panoramic video joining method, it is characterised in that comprise the following steps：
Obtain one group of image and generate template；
The each frame of video is carried out into space reflection using template；
Frame of video after mapping is rendered to panoramic video,
Wherein, generation template is comprised the following steps：
Distortion correction is carried out to image；
Feature point extraction and Feature Points Matching are carried out to the image after correction；
Using optimized algorithm and Feature Points Matching result estimation space mapping parameters, mapping parameters are preserved into template.
2. method according to claim 1, it is characterised in that described that distortion correction is carried out to image, using longitude and latitude school
Execute, on a longitude for determination, the elliptic equation of any is on image：
Coordinate after longitude and latitude correction is (x1, y1), and correction relationship is：
3. method according to claim 1, it is characterised in that the feature point extraction is comprised the following steps：
Metric space is built with gaussian pyramid method, convolution several times and down-sampling are carried out to the image after correction, obtained
Scale space images；
Scale space images to building are scanned for, and Local modulus maxima are found as preliminary key point, then by pole
Big value suppresses, it is determined that final key point, i.e. characteristic point；
By to image-region piecemeal around characteristic point, calculating every piece of inside gradient histogram, then one is generated with these histograms
128 dimensional vectors, that is, carry out feature point description.
4. method according to claim 3, it is characterised in that the convolution, image after convolution is obtained using equation below：
L (x, y, σ)=I (x, y) * G (x, y, σ),
Wherein I (x, y) is original image, and G (x, y, σ) is Gaussian function, and L (x, y, σ) is image after convolution.
5. method according to claim 3, it is characterised in that the Feature Points Matching, selects first from an image
One characteristic point, according to its feature point description, searches out from another image and describes most like characteristic point with this feature point
Description, that is, form a pair of characteristic points of matching.
6. method according to claim 1, it is characterised in that the use optimized algorithm and Feature Points Matching result are estimated
Meter space reflection parameter, is with some characteristic points to matching, by optimized algorithm estimation space mapping parameters, including rotation
These parameters are preserved into template by amount, translational movement and amount of distortion.
7. method according to claim 1, it is characterised in that the optimized algorithm is least square method or LevenBerg-
Marquardt algorithms.
8. method according to claim 1, it is characterised in that each frame of video is carried out space and reflected by the use template
Penetrate, be the space reflection parameter in the template for being generated, each two field picture of video is projected on panoramic plane.
9. method according to claim 1, it is characterised in that the frame of video by after mapping is rendered to panoramic video,
It is, by the linear fusion of image, to generate panoramic video.
10. method according to claim 9, it is characterised in that the linear fusion, be some images are carried out linearly plus
Power, comprises the following steps：
When mapping an image to panoramic picture, the weight coefficient w of each effective pixel points is generated, the weight coefficient of dead space is 0；
There is overlapping region between different camera lenses, that is, there is the same point during different width images are mapped to panoramic picture, panorama sketch
The pixel value of picture is obtained by equation below：
Wherein, N is the same point in having N number of image to be mapped to panoramic picture, In(x, y) is the pixel value of n-th image, wn
(x, y) is weight of the n-th image in point (x, y).
11. methods according to claim 1, it is characterised in that the frame of video by after mapping is rendered to panoramic video,
It is the multiple fusion by image or seamless fusion or graph cut, generates panoramic video.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710047613.4A CN106846249A (en) | 2017-01-22 | 2017-01-22 | A kind of panoramic video joining method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710047613.4A CN106846249A (en) | 2017-01-22 | 2017-01-22 | A kind of panoramic video joining method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106846249A true CN106846249A (en) | 2017-06-13 |
Family
ID=59120032
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710047613.4A Pending CN106846249A (en) | 2017-01-22 | 2017-01-22 | A kind of panoramic video joining method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106846249A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108833874A (en) * | 2018-07-04 | 2018-11-16 | 长安大学 | A kind of panoramic picture color correcting method for automobile data recorder |
| CN109272442A (en) * | 2018-09-27 | 2019-01-25 | 百度在线网络技术（北京）有限公司 | Processing method, device, equipment and the storage medium of panorama spherical surface image |
| CN109697705A (en) * | 2018-12-24 | 2019-04-30 | 北京天睿空间科技股份有限公司 | Chromatic aberration correction method suitable for video-splicing |
| CN110278405A (en) * | 2018-03-18 | 2019-09-24 | 北京图森未来科技有限公司 | A kind of lateral image processing method of automatic driving vehicle, device and system |
