CN107330856A - A kind of method for panoramic imaging based on projective transformation and thin plate spline - Google Patents
A kind of method for panoramic imaging based on projective transformation and thin plate spline Download PDFInfo
- Publication number
- CN107330856A CN107330856A CN201710459891.0A CN201710459891A CN107330856A CN 107330856 A CN107330856 A CN 107330856A CN 201710459891 A CN201710459891 A CN 201710459891A CN 107330856 A CN107330856 A CN 107330856A
- Authority
- CN
- China
- Prior art keywords
- image
- msub
- mrow
- thin plate
- projective transformation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000009466 transformation Effects 0.000 title claims abstract description 31
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000003384 imaging method Methods 0.000 title claims abstract description 13
- 239000011159 matrix material Substances 0.000 claims abstract description 14
- 239000013598 vector Substances 0.000 claims description 14
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 3
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 2
- 238000012790 confirmation Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000000638 solvent extraction 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 transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4038—Scaling the whole image or part thereof for image mosaicing, i.e. plane images composed of plane sub-images
-
- G06T3/14—
-
- G06T3/18—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/20—Linear translation of a whole image or part thereof, e.g. panning
-
- 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
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Abstract
The invention discloses a kind of method for panoramic imaging based on projective transformation and thin plate spline, three width images are gathered first;Then image characteristic point is extracted using sift algorithms, verified using brute force searching algorithms combining ratio and the characteristic point of adjacent image is matched, characteristic matching is screened using many structure fitting algorithms, fit two areal models of ground level and distant view, by the characteristic point of each image be divided into above and below two parts;Then projective transformation matrix is estimated using DLT algorithms to the characteristic matching of part in adjacent two images, the characteristic point of part under second image is respectively mapped in the first, the 3rd image lower molecules image, and deformation process is carried out to the first, the 3rd image based on thin plate spline function;First after deformation, the 3rd image are transformed in the coordinate system of the second image finally according to projection matrix, panorama sketch is obtained.The present invention carries out panoramic mosaic to the image comprising biplane, improves the alignment capabilities of image, alleviates projection distortion.
Description
Technical field
The invention belongs to computer vision and image processing field, and in particular to one kind is based on projective transformation and thin plate spline
Method for panoramic imaging
Background technology
Outdoor Scene image mosaic is widely used in terms of ground on-vehicle panorama, tourist attraction photography, streetscape car.
The Normal practice for splicing this kind of Outdoor Scene is to gather image using camera rotating model, is alignd and schemed using projective transformation model
Picture, the problem of this Normal practice is present be:Parallax is easily introduced when the photocentre of camera shifts, using projective transformation
Model alignment capabilities are not enough.
Combine partial transformation more in order to solve the problems, such as the image mosaic under the inconsistent situation of photocentre, prior art and image becomes
The method of shape, such as APAP (As-Projective-As-Possible) algorithm and DHW (Dual-Homography
Warping) algorithm.DHW algorithms need, according to two-part characteristic matching above and below image, two homographs to be assessed respectively, but
It is the lower part usually road surface of Outdoor Scene image, texture information is less, single strain is assessed according to the characteristic matching of this part
Change robustness not high.Although APAP algorithms optimize local alignment ability, but remain global homography constraint, panoramic mosaic knot
The problem of fruit still has projection distortion and poor quality of alignment.
The content of the invention
It is double to solve it is an object of the invention to provide a kind of method for panoramic imaging based on projective transformation and thin plate spline
The problem of projection distortion present in plane scene image mosaic is excessive and quality of alignment is poor.
The solution for realizing the object of the invention is:A kind of method for panoramic imaging based on projective transformation and thin plate spline,
Comprise the following steps:
Step 1, three width images are from left to right gathered, be followed successively by the first image, the second image and the 3rd image, adjacent two width
Image exists in the horizontal direction to partly overlap;
Step 2, characteristic point extracted to each image using sift algorithms, and generate feature description vectors;
Step 3, the characteristic point progress using brute-force searching algorithm combining ratio verification methods to adjacent image
Match somebody with somebody;
Step 4, screened using many structure fitting algorithms are further to the characteristic matching of adjacent two images, fitted
Two areal models of ground level and distant view;
Step 5, a horizontal boundary is determined according to the areal model of fitting, by the characteristic point of each image be divided into above and below two
Part, obtain above and below two subgraphs;
Step 6, using direct linear transformation (DLT) algorithm is estimated to the characteristic matching of part in adjacent two images
The projective transformation matrix of one image, the 3rd image to the second image, and according to projective transformation matrix by part under the second image
Characteristic point is respectively mapped in the first, the 3rd image lower molecules image;
Step 7, using the characteristic point and mapping point of the first, the 3rd image lower molecules image as thin plate spline deform
Front and rear control point, deformation process is carried out based on thin plate spline function to the first, the 3rd image;
Step 8, according to projection matrix first after deformation, the 3rd image are transformed in the coordinate system of the second image, made
Obtain two parts above and below 3 width images and be obtained for alignment, obtain panorama sketch.
There is 20%-60% overlapping region in step 1 between adjacent image.
Step 3 carry out Feature Points Matching specific method be:
Step 3.1, the feature description vectors set to adjacent two images are done European in searching loop, selection two images
Distance is less than the characteristic vector of distance threshold to being used as unidirectional matching candidate;
Step 3.2, two arest neighbors description vectors for concentrating from unidirectional matching candidate each characteristic vector of selection, only when the
When the Euclidean distance of one neighborhood matching and the ratio of the second neighbour are less than fractional threshold, then confirm it is correct matching, otherwise really
It is wrong matching to recognize, and is rejected.
Many structure fitting algorithms of step 4 use the information guidance of residual error sequence acquisition and accelerate to assume sampling process.
The DLT algorithmic notations of step 6 are as follows:
AiH=0
Wherein,X and y, u and v are two images respectively
The transverse and longitudinal coordinate of middle characteristic point, h=(h1,h2,h3,h4,h5,h6,h7,h8,h9)TRepresent the element of projective transformation matrix.
Thin plate spline transforming function transformation function represents as follows in step 7:
In formula, φ (ri)=ri 2log riFor RBF, whereinX and y are respectively
The transverse and longitudinal coordinate of pixel, x in two imagesiAnd yiIt is the transverse and longitudinal coordinate at control point, w respectively1iAnd w2iRepresent RBF
Weights;a11、a12、a21、a22、b1、b2It is 6 affine coefficients, n is the number at control point.
Compared with prior art, its remarkable advantage is the present invention:The inventive method is for including distant view plane and ground level
Scene carry out panoramic mosaic, the homography matrix that distant view planar section is estimated using the characteristic matching of the part is become
Change, obtained the effect of " protecting straightforward ", reduced projection distortion;Planar section employs the algorithm of thin-plate spline interpolation over the ground,
Need less characteristic point just to obtain preferable alignment effect, agree with the characteristics of ground texture information is few;The inventive method is carried
The high alignment capabilities of image, alleviate projection distortion.
Brief description of the drawings
Fig. 1 is flow chart of the method for the present invention.
Fig. 2 is the multi-source input picture of the present invention.
Fig. 3 is the result figure of the characteristic matching screening and segmentation of the present invention.
Fig. 4 is the thin plate spline anamorphose schematic diagram of the present invention.
Fig. 5 is the global projective transformation result figure of the present invention.
Fig. 6 is the splicing figure of the adjacent two images of the present invention.
Fig. 7 is the panoramic design sketch of the present invention.
Embodiment
Below in conjunction with the accompanying drawings and specific embodiment, the present invention program is described further.
As shown in figure 1, the method for panoramic imaging based on projective transformation and thin plate spline, using intermediate image as reference data,
Left and right sides image difference is aligned, generates panorama sketch, comprises the following steps:
Step 1, camera do horizontal revolving motion, from left to right gather three width images, are followed successively by the first image, the second image
With the 3rd image, each image be resolution ratio be 730 × 487 3 channel images.As shown in Fig. 2 image contains ground letter
There is 50% and 57% overlapping region respectively between breath, the image of first and second image, second and the 3rd.In order to produce area
Point result, the photocentre position of the camera of photographed scene allows have certain displacement, this broken splicing scene have to comply with it is single
The hypothesis of homography conversion.
Step 2, the Corner Feature for extracting using sift feature extraction algorithms 3 width images respectively;
Step 3, the initial matching added using brute-force algorithms between ratio checking acquisition adjacent image, specifically
Method is:
Step 3.1, the feature description vectors set to adjacent two images are done European in searching loop, selection two images
Distance is less than the characteristic vector of distance threshold to being used as unidirectional matching candidate;
Step 3.2, two arest neighbors description vectors for concentrating from unidirectional matching candidate each characteristic vector of selection, only when the
When the Euclidean distance of one neighborhood matching and the ratio of the second neighbour are less than fractional threshold, then confirm it is correct matching, otherwise really
It is wrong matching to recognize, and is rejected.