| CN111507902A (en) * | 2020-04-15 | 2020-08-07 | 京东城市(北京)数字科技有限公司 | High-resolution image acquisition method and device |
| CN111681190A (en) * | 2020-06-18 | 2020-09-18 | 深圳天海宸光科技有限公司 | High-precision coordinate mapping method for panoramic video |
| CN112437327A (en) * | 2020-11-23 | 2021-03-02 | 北京瞰瞰科技有限公司 | Real-time panoramic live broadcast splicing method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103971352A (en) * | 2014-04-18 | 2014-08-06 | 华南理工大学 | Rapid image splicing method based on wide-angle lenses |
| CN105957008A (en) * | 2016-05-10 | 2016-09-21 | 厦门美图之家科技有限公司 | Panoramic image real-time stitching method and panoramic image real-time stitching system based on mobile terminal |
| CN106056539A (en) * | 2016-06-24 | 2016-10-26 | 中国南方电网有限责任公司 | Panoramic video splicing method |
-
2017
- 2017-01-22 CN CN201710047613.4A patent/CN106846249A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103971352A (en) * | 2014-04-18 | 2014-08-06 | 华南理工大学 | Rapid image splicing method based on wide-angle lenses |
| CN105957008A (en) * | 2016-05-10 | 2016-09-21 | 厦门美图之家科技有限公司 | Panoramic image real-time stitching method and panoramic image real-time stitching system based on mobile terminal |
| CN106056539A (en) * | 2016-06-24 | 2016-10-26 | 中国南方电网有限责任公司 | Panoramic video splicing method |
Non-Patent Citations (1)
| Title |
|---|
| 廖训佚: "鱼眼图像全景拼接系统", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110278405A (en) * | 2018-03-18 | 2019-09-24 | 北京图森未来科技有限公司 | A kind of lateral image processing method of automatic driving vehicle, device and system |
| CN110278405B (en) * | 2018-03-18 | 2021-06-08 | 北京图森未来科技有限公司 | Method, device and system for processing lateral image of automatic driving vehicle |
| CN108833874A (en) * | 2018-07-04 | 2018-11-16 | 长安大学 | A kind of panoramic picture color correcting method for automobile data recorder |
| CN108833874B (en) * | 2018-07-04 | 2020-11-03 | 长安大学 | Panoramic image color correction method for automobile data recorder |
| CN109272442A (en) * | 2018-09-27 | 2019-01-25 | 百度在线网络技术（北京）有限公司 | Processing method, device, equipment and the storage medium of panorama spherical surface image |
| CN109697705A (en) * | 2018-12-24 | 2019-04-30 | 北京天睿空间科技股份有限公司 | Chromatic aberration correction method suitable for video-splicing |
| CN109697705B (en) * | 2018-12-24 | 2019-09-03 | 北京天睿空间科技股份有限公司 | Chromatic aberration correction method suitable for video-splicing |
| CN111507902A (en) * | 2020-04-15 | 2020-08-07 | 京东城市(北京)数字科技有限公司 | High-resolution image acquisition method and device |
| CN111507902B (en) * | 2020-04-15 | 2023-09-26 | 京东城市(北京)数字科技有限公司 | High-resolution image acquisition method and device |
| CN111681190A (en) * | 2020-06-18 | 2020-09-18 | 深圳天海宸光科技有限公司 | High-precision coordinate mapping method for panoramic video |
| CN112437327A (en) * | 2020-11-23 | 2021-03-02 | 北京瞰瞰科技有限公司 | Real-time panoramic live broadcast splicing method and system |
| CN112437327B (en) * | 2020-11-23 | 2023-05-16 | 瞰瞰技术(深圳)有限公司 | Real-time panoramic live broadcast splicing method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106846249A (en) | A kind of panoramic video joining method | |
| KR102319177B1 (en) | Method and apparatus, equipment, and storage medium for determining object pose in an image | |
| US11457138B2 (en) | Method and device for image processing, method for training object detection model | |
| CN106651938B (en) | A kind of depth map Enhancement Method merging high-resolution colour picture | |
| Zhang et al. | An image stitching algorithm based on histogram matching and SIFT algorithm | |
| CN103325112B (en) | Moving target method for quick in dynamic scene | |
| KR101643607B1 (en) | Method and apparatus for generating of image data | |
| CN104966270B (en) | A kind of more image split-joint methods | |
| US20190057513A1 (en) | Systems and Methods for Hybrid Depth Regularization | |
| CN110070598B (en) | Mobile terminal for 3D scanning reconstruction and 3D scanning reconstruction method thereof | |
| CN110211043A (en) | A kind of method for registering based on grid optimization for Panorama Mosaic | |
| WO2021017588A1 (en) | Fourier spectrum extraction-based image fusion method | |
| TW202117611A (en) | Computer vision training system and method for training computer vision system | |
| CN108460792B (en) | Efficient focusing stereo matching method based on image segmentation | |
| CN110956661A (en) | Method for calculating dynamic pose of visible light and infrared camera based on bidirectional homography matrix | |
| CN103080979A (en) | System and method for synthesizing portrait sketch from photo | |
| Luo et al. | Wavelet synthesis net for disparity estimation to synthesize dslr calibre bokeh effect on smartphones | |
| Liu et al. | High quality depth map estimation of object surface from light-field images | |
| CN112016478A (en) | Complex scene identification method and system based on multispectral image fusion | |
| CN114519772A (en) | Three-dimensional reconstruction method and system based on sparse point cloud and cost aggregation | |
| CN116579962A (en) | Panoramic sensing method, device, equipment and medium based on fisheye camera | |
| Zhou et al. | NeRFLix: High-quality neural view synthesis by learning a degradation-driven inter-viewpoint mixer | |
| CN113628134B (en) | Image noise reduction method and device, electronic equipment and storage medium | |
| KR101921608B1 (en) | Apparatus and method for generating depth information | |
| Ajith et al. | Dark channel prior based single image dehazing of daylight captures |
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 |
Application publication date: 20170613 |
|
| RJ01 | Rejection of invention patent application after publication |