Step 4, in order to screen out point not in the know, while retain be not belonging to same model but correct match point, using many structures
Fitting algorithm is filtered to the characteristic matching of adjacent two images, fits two areal models of ground level and distant view.Many knots
Structure fitting algorithm uses the information guidance of residual error sequence acquisition and accelerates to assume sampling process.Second image and the 3rd image pass through
Characteristic matching result after many structure fitting algorithm filterings is as shown in figure 3, the characteristic matching of its medium long shot plane has 42 pairs, Horizon
The characteristic matching in face has 21 pairs.
Step 5, the corresponding interior point of two different homography conversion models is obtained according to many structure fitting algorithms, by match point
Two parts above and below being divided into, and partitioning boundary above and below image delimited accordingly.
The characteristic point of part in step 6, selected digital image, the first image, the matching on the 3rd image are calculated with DLT algorithms
The homography matrix H1 and H2 of matching point set on pointto-set map to the second image.
DLT algorithmic notations are as follows:
AiH=0
Wherein,X and y, u and v are two images respectively
The transverse and longitudinal coordinate of middle characteristic point, h=(h1,h2,h3,h4,h5,h6,h7,h8,h9)TRepresent the element of projective transformation matrix.
H1 and H2 result is respectively:
Step 7, road pavement scene do the non-rigid registration splicing based on thin plate spline function, from given sparse features
With the dense correspondence for obtaining two images.In order to retain global homography constraint, according to two projective transformation matrixs, by the second figure
As the characteristic point of lower part is respectively mapped in the first, the 3rd image lower molecules image.In the first, the 3rd image lower molecules
In image, the control point before and after being deformed using initial characteristic point and mapping point as thin plate spline, to the first, the 3rd image
Do deformation process.The result that thin plate spline deformation is done to the 3rd image is as shown in Figure 4.
Thin plate spline transforming function transformation function represents as follows:
In formula, φ (ri)=ri 2log riFor RBF, whereinX and y are respectively
The transverse and longitudinal coordinate of pixel, x in two imagesiAnd yiIt is the transverse and longitudinal coordinate at control point, w respectively1iAnd w2iRepresent RBF
Weights;a11、a12、a21、a22、b1、b2It is 6 affine coefficients, n is the number at control point.
Step 8, first after deformation, the 3rd image transformed into reference picture via homography matrix H1 and H2 respectively,
I.e. in the coordinate system of the second image, the scene in such 3 width image falls under unified coordinate system, realizes pair of image
Together, panorama sketch has been obtained.
Fig. 5 shows the location and shape using 3 width images after global projective transformation alignment.Fig. 6 shows adjacent two
The splicing of width image, Fig. 7 shows final panorama sketch.From figure 7 it can be seen that the inventive method has " protecting straightforward "
Effect, projection distortion are smaller, and obtained panorama sketch whole structure and alignment details is all preferable.
Claims (6)
1. a kind of method for panoramic imaging based on projective transformation and thin plate spline, it is characterised in that comprise the following steps:
Step 1, three width images are from left to right gathered, be followed successively by the first image, the second image and the 3rd image, adjacent two images
Exist in the horizontal direction and partly overlap;
Step 2, characteristic point extracted to each image using sift algorithms, and generate feature description vectors;
Step 3, using brute-force searching algorithm combining ratio verification methods the characteristic point of adjacent image is matched;
Step 4, screened using many structure fitting algorithms are further to the characteristic matching of adjacent two images, fit Horizon
Two areal models in face and distant view;
Step 5, a horizontal boundary is determined according to the areal model of fitting, by the characteristic point of each image be divided into above and below two
Point, obtain above and below two subgraphs;
Step 6, the characteristic matching to part in adjacent two images estimate the first image, the 3rd image using DLT algorithms and arrived
The projective transformation matrix of second image, and the characteristic point of part under the second image is respectively mapped to according to projective transformation matrix
First, in the 3rd image lower molecules image;
Step 7, using the characteristic point and mapping point of the first, the 3rd image lower molecules image as thin plate spline deformation before and after
Control point, based on thin plate spline function to the first, the 3rd image carry out deformation process;
Step 8, according to projection matrix first after deformation, the 3rd image are transformed in the coordinate system of the second image so that 3 width
Two parts are obtained for alignment above and below image, obtain panorama sketch.
2. the method for panoramic imaging according to claim 1 based on projective transformation and thin plate spline, it is characterised in that step
There is 20%-60% overlapping region in 1 between adjacent image.
3. the method for panoramic imaging according to claim 1 based on projective transformation and thin plate spline, it is characterised in that step
3 progress Feature Points Matchings specific methods be:
Step 3.1, the feature description vectors set to adjacent two images do Euclidean distance in searching loop, selection two images
Less than distance threshold characteristic vector to being used as unidirectional matching candidate;
Step 3.2, two arest neighbors description vectors from each characteristic vector of unidirectional matching candidate concentration selection, only when first is near
When the Euclidean distance and the ratio of the second neighbour that neighbour matches are less than fractional threshold, then confirm it is correct matching, otherwise confirmation is
The matching of mistake, is rejected.
4. the method for panoramic imaging according to claim 1 based on projective transformation and thin plate spline, it is characterised in that step
4 many structure fitting algorithms use the information guidance of residual error sequence acquisition and accelerate to assume sampling process.
5. the method for panoramic imaging according to claim 1 based on projective transformation and thin plate spline, it is characterised in that step
6 DLT algorithmic notations are as follows:
AiH=0
Wherein,X and y, u and v are feature in two images respectively
The transverse and longitudinal coordinate of point, h=(h1,h2,h3,h4,h5,h6,h7,h8,h9)TRepresent the element of projective transformation matrix.
6. the method for panoramic imaging according to claim 1 based on projective transformation and thin plate spline, it is characterised in that step
Thin plate spline transforming function transformation function represents as follows in 7:
<mrow>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msub>
<mi>w</mi>
<mrow>
<mn>1</mn>
<mi>i</mi>
</mrow>
</msub>
<mi>&phi;</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>r</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>a</mi>
<mn>11</mn>
</msub>
<mi>x</mi>
<mo>+</mo>
<msub>
<mi>a</mi>
<mn>12</mn>
</msub>
<mi>y</mi>
<mo>+</mo>
<msub>
<mi>b</mi>
<mn>1</mn>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msub>
<mi>w</mi>
<mrow>
<mn>2</mn>
<mi>i</mi>
</mrow>
</msub>
<mi>&phi;</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>r</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>a</mi>
<mn>21</mn>
</msub>
<mi>x</mi>
<mo>+</mo>
<msub>
<mi>a</mi>
<mn>22</mn>
</msub>
<mi>y</mi>
<mo>+</mo>
<msub>
<mi>b</mi>
<mn>2</mn>
</msub>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
In formula, φ (ri)=ri 2log riFor RBF, whereinX and y are two width respectively
The transverse and longitudinal coordinate of pixel, x in imageiAnd yiIt is the transverse and longitudinal coordinate w at control point respectively1iAnd w2iRepresent the power of RBF
Value;a11、a12、a21、a22、b1、b2It is 6 affine coefficients, n is the number at control point.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710459891.0A CN107330856B (en) | 2017-06-17 | 2017-06-17 | Panoramic imaging method based on projective transformation and thin plate spline |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710459891.0A CN107330856B (en) | 2017-06-17 | 2017-06-17 | Panoramic imaging method based on projective transformation and thin plate spline |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107330856A true CN107330856A (en) | 2017-11-07 |
CN107330856B CN107330856B (en) | 2020-11-13 |
Family
ID=60195224
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710459891.0A Active CN107330856B (en) | 2017-06-17 | 2017-06-17 | Panoramic imaging method based on projective transformation and thin plate spline |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107330856B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109544447A (en) * | 2018-10-26 | 2019-03-29 | 广西师范大学 | A kind of image split-joint method, device and storage medium |
CN110059711A (en) * | 2019-01-28 | 2019-07-26 | 阿里巴巴集团控股有限公司 | Alignment schemes, device and the equipment of image |
CN112215749A (en) * | 2020-04-30 | 2021-01-12 | 北京的卢深视科技有限公司 | Image splicing method, system and equipment based on cylindrical projection and storage medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7623731B2 (en) * | 2005-06-20 | 2009-11-24 | Honda Motor Co., Ltd. | Direct method for modeling non-rigid motion with thin plate spline transformation |
CN102208025A (en) * | 2011-05-27 | 2011-10-05 | 中国科学院自动化研究所 | Method for correcting geometric distortion of text image |
US8055028B2 (en) * | 2007-02-14 | 2011-11-08 | Samsung Electronics Co., Ltd. | Object pose normalization method and apparatus and object recognition method |
US20170046833A1 (en) * | 2015-08-10 | 2017-02-16 | The Board Of Trustees Of The Leland Stanford Junior University | 3D Reconstruction and Registration of Endoscopic Data |
CN106548462A (en) * | 2016-11-02 | 2017-03-29 | 西安电子科技大学 | Non-linear SAR image geometric correction method based on thin-plate spline interpolation |
CN106657789A (en) * | 2016-12-29 | 2017-05-10 | 核动力运行研究所 | Thread panoramic image synthesis method |
CN106780338A (en) * | 2016-12-27 | 2017-05-31 | 南京理工大学 | Based on anisotropic quick super-resolution method for reconstructing |
-
2017
- 2017-06-17 CN CN201710459891.0A patent/CN107330856B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7623731B2 (en) * | 2005-06-20 | 2009-11-24 | Honda Motor Co., Ltd. | Direct method for modeling non-rigid motion with thin plate spline transformation |
US8055028B2 (en) * | 2007-02-14 | 2011-11-08 | Samsung Electronics Co., Ltd. | Object pose normalization method and apparatus and object recognition method |
CN102208025A (en) * | 2011-05-27 | 2011-10-05 | 中国科学院自动化研究所 | Method for correcting geometric distortion of text image |
US20170046833A1 (en) * | 2015-08-10 | 2017-02-16 | The Board Of Trustees Of The Leland Stanford Junior University | 3D Reconstruction and Registration of Endoscopic Data |
CN106548462A (en) * | 2016-11-02 | 2017-03-29 | 西安电子科技大学 | Non-linear SAR image geometric correction method based on thin-plate spline interpolation |
CN106780338A (en) * | 2016-12-27 | 2017-05-31 | 南京理工大学 | Based on anisotropic quick super-resolution method for reconstructing |
CN106657789A (en) * | 2016-12-29 | 2017-05-10 | 核动力运行研究所 | Thread panoramic image synthesis method |
Non-Patent Citations (3)
Title |
---|
QU H M等: "《Thermal management technology of high-power light-emitting diodes for automotive headlights》", 《 IEICE ELECTRONICS EXPRESS》 * |
STAYTON C T: "《 Application of thin‐plate spline transformations to finite element models, or, how to turn a bog turtle into a spotted turtle to analyze both》", 《 EVOLUTION: INTERNATIONAL JOURNAL OF ORGANIC EVOLUTION》 * |
颜无瑕 等: "《三次透视投影变换图像拼接》", 《光学精密工程》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109544447A (en) * | 2018-10-26 | 2019-03-29 | 广西师范大学 | A kind of image split-joint method, device and storage medium |
CN109544447B (en) * | 2018-10-26 | 2022-10-21 | 广西师范大学 | Image splicing method and device and storage medium |
CN110059711A (en) * | 2019-01-28 | 2019-07-26 | 阿里巴巴集团控股有限公司 | Alignment schemes, device and the equipment of image |
CN112215749A (en) * | 2020-04-30 | 2021-01-12 | 北京的卢深视科技有限公司 | Image splicing method, system and equipment based on cylindrical projection and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN107330856B (en) | 2020-11-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10540806B2 (en) | Systems and methods for depth-assisted perspective distortion correction | |
Steedly et al. | Efficiently registering video into panoramic mosaics | |
US9576403B2 (en) | Method and apparatus for fusion of images | |
US9900505B2 (en) | Panoramic video from unstructured camera arrays with globally consistent parallax removal | |
US8447140B1 (en) | Method and apparatus for estimating rotation, focal lengths and radial distortion in panoramic image stitching | |
CN104732482B (en) | A kind of multi-resolution image joining method based on control point | |
US6348918B1 (en) | Stereo reconstruction employing a layered approach | |
CN110111250B (en) | Robust automatic panoramic unmanned aerial vehicle image splicing method and device | |
US20060082644A1 (en) | Image processing apparatus and image processing program for multi-viewpoint image | |
US20110170784A1 (en) | Image registration processing apparatus, region expansion processing apparatus, and image quality improvement processing apparatus | |
CN111553841B (en) | Real-time video splicing method based on optimal suture line updating | |
CN107330856A (en) | A kind of method for panoramic imaging based on projective transformation and thin plate spline | |
CN107220955A (en) | A kind of brightness of image equalization methods based on overlapping region characteristic point pair | |
CN109801212A (en) | A kind of fish eye images joining method based on SIFT feature | |
Bartoli et al. | From video sequences to motion panoramas | |
Pham et al. | Fast and efficient method for large-scale aerial image stitching | |
CN112862683A (en) | Adjacent image splicing method based on elastic registration and grid optimization | |
CN112801870A (en) | Image splicing method based on grid optimization, splicing system and readable storage medium | |
US20220207679A1 (en) | Method and apparatus for stitching images | |
Jin | A three-point minimal solution for panoramic stitching with lens distortion | |
CN110580715A (en) | Image alignment method based on illumination constraint and grid deformation | |
CN110147809B (en) | Image processing method and device, storage medium and image equipment | |
Luo et al. | Drone image stitching guided by robust elastic warping and locality preserving matching | |
KR101028171B1 (en) | Determination of aspect ratio from perspective projection images | |
Dervişoğlu et al. | Interpolation-based smart video stabilization |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |